kpriv/pkgsrc
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
pkgsrc/mk/bsd.pkg.mk

5108 lines
165 KiB
Makefile
Raw Normal View History

* Push the imake- and xmkmf-handling into the new tools framework. * Get rid of an explicit check for ${_IMAKE_MAKE} == ${GMAKE} in bsd.pkg.mk to check for whether we need to depend on gmake or not. Instead, we now note in Linux.mk that packages that need imake will also need to use gmake by setting _IMAKE_TOOLS+=gmake. * Push the definition of MAKE_PROGRAM from bsd.pkg.mk into make.mk where it's closer to related code.
2005-05-02 23:10:02 +02:00
# $NetBSD: bsd.pkg.mk,v 1.1622 2005/05/02 21:10:02 jlam Exp $
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
#
move documentation to packages.7. also, remove ts=4 crap.
1998-06-05 14:45:53 +02:00
# This file is in the public domain.
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
#
move documentation to packages.7. also, remove ts=4 crap.
1998-06-05 14:45:53 +02:00
# This file is derived from bsd.port.mk - 940820 Jordan K. Hubbard.
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
#
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
# Please see the pkgsrc/doc/guide manual for details on the
# variables used in this make file template.
#
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Default sequence for "all" is: fetch checksum extract patch configure build
#
# Please read the comments in the targets section below, you
# should be able to use the pre-* or post-* targets/scripts
# (which are available for every stage except checksum) or
# override the do-* targets to do pretty much anything you want.
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
############################################################################
# Include any preferences, if not already included, and common definitions
############################################################################
Implement some parts of proposals posted to tech-pkg: - Define OS_VERSION in bsd.prefs.mk (...if not already defined). Previously was only set in a conditional block in bsd.pkg.mk. - Define new values, MACHINE_PLATFORM and MACHINE_GNU_PLATFORM, which default to ${OPSYS}-${OS_VERSION}-${MACHINE_ARCH} and ${LOWER_OPSYS}-${LOWER_VENDOR}-${LOWER_ARCH} respectively. (Probably should nuke LOWER_ARCH and rename LOWER_VENDOR to something like MACHINE_GNU_VENDOR.) - Use ${MACHINE_GNU_PLATFORM} as the argument to `--host' in GNU_CONFIGURE packages. - Allow for ONLY_FOR_PLATFORM and NOT_FOR_PLATFORM, lists of patterns in csh(1) glob style which are matched against MACHINE_PLATFORM above and take the place of {ONLY,NOT}_FOR_{ARCHS,OPSYS}. (For now, until all pkgs are converted, provide backwards compatibility.) *Both* may be defined, in which case NOT_FOR_PLATFORM takes precedence. - Also do some significant aesthetic cleanup, based on the idea `define platform-specific overrides first, then define a generic case with the ?= operator'. Lots more of this to do yet.
1999-05-24 20:42:00 +02:00
.include "../../mk/bsd.prefs.mk"
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
.include "../../mk/bsd.hacks.mk"
Prevent /etc/mk.conf from being included by BSD-style makefiles used by packages themselves, as /etc/mk.conf may contain harmful settings for package builds. Just to clarify, /etc/mk.conf is still included by the Makefiles in pkgsrc, just not by the Makefiles used to build the software itself. This should fix pkg/15297 by gabriel rosenkoetter <gr@grappa.eclipsed.net>.
2002-01-27 21:01:56 +01:00
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
# This has to come first to avoid showing all BUILD_DEFS added by this
# Makefile, which are usually not customizable.
Sometimes, the dependencies installed are a function of which variables are set, and it'd be nice to know which ones before the default dependencies are installed. Create a new target "pre-install-depends" that is executed before dependencies are installed. Move the display of BUILD_DEFS (build-defs-message) into pre-install-depends so that they are displayed before dependencies are installed.
2004-08-05 01:22:38 +02:00
.PHONY: pre-install-depends build-defs-message
pre-install-depends: build-defs-message
oops.. reversed a test
2004-08-05 01:29:50 +02:00
.if empty(PKGSRC_SHOW_BUILD_DEFS:M[yY][eE][sS])
Re-model the BUILD_DEFS display during build: if PKGSRC_SHOW_BUILD_DEFS is defined, then show the BUILD_DEFS to the user. If not, then just keep quiet. Whilst I'm here, re-indent the cpp-like .if/.else/.endif/.for/.endfor lines properly. Also, remove the ":u" modifier in the build-defs-message code - it's not strictly necessary (duplicate definitions will be displayed, but this isn't a major problem), and it removes the requirement to have a make(1) binary which supports ":u". This should fix PR 22402 from Alan Barrett, and also addresses some concerns raised by Robert Elz.
2003-08-08 10:43:37 +02:00
build-defs-message:
.elif !target(build-defs-message)
Modify the build-defs-message target so that it is only printed once, in the pre-fetch stage. Do this by creating a cookie file inside the WRKDIR. Also let the user override the target from mk.conf to disable these messages. This fixes PR pkg/22305 by Robert Elz.
2003-07-30 17:04:21 +02:00
build-defs-message: ${WRKDIR}
Re-model the BUILD_DEFS display during build: if PKGSRC_SHOW_BUILD_DEFS is defined, then show the BUILD_DEFS to the user. If not, then just keep quiet. Whilst I'm here, re-indent the cpp-like .if/.else/.endif/.for/.endfor lines properly. Also, remove the ":u" modifier in the build-defs-message code - it's not strictly necessary (duplicate definitions will be displayed, but this isn't a major problem), and it removes the requirement to have a make(1) binary which supports ":u". This should fix PR 22402 from Alan Barrett, and also addresses some concerns raised by Robert Elz.
2003-08-08 10:43:37 +02:00
. if defined(BUILD_DEFS) && !empty(BUILD_DEFS)
. if !exists(${WRKDIR}/.bdm_done)
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
@${ECHO} "=========================================================================="
@${ECHO} "The following variables will affect the build process of this package,"
@${ECHO} "${PKGNAME}. Their current value is shown below:"
@${ECHO} ""
Re-model the BUILD_DEFS display during build: if PKGSRC_SHOW_BUILD_DEFS is defined, then show the BUILD_DEFS to the user. If not, then just keep quiet. Whilst I'm here, re-indent the cpp-like .if/.else/.endif/.for/.endfor lines properly. Also, remove the ":u" modifier in the build-defs-message code - it's not strictly necessary (duplicate definitions will be displayed, but this isn't a major problem), and it removes the requirement to have a make(1) binary which supports ":u". This should fix PR 22402 from Alan Barrett, and also addresses some concerns raised by Robert Elz.
2003-08-08 10:43:37 +02:00
. for var in ${BUILD_DEFS:O}
. if !defined(${var})
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
@${ECHO} " * ${var} (not defined)"
Re-model the BUILD_DEFS display during build: if PKGSRC_SHOW_BUILD_DEFS is defined, then show the BUILD_DEFS to the user. If not, then just keep quiet. Whilst I'm here, re-indent the cpp-like .if/.else/.endif/.for/.endfor lines properly. Also, remove the ":u" modifier in the build-defs-message code - it's not strictly necessary (duplicate definitions will be displayed, but this isn't a major problem), and it removes the requirement to have a make(1) binary which supports ":u". This should fix PR 22402 from Alan Barrett, and also addresses some concerns raised by Robert Elz.
2003-08-08 10:43:37 +02:00
. elif defined(${var}) && empty(${var})
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
@${ECHO} " * ${var} (defined)"
Re-model the BUILD_DEFS display during build: if PKGSRC_SHOW_BUILD_DEFS is defined, then show the BUILD_DEFS to the user. If not, then just keep quiet. Whilst I'm here, re-indent the cpp-like .if/.else/.endif/.for/.endfor lines properly. Also, remove the ":u" modifier in the build-defs-message code - it's not strictly necessary (duplicate definitions will be displayed, but this isn't a major problem), and it removes the requirement to have a make(1) binary which supports ":u". This should fix PR 22402 from Alan Barrett, and also addresses some concerns raised by Robert Elz.
2003-08-08 10:43:37 +02:00
. else
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
@${ECHO} " * ${var} = ${${var}}"
Re-model the BUILD_DEFS display during build: if PKGSRC_SHOW_BUILD_DEFS is defined, then show the BUILD_DEFS to the user. If not, then just keep quiet. Whilst I'm here, re-indent the cpp-like .if/.else/.endif/.for/.endfor lines properly. Also, remove the ":u" modifier in the build-defs-message code - it's not strictly necessary (duplicate definitions will be displayed, but this isn't a major problem), and it removes the requirement to have a make(1) binary which supports ":u". This should fix PR 22402 from Alan Barrett, and also addresses some concerns raised by Robert Elz.
2003-08-08 10:43:37 +02:00
. endif
. endfor
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
@${ECHO} ""
wrap a long line in BUILD_DEFS output, tweak the message slightly.
2004-03-27 09:06:24 +01:00
@${ECHO} "You may want to abort the process now with CTRL-C and change their value"
@${ECHO} "before continuing. Be sure to run \`${MAKE} clean' after"
@${ECHO} "the changes."
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
@${ECHO} "=========================================================================="
Modify the build-defs-message target so that it is only printed once, in the pre-fetch stage. Do this by creating a cookie file inside the WRKDIR. Also let the user override the target from mk.conf to disable these messages. This fixes PR pkg/22305 by Robert Elz.
2003-07-30 17:04:21 +02:00
@${TOUCH} ${WRKDIR}/.bdm_done
Re-model the BUILD_DEFS display during build: if PKGSRC_SHOW_BUILD_DEFS is defined, then show the BUILD_DEFS to the user. If not, then just keep quiet. Whilst I'm here, re-indent the cpp-like .if/.else/.endif/.for/.endfor lines properly. Also, remove the ":u" modifier in the build-defs-message code - it's not strictly necessary (duplicate definitions will be displayed, but this isn't a major problem), and it removes the requirement to have a make(1) binary which supports ":u". This should fix PR 22402 from Alan Barrett, and also addresses some concerns raised by Robert Elz.
2003-08-08 10:43:37 +02:00
. endif
. endif
Add the build-defs-message target, which shows a list of all variables that can be customized for the package beeing built. This list is extracted from the contents of BUILD_DEFS, and the message is not shown if it's empty. Each variable is printed together with its actual value. The new target is called in the pre-fetch stage, so it's shown early to the user.
2003-07-28 20:15:56 +02:00
.endif
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
############################################################################
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
# Transform package Makefile variables and set defaults
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
############################################################################
Fail-safe check for circular dependencies.
2003-10-02 16:41:42 +02:00
- Move all CHECK_FILES_SKIP definitions to a central place in bsd.pkg.mk. - Move the previous block of code down in the file so that all used variables are defined (specially PKG_SYSCONFDIR). Fixes a problem noticed by kristerw@'s bulk build in comms/minicom. - Ignore diff's return code, which aborts make in NetBSD 1.6.2. Also noticed by kristerw@'s bulk build. - Use full paths to do the checks, instead of relative to ${PREFIX}. Less ambiguity. Matches should be turned into regular expressions that anchor to a whole line (tried that, but found some problems). - Turn CHECK_FILES to NO by default. As said in the previous point, there are still some problems that have to be fixed and minor improvements to be done. And I have no time to fix this ATM. Yes, this definitely needs more testing. I'm sorry for all the noise. (But hey! you should set and try this feature locally! ;-)
2005-01-15 11:50:17 +01:00
CHECK_FILES?= NO # run check-files after install
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
CHECK_FILES_STRICT?= NO # make check-files very strict on errors
* add CHECK_SHLIBS variable that can be set to "no" in emul/compat packages, to precent check-shlibs from discovering that ldd(8) can't deal with e.g. Linux binaries * On "make reinstall", make sure the package gets deinstalled first. No more 'make deinstall ; make reinstall'.
2000-08-31 04:32:32 +02:00
CHECK_SHLIBS?= YES # run check-shlibs after install
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
CLEANDEPENDS?= NO
DEINSTALLDEPENDS?= NO # add -R to pkg_delete
MKCRYPTO?= YES # build crypto packages by default
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
NOCLEAN?= NO # don't clean up after update
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
REINSTALL?= NO # reinstall upon update
SHLIB_HANDLING?= YES # do automatic shared lib handling
##### Variant spellings
.if defined(LICENCE) && !defined(LICENSE)
LICENSE= ${LICENCE}
.endif
.if defined(ACCEPTABLE_LICENCES) && !defined(ACCEPTABLE_LICENSES)
ACCEPTABLE_LICENSES= ${ACCEPTABLE_LICENCES}
.endif
##### PKGBASE, PKGNAME[_NOREV], PKGVERSION
Implement some parts of proposals posted to tech-pkg: - Define OS_VERSION in bsd.prefs.mk (...if not already defined). Previously was only set in a conditional block in bsd.pkg.mk. - Define new values, MACHINE_PLATFORM and MACHINE_GNU_PLATFORM, which default to ${OPSYS}-${OS_VERSION}-${MACHINE_ARCH} and ${LOWER_OPSYS}-${LOWER_VENDOR}-${LOWER_ARCH} respectively. (Probably should nuke LOWER_ARCH and rename LOWER_VENDOR to something like MACHINE_GNU_VENDOR.) - Use ${MACHINE_GNU_PLATFORM} as the argument to `--host' in GNU_CONFIGURE packages. - Allow for ONLY_FOR_PLATFORM and NOT_FOR_PLATFORM, lists of patterns in csh(1) glob style which are matched against MACHINE_PLATFORM above and take the place of {ONLY,NOT}_FOR_{ARCHS,OPSYS}. (For now, until all pkgs are converted, provide backwards compatibility.) *Both* may be defined, in which case NOT_FOR_PLATFORM takes precedence. - Also do some significant aesthetic cleanup, based on the idea `define platform-specific overrides first, then define a generic case with the ?= operator'. Lots more of this to do yet.
1999-05-24 20:42:00 +02:00
Oops; use ?= for PKGBASE and PKGVERSION, so pkgs can derive their PKGNAMEs from them.
2000-06-03 21:03:20 +02:00
PKGBASE?= ${PKGNAME:C/-[^-]*$//}
PKGVERSION?= ${PKGNAME:C/^.*-//}
Move definition of PKGNAME and PKGNAME_NOREV higher in the file since their expanded values are used somewhere in the middle.
2004-09-15 17:26:10 +02:00
.if defined(PKGREVISION) && !empty(PKGREVISION) && (${PKGREVISION} != "0")
. if defined(PKGNAME)
PKGNAME_NOREV:= ${PKGNAME}
PKGNAME:= ${PKGNAME}nb${PKGREVISION}
. else
PKGNAME?= ${DISTNAME}nb${PKGREVISION}
PKGNAME_NOREV= ${DISTNAME}
. endif
.else
PKGNAME?= ${DISTNAME}
PKGNAME_NOREV= ${PKGNAME}
.endif
Add PKGPATH, PKGBASE, PKGVERSION and PKGWILDCARD make(1) variables, so that the substitution hacks are all done in one place. Protect a few things with double quotes just to be safe. Fix `readme' target that was left broken in last commit.
2000-06-03 20:51:34 +02:00
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
##### PLIST
.if ${PKG_INSTALLATION_TYPE} == "pkgviews"
PLIST_TYPE?= dynamic
.endif
PLIST_TYPE?= static
.if !defined(PLIST_SRC)
. if exists(${PKGDIR}/PLIST.common)
PLIST_SRC+= ${PKGDIR}/PLIST.common
. endif
. if exists(${PKGDIR}/PLIST.${OPSYS})
PLIST_SRC+= ${PKGDIR}/PLIST.${OPSYS}
Fix PLIST_SRC error introduced in rev 1.1593. Before this fix it would include PLIST.common then _only_ one of the following PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST and then PLIST.common_end. Now, uses all of the following PLIST files, in that order: PLIST.common PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST PLIST.common_end
2005-02-27 23:10:25 +01:00
. endif
. if exists(${PKGDIR}/PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g})
Change the files to be used for PLIST_SRC from PLIST.common PLIST.${OPSYS} PLIST PLIST.common_end to PLIST.common PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST PLIST.common_end Which makes it possible to remove duplicated code in some packages. This doesn't change anything, because all affected packages define PLIST_SRC, so this code would never be used for them. And most of them will be fixed with the next commit. ToDO: ${MACHINE_ARCH:C/i[3-6]86/i386/g} looks wrong and should be changed, but thats what we allready use for MESSAGE_SRC.
2005-02-21 21:12:01 +01:00
PLIST_SRC+= ${PKGDIR}/PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g}
Fix PLIST_SRC error introduced in rev 1.1593. Before this fix it would include PLIST.common then _only_ one of the following PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST and then PLIST.common_end. Now, uses all of the following PLIST files, in that order: PLIST.common PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST PLIST.common_end
2005-02-27 23:10:25 +01:00
. endif
. if exists(${PKGDIR}/PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g})
Change the files to be used for PLIST_SRC from PLIST.common PLIST.${OPSYS} PLIST PLIST.common_end to PLIST.common PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST PLIST.common_end Which makes it possible to remove duplicated code in some packages. This doesn't change anything, because all affected packages define PLIST_SRC, so this code would never be used for them. And most of them will be fixed with the next commit. ToDO: ${MACHINE_ARCH:C/i[3-6]86/i386/g} looks wrong and should be changed, but thats what we allready use for MESSAGE_SRC.
2005-02-21 21:12:01 +01:00
PLIST_SRC+= ${PKGDIR}/PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g}
Fix PLIST_SRC error introduced in rev 1.1593. Before this fix it would include PLIST.common then _only_ one of the following PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST and then PLIST.common_end. Now, uses all of the following PLIST files, in that order: PLIST.common PLIST.${OPSYS} PLIST.${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g} PLIST PLIST.common_end
2005-02-27 23:10:25 +01:00
. endif
. if exists(${PKGDIR}/PLIST)
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
PLIST_SRC+= ${PKGDIR}/PLIST
. endif
. if exists(${PKGDIR}/PLIST.common_end)
PLIST_SRC+= ${PKGDIR}/PLIST.common_end
. endif
.endif # !PLIST_SRC
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
##### Others
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
_DISTDIR?= ${DISTDIR}/${DIST_SUBDIR}
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
BUILD_DEPENDS?= # empty
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
BUILD_TARGET?= all
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
COMMENT?= (no description)
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
CONFIGURE_DIRS?= ${WRKSRC}
CONFIGURE_SCRIPT?= ./configure
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
DEPENDS?= # empty
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
DESCR_SRC?= ${PKGDIR}/DESCR
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
DIGEST_ALGORITHMS?= SHA1 RMD160
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
DISTFILES?= ${DISTNAME}${EXTRACT_SUFX}
DISTINFO_FILE?= ${PKGDIR}/distinfo
EXTRACT_ONLY?= ${DISTFILES}
EXTRACT_SUFX?= .tar.gz
INSTALL_DIRS?= ${BUILD_DIRS}
INSTALL_MAKE_FLAGS?= ${MAKE_FLAGS}
INSTALL_TARGET?= install
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
INTERACTIVE_STAGE?= none
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
MAINTAINER?= tech-pkg@NetBSD.org
MAKE_FLAGS?= # empty
MAKEFILE?= Makefile
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
PATCH_DIGEST_ALGORITHM?=SHA1
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
PKG_SUFX?= .tgz
PKGFILE?= ${PKGREPOSITORY}/${PKGNAME}${PKG_SUFX}
PKGREPOSITORY?= ${PACKAGES}/${PKGREPOSITORYSUBDIR}
PKGREPOSITORYSUBDIR?= All
PKGWILDCARD?= ${PKGBASE}-[0-9]*
SVR4_PKGNAME?= ${PKGNAME}
USE_DIGEST?= YES
USE_GNU_TOOLS?= # empty
WRKSRC?= ${WRKDIR}/${DISTNAME}
Fix a gleaming illustration of why using YES/NO variables everywhere possible, just for the sake of doing so, is not a good thing to do: The platform files define _STRIPFLAG_* to determine whether to strip things. But since this is included in bsd.prefs.mk, ".if ..." checks cannot take things set in the Makefile into account. So convert INSTALL_UNSTRIPPED=YES to a defined/undefined variable check in bsd.pkg.mk, and use the :D:U idiom in the _STRIPFLAG_* variables. This should fix PR pkg/28772 and PR pkg/29031.
2005-03-18 19:16:34 +01:00
.if (defined(INSTALL_UNSTRIPPED) && !empty(INSTALL_UNSTRIPPED:M[yY][eE][sS])) || defined(DEBUG_FLAGS)
_INSTALL_UNSTRIPPED= # set (flag used by platform/*.mk)
.endif
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
##### Non-overridable constants
# Latest versions of tools required for correct pkgsrc operation.
DIGEST_REQD= 20010302
PKGTOOLS_REQD= ${_OPSYS_PKGTOOLS_REQD:U20050204}
PKG_DB_TMPDIR= ${WRKDIR}/.pkgdb
DDIR= ${WRKDIR}/.DDIR
DESCR= ${PKG_DB_TMPDIR}/+DESC
DLIST= ${WRKDIR}/.DLIST
PLIST= ${WRKDIR}/.PLIST
# Files to create for versioning and build information
BUILD_VERSION_FILE= ${PKG_DB_TMPDIR}/+BUILD_VERSION
BUILD_INFO_FILE= ${PKG_DB_TMPDIR}/+BUILD_INFO
# Files containing size of pkg w/o and w/ all required pkgs
SIZE_PKG_FILE= ${PKG_DB_TMPDIR}/+SIZE_PKG
SIZE_ALL_FILE= ${PKG_DB_TMPDIR}/+SIZE_ALL
# File to denote "no deletion of a package"
PRESERVE_FILE= ${PKG_DB_TMPDIR}/+PRESERVE
##### Transform USE_* into dependencies
.include "../../mk/bsd.pkg.use.mk"
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
############################################################################
# Sanity checks
############################################################################
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
# Fail-safe in the case of circular dependencies
.if defined(_PKGSRC_DEPS) && defined(PKGNAME) && !empty(_PKGSRC_DEPS:M${PKGNAME})
Whitespace formatting nits.
2005-02-11 18:00:07 +01:00
PKG_FAIL_REASON+= "Circular dependency detected"
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
.endif
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
# PKG_INSTALLATION_TYPE can only be one of two values: "pkgviews" or
# "overwrite".
.if (${PKG_INSTALLATION_TYPE} != "pkgviews") && \
(${PKG_INSTALLATION_TYPE} != "overwrite")
PKG_FAIL_REASON+= "PKG_INSTALLATION_TYPE must be \`\`pkgviews'' or \`\`overwrite''."
.endif
.if empty(PKG_INSTALLATION_TYPES:M${PKG_INSTALLATION_TYPE})
PKG_FAIL_REASON+= "This package doesn't support PKG_INSTALLATION_TYPE=${PKG_INSTALLATION_TYPE}."
.endif
.if (${PLIST_TYPE} != "dynamic") && (${PLIST_TYPE} != "static")
PKG_FAIL_REASON+= "PLIST_TYPE must be \`\`dynamic'' or \`\`static''."
.endif
.if (${PKG_INSTALLATION_TYPE} == "overwrite") && (${PLIST_TYPE} != "static")
PKG_FAIL_REASON+= "PLIST_TYPE must be \`\`static'' for \`\`overwrite'' packages."
.endif
# Check that we are using up-to-date pkg_* tools with this file.
.if !defined(NO_PKGTOOLS_REQD_CHECK)
. if ${PKGTOOLS_VERSION} < ${PKGTOOLS_REQD}
PKG_FAIL_REASON+='Error: The package tools installed on this system are out of date.'
PKG_FAIL_REASON+='The installed package tools are dated ${PKGTOOLS_VERSION:C|(....)(..)(..)|\1/\2/\3|} and you must update'
PKG_FAIL_REASON+='them to at least ${PKGTOOLS_REQD:C|(....)(..)(..)|\1/\2/\3|} using the following command:'
PKG_FAIL_REASON+=''
PKG_FAIL_REASON+=' cd ${PKGSRCDIR}/pkgtools/pkg_install && ${MAKE} clean && ${MAKE} install'
. endif
.endif # !NO_PKGTOOLS_REQD_CHECK
.if defined(ALL_TARGET)
PKG_FAIL_REASON+='ALL_TARGET is deprecated and must be replaced with BUILD_TARGET.'
.endif
.if defined(NO_WRKSUBDIR)
PKG_FAIL_REASON+='NO_WRKSUBDIR has been deprecated - please replace it with an explicit'
PKG_FAIL_REASON+='assignment of WRKSRC= $${WRKDIR}'
.endif # NO_WRKSUBDIR
# We need to make sure the buildlink-x11 package is not installed since it
# breaks builds that use imake.
.if defined(USE_IMAKE)
. if exists(${LOCALBASE}/lib/X11/config/buildlinkX11.def) || \
exists(${X11BASE}/lib/X11/config/buildlinkX11.def)
PKG_FAIL_REASON+= "${PKGNAME} uses imake, but the buildlink-x11 package was found." \
" Please deinstall it (pkg_delete buildlink-x11)."
. endif
.endif # USE_IMAKE
.if !defined(CATEGORIES) || !defined(DISTNAME)
PKG_FAIL_REASON+='CATEGORIES and DISTNAME are mandatory.'
.endif
.if defined(LIB_DEPENDS)
PKG_FAIL_REASON+='LIB_DEPENDS is deprecated and must be replaced with DEPENDS.'
.endif
.if defined(PKG_PATH)
PKG_FAIL_REASON+='Please unset PKG_PATH before doing pkgsrc work!'
.endif
.if defined(MASTER_SITE_SUBDIR)
PKG_FAIL_REASON+='MASTER_SITE_SUBDIR is deprecated and must be replaced with MASTER_SITES.'
.endif
.if defined(PATCH_SITE_SUBDIR)
PKG_FAIL_REASON+='PATCH_SITE_SUBDIR is deprecated and must be replaced with PATCH_SITES.'
.endif
.if defined(ONLY_FOR_ARCHS) || defined(NOT_FOR_ARCHS) \
|| defined(ONLY_FOR_OPSYS) || defined(NOT_FOR_OPSYS)
PKG_FAIL_REASON+='ONLY/NOT_FOR_ARCHS/OPSYS are deprecated and must be replaced with ONLY/NOT_FOR_PLATFORM.'
.endif
.if (${PKGSRC_LOCKTYPE} == "sleep" || ${PKGSRC_LOCKTYPE} == "once")
Whitespace formatting nits.
2005-02-11 18:00:07 +01:00
. if !defined(OBJHOSTNAME)
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
PKG_FAIL_REASON+='PKGSRC_LOCKTYPE needs OBJHOSTNAME defined.'
Whitespace formatting nits.
2005-02-11 18:00:07 +01:00
. elif !exists(${SHLOCK})
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
PKG_FAIL_REASON+='The ${SHLOCK} utility does not exist, and is necessary for locking.'
PKG_FAIL_REASON+='Please "${MAKE} install" in ../../pkgtools/shlock.'
Whitespace formatting nits.
2005-02-11 18:00:07 +01:00
. endif
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
.endif
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
# Allow variables to be set on a per-OS basis
OPSYSVARS+= CFLAGS CXXFLAGS CPPFLAGS LDFLAGS LIBS
.for _var_ in ${OPSYSVARS:O}
. if defined(${_var_}.${OPSYS})
${_var_}+= ${${_var_}.${OPSYS}}
. elif defined(${_var_}.*)
${_var_}+= ${${_var_}.*}
. endif
.endfor
Visual cleanup from tv-derecurse: Consolidate many MAKE_ENVs and SCRIPTS_ENVs into a common block. (CONFIGURE_ENVs to be done later.) Introduce new variable ALL_ENV which is automatically included into all of MAKE_ENV, SCRIPTS_ENV, and CONFIGURE_ENV; this allows much cleaner addition of the common CC/CXX/CFLAGS/etc. variables needed by all of these.
2005-02-11 17:36:49 +01:00
CPPFLAGS+= ${CPP_PRECOMP_FLAGS}
ALL_ENV+= CC=${CC:Q}
Work around brain damaged autoconf by stripping leading and trailing whitespace from *FLAGS and LIBS before doing :Q quoting. Should fix PR pkg/29945.
2005-04-21 17:53:53 +02:00
ALL_ENV+= CFLAGS=${CFLAGS:C/^ *//:C/ *$//:Q}
ALL_ENV+= CPPFLAGS=${CPPFLAGS:C/^ *//:C/ *$//:Q}
ALL_ENV+= CXX=${CXX:C/^ *//:C/ *$//:Q}
ALL_ENV+= CXXFLAGS=${CXXFLAGS:C/^ *//:C/ *$//:Q}
Visual cleanup from tv-derecurse: Consolidate many MAKE_ENVs and SCRIPTS_ENVs into a common block. (CONFIGURE_ENVs to be done later.) Introduce new variable ALL_ENV which is automatically included into all of MAKE_ENV, SCRIPTS_ENV, and CONFIGURE_ENV; this allows much cleaner addition of the common CC/CXX/CFLAGS/etc. variables needed by all of these.
2005-02-11 17:36:49 +01:00
ALL_ENV+= COMPILER_RPATH_FLAG=${COMPILER_RPATH_FLAG:Q}
make sure we correctly set F77. Addresses PR29666
2005-03-16 03:09:54 +01:00
ALL_ENV+= F77=${FC:Q}
Visual cleanup from tv-derecurse: Consolidate many MAKE_ENVs and SCRIPTS_ENVs into a common block. (CONFIGURE_ENVs to be done later.) Introduce new variable ALL_ENV which is automatically included into all of MAKE_ENV, SCRIPTS_ENV, and CONFIGURE_ENV; this allows much cleaner addition of the common CC/CXX/CFLAGS/etc. variables needed by all of these.
2005-02-11 17:36:49 +01:00
ALL_ENV+= FC=${FC:Q}
Work around brain damaged autoconf by stripping leading and trailing whitespace from *FLAGS and LIBS before doing :Q quoting. Should fix PR pkg/29945.
2005-04-21 17:53:53 +02:00
ALL_ENV+= FFLAGS=${FFLAGS:C/^ *//:C/ *$//:Q}
Reset the current locale to a sane value (that is, 'C') during the build of packages. Several utilities behave differently or even incorrectly if a locale different than 'C' is set. One of these is perl, which spits a big warning and then breaks some packages. An example is irssi, as described in PR pkg/29613 by Hanno Wagner.
2005-03-20 19:18:09 +01:00
ALL_ENV+= LANG=C
ALL_ENV+= LC_COLLATE=C
ALL_ENV+= LC_CTYPE=C
ALL_ENV+= LC_MESSAGES=C
ALL_ENV+= LC_MONETARY=C
ALL_ENV+= LC_NUMERIC=C
ALL_ENV+= LC_TIME=C
Work around brain damaged autoconf by stripping leading and trailing whitespace from *FLAGS and LIBS before doing :Q quoting. Should fix PR pkg/29945.
2005-04-21 17:53:53 +02:00
ALL_ENV+= LDFLAGS=${LDFLAGS:C/^ *//:C/ *$//:Q}
Visual cleanup from tv-derecurse: Consolidate many MAKE_ENVs and SCRIPTS_ENVs into a common block. (CONFIGURE_ENVs to be done later.) Introduce new variable ALL_ENV which is automatically included into all of MAKE_ENV, SCRIPTS_ENV, and CONFIGURE_ENV; this allows much cleaner addition of the common CC/CXX/CFLAGS/etc. variables needed by all of these.
2005-02-11 17:36:49 +01:00
ALL_ENV+= LINKER_RPATH_FLAG=${LINKER_RPATH_FLAG:Q}
ALL_ENV+= PATH=${PATH:Q}:${LOCALBASE}/bin:${X11BASE}/bin
ALL_ENV+= PREFIX=${PREFIX}
MAKE_ENV+= ${ALL_ENV}
MAKE_ENV+= ${NO_EXPORT_CPP:D:UCPP=${CPP:Q}}
MAKE_ENV+= LINK_ALL_LIBGCC_HACK=${LINK_ALL_LIBGCC_HACK:Q}
MAKE_ENV+= LOCALBASE=${LOCALBASE}
MAKE_ENV+= NO_WHOLE_ARCHIVE_FLAG=${NO_WHOLE_ARCHIVE_FLAG:Q}
MAKE_ENV+= WHOLE_ARCHIVE_FLAG=${WHOLE_ARCHIVE_FLAG:Q}
MAKE_ENV+= X11BASE=${X11BASE}
MAKE_ENV+= X11PREFIX=${X11PREFIX}
# Constants to provide a consistent environment for packages using
# BSD-style Makefiles.
MAKE_ENV+= MAKECONF=${PKGMAKECONF:U/dev/null}
MAKE_ENV+= OBJECT_FMT=${OBJECT_FMT:Q}
MAKE_ENV+= ${USETOOLS:DUSETOOLS=${USETOOLS:Q}}
SCRIPTS_ENV+= ${ALL_ENV}
SCRIPTS_ENV+= _PKGSRCDIR=${_PKGSRCDIR}
SCRIPTS_ENV+= ${BATCH:DBATCH=yes}
SCRIPTS_ENV+= CURDIR=${.CURDIR}
SCRIPTS_ENV+= DEPENDS=${DEPENDS:Q}
SCRIPTS_ENV+= DISTDIR=${DISTDIR}
SCRIPTS_ENV+= FILESDIR=${FILESDIR}
SCRIPTS_ENV+= LOCALBASE=${LOCALBASE}
SCRIPTS_ENV+= PATCHDIR=${PATCHDIR}
SCRIPTS_ENV+= PKGSRCDIR=${PKGSRCDIR}
SCRIPTS_ENV+= SCRIPTDIR=${SCRIPTDIR}
SCRIPTS_ENV+= VIEWBASE=${VIEWBASE}
SCRIPTS_ENV+= WRKDIR=${WRKDIR}
SCRIPTS_ENV+= WRKSRC=${WRKSRC}
SCRIPTS_ENV+= X11BASE=${X11BASE}
CONFIGURE_ENV+= ${ALL_ENV}
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
# Store the result in the +BUILD_INFO file so we can query for the build
# options using "pkg_info -Q PKG_OPTIONS <pkg>".
#
.if defined(PKG_SUPPORTED_OPTIONS) && defined(PKG_OPTIONS)
BUILD_DEFS+= PKG_OPTIONS
.endif
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if empty(DEPOT_SUBDIR)
PKG_FAIL_REASON+= "DEPOT_SUBDIR may not be empty."
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
- Integrate xpkgwedge with a switch, USE_LOCALBASE_FOR_X11, which leaves X11BASE alone and, when set, sets PREFIX to LOCALBASE and adds a build dependency on xpkgwedge - Nuke last vestiges of the ${ARCH} variable, and replace it with the binary compatibility identifier ${MACHINE_ARCH}. - Include bsd.prefs.mk early in bsd.pkg.mk. (Needed by xpkgwedge change.) - Look for Motif/lesstif in both X11BASE and LOCALBASE - Add rpath (-Wl,-R) options for MOTIFLIB - Nuke obsolete EXEC_DEPENDS cruft
1998-07-31 16:55:38 +02:00
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if ${PKG_INSTALLATION_TYPE} == "pkgviews"
#
# _PLIST_IGNORE_FILES basically mirrors the list of ignored files found
# in pkg_views(1). It's used by the dynamic PLIST generator to skip
# adding the named files to the PLIST.
#
_PLIST_IGNORE_FILES= +* # package metadata files
_PLIST_IGNORE_FILES+= info/dir
.if defined(INFO_DIR) && empty(INFO_DIR:Minfo)
_PLIST_IGNORE_FILES+= ${INFO_DIR}/dir
.endif
_PLIST_IGNORE_FILES+= *[~\#] *.OLD *.orig *,v # scratch config files
_PLIST_IGNORE_FILES+= ${PLIST_IGNORE_FILES}
.endif
BUILD_DEFS+= _PLIST_IGNORE_FILES
Merge a little visual-only cleanup from the tv-derecurse branch: Collect many modifiable hard-to-find variables scattered throughout bsd.pkg.mk into a common section near the top of the file. (Not necessarily "complete", but helps to reduce HEAD-branch divergence.)
2005-01-24 20:57:42 +01:00
.if !empty(USE_GNU_TOOLS:Mmake)
Create a symlink from the "make" requested by the package into ${TOOLS_DIR}/bin/make. If a package wants GNU make, then it should set: USE_GNU_TOOLS+= make in the package Makefile. Allow for USE_GMAKE to continue to work until we get a chance to sweep through pkgsrc and remove the usage of this now-redundant variable. This change allows us to avoid patching makefiles that use a bare "make" command to invoke sub-make processes. Idea suggested by salo@netbsd.org in pkg/22509.
2003-08-26 22:15:23 +02:00
_USE_GMAKE= yes
.endif
* Push the imake- and xmkmf-handling into the new tools framework. * Get rid of an explicit check for ${_IMAKE_MAKE} == ${GMAKE} in bsd.pkg.mk to check for whether we need to depend on gmake or not. Instead, we now note in Linux.mk that packages that need imake will also need to use gmake by setting _IMAKE_TOOLS+=gmake. * Push the definition of MAKE_PROGRAM from bsd.pkg.mk into make.mk where it's closer to related code.
2005-05-02 23:10:02 +02:00
.if empty(_USE_NEW_TOOLS:M[yY][eE][sS])
Create a symlink from the "make" requested by the package into ${TOOLS_DIR}/bin/make. If a package wants GNU make, then it should set: USE_GNU_TOOLS+= make in the package Makefile. Allow for USE_GMAKE to continue to work until we get a chance to sweep through pkgsrc and remove the usage of this now-redundant variable. This change allows us to avoid patching makefiles that use a bare "make" command to invoke sub-make processes. Idea suggested by salo@netbsd.org in pkg/22509.
2003-08-26 22:15:23 +02:00
.if defined(_USE_GMAKE)
Let make do the work: + set MAKE_PROGRAM variable in bsd.port.mk to either ${GMAKE} or ${MAKE}, and use it in do-build and do-install targets + if X11 package, and man pages are installed, add install.man to ${INSTALL_TARGET}, rather than using a separate command to do this. Also added CATn= blah.0, similar to MANn= blah.1, which installs pre-formatted man pages, like the ones imake produces.
1997-11-05 14:36:31 +01:00
MAKE_PROGRAM= ${GMAKE}
Create a symlink from the "make" requested by the package into ${TOOLS_DIR}/bin/make. If a package wants GNU make, then it should set: USE_GNU_TOOLS+= make in the package Makefile. Allow for USE_GMAKE to continue to work until we get a chance to sweep through pkgsrc and remove the usage of this now-redundant variable. This change allows us to avoid patching makefiles that use a bare "make" command to invoke sub-make processes. Idea suggested by salo@netbsd.org in pkg/22509.
2003-08-26 22:15:23 +02:00
.elif defined(USE_IMAKE)
Use GNU make on Linux when USE_IMAKE=yes, because Makefile generated by imake on Linux requires GNU make.
2004-12-30 00:19:42 +01:00
. if ${_IMAKE_MAKE} == ${GMAKE}
USE_GNU_TOOLS+= make
. endif
Add a definition of _IMAKE_MAKE to the individual defs.${OPSYS}.mk files, defined to be the make program which is invoked by imake, and use it accordingly in bsd.pkg.mk
2001-11-21 13:53:04 +01:00
MAKE_PROGRAM= ${_IMAKE_MAKE}
Create a symlink from the "make" requested by the package into ${TOOLS_DIR}/bin/make. If a package wants GNU make, then it should set: USE_GNU_TOOLS+= make in the package Makefile. Allow for USE_GMAKE to continue to work until we get a chance to sweep through pkgsrc and remove the usage of this now-redundant variable. This change allows us to avoid patching makefiles that use a bare "make" command to invoke sub-make processes. Idea suggested by salo@netbsd.org in pkg/22509.
2003-08-26 22:15:23 +02:00
.else
Fix thinko in last commit.
2001-06-07 17:40:15 +02:00
MAKE_PROGRAM= ${MAKE}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Several GNU configure scripts (old ones, typically) can't handle it when "${MAKE}" contains forward slashes. Just pass the bare name of ${MAKE} to the configure phase as we ensure that "make" is always calling the correct one symlinked into the .tools directory.
2004-02-12 10:59:48 +01:00
CONFIGURE_ENV+= MAKE="${MAKE_PROGRAM:T}"
* Push the imake- and xmkmf-handling into the new tools framework. * Get rid of an explicit check for ${_IMAKE_MAKE} == ${GMAKE} in bsd.pkg.mk to check for whether we need to depend on gmake or not. Instead, we now note in Linux.mk that packages that need imake will also need to use gmake by setting _IMAKE_TOOLS+=gmake. * Push the definition of MAKE_PROGRAM from bsd.pkg.mk into make.mk where it's closer to related code.
2005-05-02 23:10:02 +02:00
.endif
Implement some parts of proposals posted to tech-pkg: - Define OS_VERSION in bsd.prefs.mk (...if not already defined). Previously was only set in a conditional block in bsd.pkg.mk. - Define new values, MACHINE_PLATFORM and MACHINE_GNU_PLATFORM, which default to ${OPSYS}-${OS_VERSION}-${MACHINE_ARCH} and ${LOWER_OPSYS}-${LOWER_VENDOR}-${LOWER_ARCH} respectively. (Probably should nuke LOWER_ARCH and rename LOWER_VENDOR to something like MACHINE_GNU_VENDOR.) - Use ${MACHINE_GNU_PLATFORM} as the argument to `--host' in GNU_CONFIGURE packages. - Allow for ONLY_FOR_PLATFORM and NOT_FOR_PLATFORM, lists of patterns in csh(1) glob style which are matched against MACHINE_PLATFORM above and take the place of {ONLY,NOT}_FOR_{ARCHS,OPSYS}. (For now, until all pkgs are converted, provide backwards compatibility.) *Both* may be defined, in which case NOT_FOR_PLATFORM takes precedence. - Also do some significant aesthetic cleanup, based on the idea `define platform-specific overrides first, then define a generic case with the ?= operator'. Lots more of this to do yet.
1999-05-24 20:42:00 +02:00
Add a new variable "UNLIMIT_RESOURCES" which can be used to raise process limits for building a package.
2002-01-04 13:00:56 +01:00
# Automatically increase process limit where necessary for building.
_ULIMIT_CMD=
.if defined(UNLIMIT_RESOURCES)
Allow more flexibility in the handling of UNLIMIT_RESOURCES. Each word of UNLIMIT_RESOURCES is supposed to be a knob on ULIMIT_CMD_<word> variable which value if defined is added to _ULIMIT_CMD. The ULIMIT_CMD_* variables are set per $OPSYS in defs.*.mk and are overridable by the user. Solaris' default value of ULIMIT_CMD_memorysize adjusted as suggested in PR pkg/18087 by Jonathan Perkin <sketch@rd.bbc.co.uk>.
2002-08-28 13:06:17 +02:00
. for __tmp__ in ${UNLIMIT_RESOURCES}
. if defined(ULIMIT_CMD_${__tmp__})
_ULIMIT_CMD+= ${ULIMIT_CMD_${__tmp__}} ;
. endif
. endfor
Add a new variable "UNLIMIT_RESOURCES" which can be used to raise process limits for building a package.
2002-01-04 13:00:56 +01:00
.endif
If GNU_CONFIGURE is defined, then pass LIBS to the GNU configure script. There are 40 packages in pkgsrc that do this.
2001-06-12 22:30:09 +02:00
# If GNU_CONFIGURE is defined, then pass LIBS to the GNU configure script.
add CONFIG_SHELL to CONFIGURE_ENV. This avoids picking up bash as the shell which complains about the buildlink vaiable names containing '.' in them.
2002-04-07 19:09:57 +02:00
# also pass in a CONFIG_SHELL to avoid picking up bash
If GNU_CONFIGURE is defined, then pass LIBS to the GNU configure script. There are 40 packages in pkgsrc that do this.
2001-06-12 22:30:09 +02:00
.if defined(GNU_CONFIGURE)
add CONFIG_SHELL to CONFIGURE_ENV. This avoids picking up bash as the shell which complains about the buildlink vaiable names containing '.' in them.
2002-04-07 19:09:57 +02:00
CONFIG_SHELL?= ${SH}
CONFIGURE_ENV+= CONFIG_SHELL=${CONFIG_SHELL}
Work around brain damaged autoconf by stripping leading and trailing whitespace from *FLAGS and LIBS before doing :Q quoting. Should fix PR pkg/29945.
2005-04-21 17:53:53 +02:00
CONFIGURE_ENV+= LIBS=${LIBS:C/^ *//:C/ *$//:Q}
Rather than work around an automake braindeadness, parameterize install-sh so that Interix can set the default mode to 0775. Then add "install_sh" to CONFIGURE_ENV so it uses ${INSTALL}, not the package-supplied install script (as is done for autoconf $INSTALL).
2004-04-23 04:03:18 +02:00
CONFIGURE_ENV+= install_sh=${INSTALL:Q}
Relax need to put GNU_CONFIGURE before inclusion of bsd.prefs.mk: Convert _OPSYS_MAX_CMDLEN to a plain _OPSYS_MAX_CMDLEN_CMD variable, which is not evaluated by a shell until CONFIGURE_ENV is expanded (and only then if USE_LIBTOOL+GNU_CONFIGURE are both set).
2005-01-25 19:15:58 +01:00
. if defined(USE_LIBTOOL) && defined(_OPSYS_MAX_CMDLEN_CMD)
CONFIGURE_ENV+= lt_cv_sys_max_cmd_len=${_OPSYS_MAX_CMDLEN_CMD:sh}
Revert previous (re-add common CPP_PRECOMP_FLAGS). For consistency, also move usage of _USER_DEPENDS and _OPSYS_MAX_CMDLEN to bsd.pkg.mk.
2004-11-16 19:04:00 +01:00
. endif
If GNU_CONFIGURE is defined, then pass LIBS to the GNU configure script. There are 40 packages in pkgsrc that do this.
2001-06-12 22:30:09 +02:00
.endif
- The comment above the _COOKIE settings says "don't change these!!", so use = to set them, not ?=. - _DEPENDS_USE doesn't actually do any macro expansion, so save some processing time by executing it exactly once (remove the .USE operator). - Merge the IGNORE targets. Add an IGNORE target for depends and check-depends (see bottom). - Don't check for installation of a package in ${DEPENDS} when doing a "make depends". Instead, let the return code of make fall through (using && instead of ; in the last clause of each depend recursion block). - Provide a variable IGNORE_FAIL: when set, causes the ignore message to be printed (regardless of IGNORE_SILENT or ECHO_MSG setting) and the build to fail with ${FALSE}. Used with the next addition: - Add a new target, "check-depends", which is called as part of "real-fetch". This validates whether the dependencies of a pkg to be built are installed, and if not, whether those dependencies will prospectively build (will not set IGNORE directly or indirectly through the various disable options). Recurses to "depends" with a DEPENDS_TARGET of "check-depends". ...And now, it is no longer necessary to mark dependent pkgs as unusable when their dependencies are unusable thanks to an incorrect LICENSE, _FOR_PLATFORM setting, and so forth. The check-depends step will validate everything before fetching a single file.
1999-05-25 01:04:52 +02:00
EXTRACT_COOKIE= ${WRKDIR}/.extract_done
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
WRAPPER_COOKIE= ${WRKDIR}/.wrapper_done
- The comment above the _COOKIE settings says "don't change these!!", so use = to set them, not ?=. - _DEPENDS_USE doesn't actually do any macro expansion, so save some processing time by executing it exactly once (remove the .USE operator). - Merge the IGNORE targets. Add an IGNORE target for depends and check-depends (see bottom). - Don't check for installation of a package in ${DEPENDS} when doing a "make depends". Instead, let the return code of make fall through (using && instead of ; in the last clause of each depend recursion block). - Provide a variable IGNORE_FAIL: when set, causes the ignore message to be printed (regardless of IGNORE_SILENT or ECHO_MSG setting) and the build to fail with ${FALSE}. Used with the next addition: - Add a new target, "check-depends", which is called as part of "real-fetch". This validates whether the dependencies of a pkg to be built are installed, and if not, whether those dependencies will prospectively build (will not set IGNORE directly or indirectly through the various disable options). Recurses to "depends" with a DEPENDS_TARGET of "check-depends". ...And now, it is no longer necessary to mark dependent pkgs as unusable when their dependencies are unusable thanks to an incorrect LICENSE, _FOR_PLATFORM setting, and so forth. The check-depends step will validate everything before fetching a single file.
1999-05-25 01:04:52 +02:00
CONFIGURE_COOKIE= ${WRKDIR}/.configure_done
INSTALL_COOKIE= ${WRKDIR}/.install_done
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
TEST_COOKIE= ${WRKDIR}/.test_done
- The comment above the _COOKIE settings says "don't change these!!", so use = to set them, not ?=. - _DEPENDS_USE doesn't actually do any macro expansion, so save some processing time by executing it exactly once (remove the .USE operator). - Merge the IGNORE targets. Add an IGNORE target for depends and check-depends (see bottom). - Don't check for installation of a package in ${DEPENDS} when doing a "make depends". Instead, let the return code of make fall through (using && instead of ; in the last clause of each depend recursion block). - Provide a variable IGNORE_FAIL: when set, causes the ignore message to be printed (regardless of IGNORE_SILENT or ECHO_MSG setting) and the build to fail with ${FALSE}. Used with the next addition: - Add a new target, "check-depends", which is called as part of "real-fetch". This validates whether the dependencies of a pkg to be built are installed, and if not, whether those dependencies will prospectively build (will not set IGNORE directly or indirectly through the various disable options). Recurses to "depends" with a DEPENDS_TARGET of "check-depends". ...And now, it is no longer necessary to mark dependent pkgs as unusable when their dependencies are unusable thanks to an incorrect LICENSE, _FOR_PLATFORM setting, and so forth. The check-depends step will validate everything before fetching a single file.
1999-05-25 01:04:52 +02:00
BUILD_COOKIE= ${WRKDIR}/.build_done
PATCH_COOKIE= ${WRKDIR}/.patch_done
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
TOOLS_COOKIE= ${WRKDIR}/.tools_done
- The comment above the _COOKIE settings says "don't change these!!", so use = to set them, not ?=. - _DEPENDS_USE doesn't actually do any macro expansion, so save some processing time by executing it exactly once (remove the .USE operator). - Merge the IGNORE targets. Add an IGNORE target for depends and check-depends (see bottom). - Don't check for installation of a package in ${DEPENDS} when doing a "make depends". Instead, let the return code of make fall through (using && instead of ; in the last clause of each depend recursion block). - Provide a variable IGNORE_FAIL: when set, causes the ignore message to be printed (regardless of IGNORE_SILENT or ECHO_MSG setting) and the build to fail with ${FALSE}. Used with the next addition: - Add a new target, "check-depends", which is called as part of "real-fetch". This validates whether the dependencies of a pkg to be built are installed, and if not, whether those dependencies will prospectively build (will not set IGNORE directly or indirectly through the various disable options). Recurses to "depends" with a DEPENDS_TARGET of "check-depends". ...And now, it is no longer necessary to mark dependent pkgs as unusable when their dependencies are unusable thanks to an incorrect LICENSE, _FOR_PLATFORM setting, and so forth. The check-depends step will validate everything before fetching a single file.
1999-05-25 01:04:52 +02:00
PACKAGE_COOKIE= ${WRKDIR}/.package_done
Don't create the ${INTERACTIVE_COOKIE}s in the ${WRKDIR} - this will fail if the interactive stage is "fetch", since there will be no ${WRKDIR} at that time. Instead, since ${INTERACTIVE_COOKIE}s are only used in bulk builds, create them in the package directory, along with the related .broken.`hostname` file.
2002-08-07 17:12:45 +02:00
INTERACTIVE_COOKIE= .interactive_stage
Introduce NULL_COOKIE and use it instead of /dev/null for touch (see PR 15597). While here remove some old comments.
2002-05-14 14:47:48 +02:00
NULL_COOKIE= ${WRKDIR}/.null
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Miscellaneous overridable commands:
Add code to include the size of a package into both the installed version and binary packages. The size can be queried then via pkg_info: xfeyrer @ noon% pkg_info -s xv Information for xv-3.10apl1: Size of this package in bytes: 4670692 xfeyrer @ noon% pkg_info -S xv Information for xv-3.10apl1: Size in bytes including required pkgs: 14610165 While doing work on the size code, support for the @src directive was removed (formerly enabled with the pkg_create -s switch, but unused in our pkg system). The new pkg_info -s and -S switches were tested on installed, local (file) and remote (ftp) packages. In bsd.pkg.mk, take special care for pkg_* versions that do not have the pkg_create -s and -S switches and do not record size information there.
1999-11-29 20:48:15 +01:00
SHCOMMENT?= ${ECHO_MSG} >/dev/null '***'
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
Introduce LIBABISUFFIX for platforms which require ABI-dependant library paths.
2004-11-12 21:25:41 +01:00
LIBABISUFFIX?=
Visual cleanup from tv-derecurse: Consolidate many MAKE_ENVs and SCRIPTS_ENVs into a common block. (CONFIGURE_ENVs to be done later.) Introduce new variable ALL_ENV which is automatically included into all of MAKE_ENV, SCRIPTS_ENV, and CONFIGURE_ENV; this allows much cleaner addition of the common CC/CXX/CFLAGS/etc. variables needed by all of these.
2005-02-11 17:36:49 +01:00
CONFIGURE_ENV+= M4="${M4}" YACC="${YACC}"
Don't use "MOTIFBASE" if "USE_MOTIF" was not defined. Patch supplied by John Darrow in PR pkg/9914.
2000-04-25 22:19:56 +02:00
Implement some parts of proposals posted to tech-pkg: - Define OS_VERSION in bsd.prefs.mk (...if not already defined). Previously was only set in a conditional block in bsd.pkg.mk. - Define new values, MACHINE_PLATFORM and MACHINE_GNU_PLATFORM, which default to ${OPSYS}-${OS_VERSION}-${MACHINE_ARCH} and ${LOWER_OPSYS}-${LOWER_VENDOR}-${LOWER_ARCH} respectively. (Probably should nuke LOWER_ARCH and rename LOWER_VENDOR to something like MACHINE_GNU_VENDOR.) - Use ${MACHINE_GNU_PLATFORM} as the argument to `--host' in GNU_CONFIGURE packages. - Allow for ONLY_FOR_PLATFORM and NOT_FOR_PLATFORM, lists of patterns in csh(1) glob style which are matched against MACHINE_PLATFORM above and take the place of {ONLY,NOT}_FOR_{ARCHS,OPSYS}. (For now, until all pkgs are converted, provide backwards compatibility.) *Both* may be defined, in which case NOT_FOR_PLATFORM takes precedence. - Also do some significant aesthetic cleanup, based on the idea `define platform-specific overrides first, then define a generic case with the ?= operator'. Lots more of this to do yet.
1999-05-24 20:42:00 +02:00
TOUCH_FLAGS?= -f
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fix comment about needing patch(1).
2004-03-31 12:00:43 +02:00
# determine if we need a working patch(1).
Follow Grant's advice: use _OPSYS_GPATCH_REQD to influence _NEED_PATCH and let tools.mk take care of the rest.
2004-05-19 05:26:09 +02:00
.if defined(_OPSYS_GPATCH_REQD) && !empty(_OPSYS_GPATCH_REQD:M[yY][eE][sS])
_NEED_PATCH= YES
.else
add a BUILD_DEPENDS on devel/patch if no suitable patch(1) is available. fixes my PR pkg/18380, with hints from jschauma.
2003-10-11 07:09:39 +02:00
_NEED_PATCH!= if [ -d ${PATCHDIR} ]; then \
if [ "`${ECHO} ${PATCHDIR}/patch-*`" != "${PATCHDIR}/patch-*" ]; then \
${ECHO} YES; \
else \
${ECHO} NO; \
fi \
else \
${ECHO} NO; \
fi
Follow Grant's advice: use _OPSYS_GPATCH_REQD to influence _NEED_PATCH and let tools.mk take care of the rest.
2004-05-19 05:26:09 +02:00
.endif
add a BUILD_DEPENDS on devel/patch if no suitable patch(1) is available. fixes my PR pkg/18380, with hints from jschauma.
2003-10-11 07:09:39 +02:00
.if defined(PATCHFILES)
_NEED_PATCH= YES
.endif
let tools.mk decide whether GNU patch is needed.
2004-01-29 11:22:01 +01:00
.if ${_NEED_PATCH} == "YES"
add GNU gnu patch(1), to override patch(1) on some platforms.
2004-01-29 10:38:10 +01:00
USE_GNU_TOOLS+= patch
add a BUILD_DEPENDS on devel/patch if no suitable patch(1) is available. fixes my PR pkg/18380, with hints from jschauma.
2003-10-11 07:09:39 +02:00
.endif
Add automatic manual page handling. This is backwards-compatible with the MANx and CATx definitions in package Makefiles - if neither MANx or CATx is specified, the PLIST will be searched for the package's manual pages. If automatic manual page handling is in operation, and PKG_VERBOSE is set, echo the manual page names.
1998-06-18 13:45:33 +02:00
.if defined(PATCH_DEBUG) || defined(PKG_VERBOSE)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
PATCH_DEBUG_TMP= yes
.else
PATCH_DEBUG_TMP= no
Add some new variables to control how "distribution" patches are extracted and applied to the package sources. The existing options are: PATCH_DIST_STRIP - option that sets the pathname strip count. PATCH_DIST_ARGS - the list of arguments to pass to patch(1). The new options are: PATCH_DIST_CAT is the command that outputs the patch to stdout. For each of these variables, there is a patch-specific variant that may be set, i.e. PATCH_DIST_STRIP.<patch>, PATCH_DIST_ARGS.<patch>, PATCH_DIST_CAT.<patch>.
2004-08-07 17:58:59 +02:00
.endif
PATCH_STRIP?= -p0
.if ${PATCH_DEBUG_TMP} == "yes"
PATCH_ARGS?= -d ${WRKSRC} -E ${PATCH_STRIP}
.else
Implement some parts of proposals posted to tech-pkg: - Define OS_VERSION in bsd.prefs.mk (...if not already defined). Previously was only set in a conditional block in bsd.pkg.mk. - Define new values, MACHINE_PLATFORM and MACHINE_GNU_PLATFORM, which default to ${OPSYS}-${OS_VERSION}-${MACHINE_ARCH} and ${LOWER_OPSYS}-${LOWER_VENDOR}-${LOWER_ARCH} respectively. (Probably should nuke LOWER_ARCH and rename LOWER_VENDOR to something like MACHINE_GNU_VENDOR.) - Use ${MACHINE_GNU_PLATFORM} as the argument to `--host' in GNU_CONFIGURE packages. - Allow for ONLY_FOR_PLATFORM and NOT_FOR_PLATFORM, lists of patterns in csh(1) glob style which are matched against MACHINE_PLATFORM above and take the place of {ONLY,NOT}_FOR_{ARCHS,OPSYS}. (For now, until all pkgs are converted, provide backwards compatibility.) *Both* may be defined, in which case NOT_FOR_PLATFORM takes precedence. - Also do some significant aesthetic cleanup, based on the idea `define platform-specific overrides first, then define a generic case with the ?= operator'. Lots more of this to do yet.
1999-05-24 20:42:00 +02:00
PATCH_ARGS?= -d ${WRKSRC} --forward --quiet -E ${PATCH_STRIP}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
.if defined(BATCH)
PATCH_ARGS+= --batch
.endif
Add some new variables to control how "distribution" patches are extracted and applied to the package sources. The existing options are: PATCH_DIST_STRIP - option that sets the pathname strip count. PATCH_DIST_ARGS - the list of arguments to pass to patch(1). The new options are: PATCH_DIST_CAT is the command that outputs the patch to stdout. For each of these variables, there is a patch-specific variant that may be set, i.e. PATCH_DIST_STRIP.<patch>, PATCH_DIST_ARGS.<patch>, PATCH_DIST_CAT.<patch>.
2004-08-07 17:58:59 +02:00
.if defined(_PATCH_CAN_BACKUP) && (${_PATCH_CAN_BACKUP} == "yes")
make _PATCH_BACKUP_ARG be more useful to callers by moving "-V simple" definition to be part of it instead of appending to PATCH_ARGS. otherwise, $VERSION_CONTROL or $PATCH_VERSION_CONTROL would still override -b/-z.
2002-03-14 03:28:22 +01:00
PATCH_ARGS+= ${_PATCH_BACKUP_ARG} .orig
Add initial support for IRIX. Again, anybody's feedback (success or failure) will be appreciated.
2002-12-02 18:07:27 +01:00
.endif
Add some new variables to control how "distribution" patches are extracted and applied to the package sources. The existing options are: PATCH_DIST_STRIP - option that sets the pathname strip count. PATCH_DIST_ARGS - the list of arguments to pass to patch(1). The new options are: PATCH_DIST_CAT is the command that outputs the patch to stdout. For each of these variables, there is a patch-specific variant that may be set, i.e. PATCH_DIST_STRIP.<patch>, PATCH_DIST_ARGS.<patch>, PATCH_DIST_CAT.<patch>.
2004-08-07 17:58:59 +02:00
PATCH_FUZZ_FACTOR?= -F0 # Default to zero fuzz
# The following variables control how "distribution" patches are extracted
# and applied to the package sources.
#
# PATCH_DIST_STRIP is a patch option that sets the pathname strip count.
# PATCH_DIST_ARGS is the list of arguments to pass to the patch command.
# PATCH_DIST_CAT is the command that outputs the patch to stdout.
#
# For each of these variables, there is a patch-specific variant that
# may be set, i.e. PATCH_DIST_STRIP.<patch>, PATCH_DIST_ARGS.<patch>,
# PATCH_DIST_CAT.<patch>.
#
PATCH_DIST_STRIP?= -p0
.for i in ${PATCHFILES}
PATCH_DIST_STRIP.${i:S/=/--/}?= ${PATCH_DIST_STRIP}
. if defined(PATCH_DIST_ARGS)
PATCH_DIST_ARGS.${i:S/=/--/}?= ${PATCH_DIST_ARGS}
. elif ${PATCH_DEBUG_TMP} == "yes"
PATCH_DIST_ARGS.${i:S/=/--/}?= -d ${WRKSRC} -E ${PATCH_DIST_STRIP.${i:S/=/--/}}
. else
PATCH_DIST_ARGS.${i:S/=/--/}?= -d ${WRKSRC} --forward --quiet -E ${PATCH_DIST_STRIP.${i:S/=/--/}}
. endif
.endfor
.if defined(BATCH)
PATCH_DIST_ARGS+= --batch
. for i in ${PATCHFILES}
PATCH_DIST_ARGS.${i:S/=/--/}+= --batch
. endfor
.endif
.if defined(_PATCH_CAN_BACKUP) && (${_PATCH_CAN_BACKUP} == "yes")
PATCH_DIST_ARGS+= ${_PATCH_BACKUP_ARG} .orig_dist
. for i in ${PATCHFILES}
PATCH_DIST_ARGS.${i:S/=/--/}+= ${_PATCH_BACKUP_ARG} .orig_dist
. endfor
.endif
PATCH_DIST_CAT?= { case $$patchfile in \
*.Z|*.gz) ${GZCAT} $$patchfile ;; \
*.bz2) ${BZCAT} $$patchfile ;; \
*) ${CAT} $$patchfile ;; \
esac; }
.for i in ${PATCHFILES}
PATCH_DIST_CAT.${i:S/=/--/}?= { patchfile=${i}; ${PATCH_DIST_CAT}; }
.endfor
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Add a knob PKGSRC_SHOW_PATCH_ERRORMSG to control whether the newly verbose message displayed on patch failure is in fact displayed on patch failure. It defaults to "no" if PKG_DEVELOPER is defined, "yes" otherwise. While here, rename PKG_PATCH_FAIL to PKGSRC_PATCH_FAIL.
2004-08-25 06:09:10 +02:00
.if empty(PKGSRC_SHOW_PATCH_ERRORMSG:M[yY][eE][sS])
PKGSRC_PATCH_FAIL= exit 1
.else
PKGSRC_PATCH_FAIL= \
If patching fails and either of PKG_OPTIONS and LOCALPATCHES are defined, note thatthe patch failure may be due to incompatible build options and/or local patches, and suggest building with fewer.
2004-08-16 03:56:03 +02:00
if [ -n "${PKG_OPTIONS}" ] || [ -n "${_LOCALPATCHFILES}" ]; then \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
${ECHO} "=========================================================================="; \
If patching fails and either of PKG_OPTIONS and LOCALPATCHES are defined, note thatthe patch failure may be due to incompatible build options and/or local patches, and suggest building with fewer.
2004-08-16 03:56:03 +02:00
${ECHO}; \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
${ECHO} "Some of the selected build options and/or local patches may be incompatible."; \
If patching fails and either of PKG_OPTIONS and LOCALPATCHES are defined, note thatthe patch failure may be due to incompatible build options and/or local patches, and suggest building with fewer.
2004-08-16 03:56:03 +02:00
${ECHO} "Please try building with fewer options or patches."; \
${ECHO}; \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
${ECHO} "=========================================================================="; \
If patching fails and either of PKG_OPTIONS and LOCALPATCHES are defined, note thatthe patch failure may be due to incompatible build options and/or local patches, and suggest building with fewer.
2004-08-16 03:56:03 +02:00
fi; exit 1
Add a knob PKGSRC_SHOW_PATCH_ERRORMSG to control whether the newly verbose message displayed on patch failure is in fact displayed on patch failure. It defaults to "no" if PKG_DEVELOPER is defined, "yes" otherwise. While here, rename PKG_PATCH_FAIL to PKGSRC_PATCH_FAIL.
2004-08-25 06:09:10 +02:00
.endif
If patching fails and either of PKG_OPTIONS and LOCALPATCHES are defined, note thatthe patch failure may be due to incompatible build options and/or local patches, and suggest building with fewer.
2004-08-16 03:56:03 +02:00
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
# We need bzip2 for PATCHFILES with .bz2 suffix.
Support bzip2 compressed patch files.
2000-05-17 05:32:43 +02:00
.if defined(PATCHFILES)
Replace the only occurrence of the :E modifier in bsd.pkg.mk with a more familiar construction. BZCAT wasn't getting defined for a bzip2-compressed patch in my tree; now it is.
2004-09-01 02:10:58 +02:00
. if !empty(PATCHFILES:M*.bz2) && ${EXTRACT_SUFX} != ".tar.bz2"
Teach the new tools framework about USE_TOOLS+=bzcat.
2005-05-02 04:50:33 +02:00
. if !empty(_USE_NEW_TOOLS:M[yY][eE][sS])
USE_TOOLS+= bzcat
. elif exists(/usr/bin/bzcat)
Support bzip2 compressed patch files.
2000-05-17 05:32:43 +02:00
BZCAT= /usr/bin/bzcat
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Support bzip2 compressed patch files.
2000-05-17 05:32:43 +02:00
BZCAT= ${LOCALBASE}/bin/bzcat
change BUILD_DEPENDS+= bzip2-0.9.0b:../../archivers/bzip2 to >=0.9.0b; otherwise, extracts fail when bzip2-1.0.1 is installed.
2001-08-16 03:49:09 +02:00
BUILD_DEPENDS+= bzip2>=0.9.0b:../../archivers/bzip2
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif # !exists bzcat
. endif
Support bzip2 compressed patch files.
2000-05-17 05:32:43 +02:00
.endif # defined(PATCHFILES)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Figure out where the local mtree file is
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
MTREE_FILE?= ${PKGSRCDIR}/mk/${OPSYS}.pkg.dist
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
MTREE_ARGS?= -U -f ${MTREE_FILE} -d -e -p
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
# Debugging levels for this file, dependent on PKG_DEBUG_LEVEL definition
# 0 == normal, default, quiet operation
# 1 == all shell commands echoed before invocation
# 2 == shell "set -x" operation
PKG_DEBUG_LEVEL?= 0
_PKG_SILENT= @
When PKG_DEBUG_LEVEL=2, set_PKG_DEBUG_SCRIPT is set to "${SH} -x". Use _PKG_DEBUG_SCRIPT when running the INSTALL/DEINSTALL scripts so that they may be easier to debug.
2002-09-26 23:04:11 +02:00
_PKG_DEBUG= # empty
_PKG_DEBUG_SCRIPT= # empty
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
.if ${PKG_DEBUG_LEVEL} > 0
When PKG_DEBUG_LEVEL=2, set_PKG_DEBUG_SCRIPT is set to "${SH} -x". Use _PKG_DEBUG_SCRIPT when running the INSTALL/DEINSTALL scripts so that they may be easier to debug.
2002-09-26 23:04:11 +02:00
_PKG_SILENT= # empty
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
.endif
.if ${PKG_DEBUG_LEVEL} > 1
_PKG_DEBUG= set -x;
When PKG_DEBUG_LEVEL=2, set_PKG_DEBUG_SCRIPT is set to "${SH} -x". Use _PKG_DEBUG_SCRIPT when running the INSTALL/DEINSTALL scripts so that they may be easier to debug.
2002-09-26 23:04:11 +02:00
_PKG_DEBUG_SCRIPT= ${SH} -x
- Add OpenBSD and FreeBSD Ids in the header - Changes from FreeBSD's V1.164: * Warn if mtree-file not found * On 'make reinstall' gets DEPENDS_TARGET set to 'reinstall' instead of 'install' - Changes from OpenBSD's V1.14: * Documentation and implementation for ONLY_FOR_ARCHS * Documentation and implementation for WRKOBJDIR * Documentation on HAVE_MOTIF and MOTIF_STATIC: set in /etc/mk.conf, not /etc/make.conf * Set NOMANCOMPRESS=no to still compress man-pages (should probably be set to 'yes') * DEF_UMASK=022 * add ${MACHINE} suffix to WRKDIR if OBJMACHINE is set * .include <bsd.own.mk>
1997-09-28 02:44:15 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# A few aliases for *-install targets
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_PROGRAM?= \
Replace STRIPFLAG with _STRIPFLAG_INSTALL to fix pkg/15467.
2002-05-29 16:52:14 +02:00
${INSTALL} ${COPY} ${_STRIPFLAG_INSTALL} -o ${BINOWN} -g ${BINGRP} -m ${BINMODE}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_GAME?= \
${INSTALL} ${COPY} ${_STRIPFLAG_INSTALL} -o ${GAMEOWN} -g ${GAMEGRP} -m ${GAMEMODE}
INSTALL_SCRIPT?= \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${INSTALL} ${COPY} -o ${BINOWN} -g ${BINGRP} -m ${BINMODE}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_LIB?= \
Add a new install macro INSTALL_LIB for use when installing libraries (mainly intended for shlib use, but for homeful use on all libraries so that currently static libs can be "provisioned" for future shlib use).
2004-04-28 00:35:47 +02:00
${INSTALL} ${COPY} -o ${BINOWN} -g ${BINGRP} -m ${BINMODE}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_DATA?= \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${INSTALL} ${COPY} -o ${SHAREOWN} -g ${SHAREGRP} -m ${SHAREMODE}
Introduce GAMEDATAMODE and INSTALL_GAME_DATA, which will take care of files installed for SETGIDGAME packages. This will help address PR pkg/25313 and PR pkg/26236.
2004-08-04 04:56:01 +02:00
INSTALL_GAME_DATA?= \
${INSTALL} ${COPY} -o ${GAMEOWN} -g ${GAMEGRP} -m ${GAMEDATAMODE}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_MAN?= \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${INSTALL} ${COPY} -o ${MANOWN} -g ${MANGRP} -m ${MANMODE}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_PROGRAM_DIR?= \
Create directories with mode 0755, since that is what the mtree files expect. Noted by Toru TAKAMIZU in PR 22007, patch from seb@netbsd.
2003-07-12 12:21:33 +02:00
${INSTALL} -d -o ${BINOWN} -g ${BINGRP} -m ${PKGDIRMODE}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_SCRIPT_DIR?= \
Add INSTALL_*_DIR macros.
1998-03-07 22:19:00 +01:00
${INSTALL_PROGRAM_DIR}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_LIB_DIR?= \
Add a new install macro INSTALL_LIB for use when installing libraries (mainly intended for shlib use, but for homeful use on all libraries so that currently static libs can be "provisioned" for future shlib use).
2004-04-28 00:35:47 +02:00
${INSTALL_PROGRAM_DIR}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_DATA_DIR?= \
Create directories with mode 0755, since that is what the mtree files expect. Noted by Toru TAKAMIZU in PR 22007, patch from seb@netbsd.
2003-07-12 12:21:33 +02:00
${INSTALL} -d -o ${SHAREOWN} -g ${SHAREGRP} -m ${PKGDIRMODE}
Introduce new variables: SETGIDGAME: specifies whether or not certain games are installed setgid GAMEGRP: the group owning games if SETGIDGAME is set GAMEOWN: the user owning games if SETFIDGAME is set GAMEMODE: the mode to install games as Use these to provide INSTALL_GAME (for example for PR pkg/25313), and set them to meaningful values under NetBSD.
2004-05-12 03:03:58 +02:00
INSTALL_MAN_DIR?= \
Create directories with mode 0755, since that is what the mtree files expect. Noted by Toru TAKAMIZU in PR 22007, patch from seb@netbsd.
2003-07-12 12:21:33 +02:00
${INSTALL} -d -o ${MANOWN} -g ${MANGRP} -m ${PKGDIRMODE}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
INSTALL_MACROS= BSD_INSTALL_PROGRAM="${INSTALL_PROGRAM}" \
BSD_INSTALL_SCRIPT="${INSTALL_SCRIPT}" \
Add a new install macro INSTALL_LIB for use when installing libraries (mainly intended for shlib use, but for homeful use on all libraries so that currently static libs can be "provisioned" for future shlib use).
2004-04-28 00:35:47 +02:00
BSD_INSTALL_LIB="${INSTALL_LIB}" \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
BSD_INSTALL_DATA="${INSTALL_DATA}" \
BSD_INSTALL_MAN="${INSTALL_MAN}" \
* pass BSD_INSTALL to programs, just like the other BSD_INSTALL_* macros, but _only_ the install program's path. * check for a few variables before using them.
2002-12-11 22:01:05 +01:00
BSD_INSTALL="${INSTALL}" \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
BSD_INSTALL_PROGRAM_DIR="${INSTALL_PROGRAM_DIR}" \
BSD_INSTALL_SCRIPT_DIR="${INSTALL_SCRIPT_DIR}" \
Add a new install macro INSTALL_LIB for use when installing libraries (mainly intended for shlib use, but for homeful use on all libraries so that currently static libs can be "provisioned" for future shlib use).
2004-04-28 00:35:47 +02:00
BSD_INSTALL_LIB_DIR="${INSTALL_LIB_DIR}" \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
BSD_INSTALL_DATA_DIR="${INSTALL_DATA_DIR}" \
INSTALL_GAME should be exported a as BSD_INSTALL_GAME
2004-05-12 03:37:42 +02:00
BSD_INSTALL_MAN_DIR="${INSTALL_MAN_DIR}" \
Introduce GAMEDATAMODE and INSTALL_GAME_DATA, which will take care of files installed for SETGIDGAME packages. This will help address PR pkg/25313 and PR pkg/26236.
2004-08-04 04:56:01 +02:00
BSD_INSTALL_GAME="${INSTALL_GAME}" \
BSD_INSTALL_GAME_DATA="${INSTALL_GAME_DATA}"
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
MAKE_ENV+= ${INSTALL_MACROS}
SCRIPTS_ENV+= ${INSTALL_MACROS}
# The user can override the NO_PACKAGE by specifying this from
# the make command line
.if defined(FORCE_PACKAGE)
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. undef NO_PACKAGE
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
s/PLIST_SUB/PLIST_SUBST/
1999-09-03 04:43:51 +02:00
# Set PLIST_SUBST to substitute "${variable}" to "value" in PLIST
Changes for building dynamic PLIST for perl packages: * Add a PERL5 variable that is always defined, set to the path of a pkgsrc-installed perl5 binary (${LOCALBASE}/bin/perl). * Define three new variables PERL5_SITELIB, PERL5_SITEARCH, and PERL5_ARCHLIB that define special locations used by perl. In particular, it seems packlists are always found under PERL5_SITEARCH. * Substitute the above three variables when generating PLISTs, but stripping away the initial ${LOCALBASE}/ part. This is for perl packages which don't generate packlists because they don't use the perl Makefile generator during the build process. * Append to the PLIST any files listed in the packlists named by PERL5_PACKLIST. Also append statements to try removing the directories installed into by the packlist-ed files.
2000-08-27 04:15:49 +02:00
PLIST_SUBST+= OPSYS=${OPSYS} \
OS_VERSION=${OS_VERSION} \
MACHINE_ARCH=${MACHINE_ARCH} \
MACHINE_GNU_ARCH=${MACHINE_GNU_ARCH} \
MACHINE_GNU_PLATFORM=${MACHINE_GNU_PLATFORM} \
Add missing LN to PLIST_SUBST
2002-07-14 01:27:54 +02:00
LN=${LN:Q} \
Changes for building dynamic PLIST for perl packages: * Add a PERL5 variable that is always defined, set to the path of a pkgsrc-installed perl5 binary (${LOCALBASE}/bin/perl). * Define three new variables PERL5_SITELIB, PERL5_SITEARCH, and PERL5_ARCHLIB that define special locations used by perl. In particular, it seems packlists are always found under PERL5_SITEARCH. * Substitute the above three variables when generating PLISTs, but stripping away the initial ${LOCALBASE}/ part. This is for perl packages which don't generate packlists because they don't use the perl Makefile generator during the build process. * Append to the PLIST any files listed in the packlists named by PERL5_PACKLIST. Also append statements to try removing the directories installed into by the packlist-ed files.
2000-08-27 04:15:49 +02:00
LOWER_VENDOR=${LOWER_VENDOR} \
LOWER_OPSYS=${LOWER_OPSYS} \
Add LOWER_OS_VERSION to match LOWER_OPSYS, LOWER_VENDOR, and LOWER_ARCH
2001-12-12 13:31:14 +01:00
LOWER_OS_VERSION=${LOWER_OS_VERSION} \
Add ${PKGBASE} to the things to substitute for in PLISTs.
2001-11-14 13:38:51 +01:00
PKGBASE=${PKGBASE} \
Invent a new variable "PKGNAME_NOREV" which holds the package name without the package revision added. Replace "PKGNAME" in package list with this value because the installed software doesn't know anything about package system revision. This fixes package list problems in e.g. the "xemacs" package.
2002-03-15 11:05:56 +01:00
PKGNAME=${PKGNAME_NOREV} \
Fix cut and paste error in last commit.
2001-09-24 13:02:49 +02:00
PKGLOCALEDIR=${PKGLOCALEDIR} \
PKGVERSION=${PKGVERSION:C/nb[0-9]*$//} \
Don't hardcode /usr/X11R6 when making directories or running X based programs such as mkfontdir use ${X11BASE} instead. Also pick up a couple of /bin/chmod -> ${CHMOD}s
2002-02-05 23:03:54 +01:00
LOCALBASE=${LOCALBASE} \
Make the value of VIEWBASE available in more settings.
2003-11-23 08:31:18 +01:00
VIEWBASE=${VIEWBASE} \
Don't hardcode /usr/X11R6 when making directories or running X based programs such as mkfontdir use ${X11BASE} instead. Also pick up a couple of /bin/chmod -> ${CHMOD}s
2002-02-05 23:03:54 +01:00
X11BASE=${X11BASE} \
X11PREFIX=${X11PREFIX} \
So many of our PLIST use mkdir, rmdir, and true, and we have full paths to these utilities in ${MKDIR}, ${RMDIR}, and ${TRUE}, so substitute for them.
2001-11-01 21:37:17 +01:00
SVR4_PKGNAME=${SVR4_PKGNAME} \
Also substitute for CHMOD, CHOWN, CHGRP in PLISTs.
2001-11-03 05:58:15 +01:00
CHGRP=${CHGRP:Q} \
CHMOD=${CHMOD:Q} \
CHOWN=${CHOWN:Q} \
So many of our PLIST use mkdir, rmdir, and true, and we have full paths to these utilities in ${MKDIR}, ${RMDIR}, and ${TRUE}, so substitute for them.
2001-11-01 21:37:17 +01:00
MKDIR=${MKDIR:Q} \
RMDIR=${RMDIR:Q} \
mkdir -> ${MKDIR} rmdir -> ${RMDIR} rm -> ${RM} (${RM} added to PLIST_SUBST) chmod -> ${CHMOD} chown -> ${CHOWN}
2002-02-15 11:12:28 +01:00
RM=${RM:Q} \
Add QMAILDIR variable that defaults to /var/qmail, may be set in /etc/mk.conf Also substitute QMAILDIR in PLIST because more packages than just mail/qmail need it.
2002-01-21 18:31:02 +01:00
TRUE=${TRUE:Q} \
QMAILDIR=${QMAILDIR}
Add support for PLIST_SUB variable. It is a list of "VAR=VALUE" pairs, and will substitute ${VAR} with the value of VALUE in the PLIST. This implementation was taken from FreeBSD. Default PLIST_SUB: OPSYS=${OPSYS} OS_VERSION=${OS_VERSION} MACHINE_ARCH=${MACHINE_ARCH} MACHINE_GNU_ARCH=${MACHINE_GNU_ARCH} MACHINE_GNU_PLATFORM=${MACHINE_GNU_PLATFORM} LOWER_VENDOR=${LOWER_VENDOR} LOWER_OPSYS=${LOWER_OPSYS} PKGNAME=${PKGNAME}
1999-08-31 11:20:21 +02:00
Add alternatives.mk, a file used to manipulate the alternatives system. A package can optionally provide an ALTERNATIVES file which contains pairs of wrapper/alternative, one per line. The file is then used by the install and deinstall scripts to register the alternatives and to create the appropriate wrappers. Make bsd.pkg.mk include this new file. This happens unconditionally to keep all the alternatives logic in an independent file. Otherwise, some of the stuff could have to be moved to bsd.pkg.mk, breaking consistency.
2005-01-25 14:13:35 +01:00
# Handle alternatives
#
.include "../../mk/alternatives.mk"
Introduce a new variable "CONF_DEPENDS" for pkgviews: a package's config directory matches the config directory for the dependency listed in CONF_DEPENDS. Use symlinks to physically point the package's config directory to the dependency's config directory, and handle all of this in the INSTALL/DEINSTALL scripts. Also make the INSTALL/DEINSTALL scripts a bit smarter about not copying files and not removing files if the source and destination file locations point to the same thing.
2003-09-17 04:38:22 +02:00
# CONF_DEPENDS notes a dependency where the config directory for the
# package matches the dependency's config directory. CONF_DEPENDS is
# only meaningful if ${PKG_INSTALLATION_TYPE} == "pkgviews".
#
CONF_DEPENDS?= # empty
.if !empty(CONF_DEPENDS)
USE_PKGINSTALL= yes
.endif
Move the inclusion of bsd.pkg.install.mk higher so that binary packages are generated correctly (with INSTALL and DEINSTALL scripts). Fix suggested by Stoned Elipot <seb@netbsd.org>.
2003-01-29 20:34:14 +01:00
.if defined(USE_PKGINSTALL) && !empty(USE_PKGINSTALL:M[yY][eE][sS])
. include "../../mk/bsd.pkg.install.mk"
.endif
Make any INSTALL or DEINSTALL scripts specifiable in the package Makefile (for read-only pkgsrc use). This is for the use of packages which build INSTALL and DEINSTALL scripts dynamically, such as pkgsrc/games/crossfire.
1998-09-17 16:03:48 +02:00
# Set INSTALL_FILE to be the name of any INSTALL file
.if !defined(INSTALL_FILE) && exists(${PKGDIR}/INSTALL)
INSTALL_FILE= ${PKGDIR}/INSTALL
.endif
# Set DEINSTALL_FILE to be the name of any DEINSTALL file
.if !defined(DEINSTALL_FILE) && exists(${PKGDIR}/DEINSTALL)
DEINSTALL_FILE= ${PKGDIR}/DEINSTALL
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Allow MESSAGE_SRC to be a space-separated list of files, like PLIST_SRC, that are concatenated together to form the whole MESSAGE file.
2001-11-14 22:04:43 +01:00
# If MESSAGE hasn't been defined, then set MESSAGE_SRC to be a space-separated
# list of files to be concatenated together to generate the MESSAGE file.
#
Add OS and arch specific MESSAGE file handling. This removes the need for packages to override MESSAGE_SRC to provide platform specific install messages.
2002-12-05 06:41:39 +01:00
.if !defined(MESSAGE_SRC) && !defined(MESSAGE)
Partially revert to behavior of revision 1.1100 wrt MESSAGE_SRC: only if if MESSAGE_SRC and MESSAGE are _not_ already set fall back to setting MESSAGE_SRC to ${PKGDIR}/MESSAGE if it exists. This fix builds of mail/sendmail{,811}/.
2002-12-07 17:03:12 +01:00
. if exists(${PKGDIR}/MESSAGE)
MESSAGE_SRC= ${PKGDIR}/MESSAGE
. else
. if exists(${PKGDIR}/MESSAGE.common)
Add OS and arch specific MESSAGE file handling. This removes the need for packages to override MESSAGE_SRC to provide platform specific install messages.
2002-12-05 06:41:39 +01:00
MESSAGE_SRC= ${PKGDIR}/MESSAGE.common
Partially revert to behavior of revision 1.1100 wrt MESSAGE_SRC: only if if MESSAGE_SRC and MESSAGE are _not_ already set fall back to setting MESSAGE_SRC to ${PKGDIR}/MESSAGE if it exists. This fix builds of mail/sendmail{,811}/.
2002-12-07 17:03:12 +01:00
. endif
. if exists(${PKGDIR}/MESSAGE.${OPSYS})
Add OS and arch specific MESSAGE file handling. This removes the need for packages to override MESSAGE_SRC to provide platform specific install messages.
2002-12-05 06:41:39 +01:00
MESSAGE_SRC+= ${PKGDIR}/MESSAGE.${OPSYS}
Partially revert to behavior of revision 1.1100 wrt MESSAGE_SRC: only if if MESSAGE_SRC and MESSAGE are _not_ already set fall back to setting MESSAGE_SRC to ${PKGDIR}/MESSAGE if it exists. This fix builds of mail/sendmail{,811}/.
2002-12-07 17:03:12 +01:00
. endif
Add architecture specific MESSAGE file handling.
2003-04-03 18:55:09 +02:00
. if exists(${PKGDIR}/MESSAGE.${MACHINE_ARCH:C/i[3-6]86/i386/g})
MESSAGE_SRC+= ${PKGDIR}/MESSAGE.${MACHINE_ARCH:C/i[3-6]86/i386/g}
. endif
use the form MESSAGE.${OPSYS}-${MACHINE_PLATFORM} for automatic MESSAGE file handling. pointed out by Masao Uebayashi.
2002-12-10 14:08:34 +01:00
. if exists(${PKGDIR}/MESSAGE.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g})
MESSAGE_SRC+= ${PKGDIR}/MESSAGE.${OPSYS}-${MACHINE_ARCH:C/i[3-6]86/i386/g}
Partially revert to behavior of revision 1.1100 wrt MESSAGE_SRC: only if if MESSAGE_SRC and MESSAGE are _not_ already set fall back to setting MESSAGE_SRC to ${PKGDIR}/MESSAGE if it exists. This fix builds of mail/sendmail{,811}/.
2002-12-07 17:03:12 +01:00
. endif
Add OS and arch specific MESSAGE file handling. This removes the need for packages to override MESSAGE_SRC to provide platform specific install messages.
2002-12-05 06:41:39 +01:00
. endif
Add automatic ${VARIABLE} handling for MESSAGE files. Convert most MESSAGE files to new syntax (${VARIABLE} gets replaced, not @VARIABLE@, nor @@VARIABLE@@). By default, substitutions are done for LOCALBASE, PKGNAME, PREFIX, X11BASE, X11PREFIX; additional patterns can be added via MESSAGE_SUBST. Clean up some packages while I'm there; add RCS tags to most MESSAGEs. Remove some uninteresting MESSAGEs.
2001-01-29 12:34:21 +01:00
.endif
.if defined(MESSAGE_SRC)
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
MESSAGE= ${PKG_DB_TMPDIR}/+DISPLAY
Add automatic ${VARIABLE} handling for MESSAGE files. Convert most MESSAGE files to new syntax (${VARIABLE} gets replaced, not @VARIABLE@, nor @@VARIABLE@@). By default, substitutions are done for LOCALBASE, PKGNAME, PREFIX, X11BASE, X11PREFIX; additional patterns can be added via MESSAGE_SUBST. Clean up some packages while I'm there; add RCS tags to most MESSAGEs. Remove some uninteresting MESSAGEs.
2001-01-29 12:34:21 +01:00
# Set MESSAGE_SUBST to substitute "${variable}" to "value" in MESSAGE
MESSAGE_SUBST+= PKGNAME=${PKGNAME} \
PREFIX=${PREFIX} \
LOCALBASE=${LOCALBASE} \
Make the value of VIEWBASE available in more settings.
2003-11-23 08:31:18 +01:00
VIEWBASE=${VIEWBASE} \
Add automatic ${VARIABLE} handling for MESSAGE files. Convert most MESSAGE files to new syntax (${VARIABLE} gets replaced, not @VARIABLE@, nor @@VARIABLE@@). By default, substitutions are done for LOCALBASE, PKGNAME, PREFIX, X11BASE, X11PREFIX; additional patterns can be added via MESSAGE_SUBST. Clean up some packages while I'm there; add RCS tags to most MESSAGEs. Remove some uninteresting MESSAGEs.
2001-01-29 12:34:21 +01:00
X11PREFIX=${X11PREFIX} \
Also substitute for ${PKG_SYSCONFDIR} in MESSAGE files.
2002-01-10 03:51:30 +01:00
X11BASE=${X11BASE} \
Add ${QMAILDIR} to MESSAGE_SUBST, as it is going to be used in all qmail-related packages.
2002-01-21 17:31:42 +01:00
PKG_SYSCONFDIR=${PKG_SYSCONFDIR} \
Recognize ROOT_USER and ROOT_GROUP in MESSAGE files.
2002-10-08 13:01:50 +02:00
ROOT_GROUP=${ROOT_GROUP} \
ROOT_USER=${ROOT_USER} \
Add ${QMAILDIR} to MESSAGE_SUBST, as it is going to be used in all qmail-related packages.
2002-01-21 17:31:42 +01:00
QMAILDIR=${QMAILDIR}
Allow MESSAGE_SRC to be a space-separated list of files, like PLIST_SRC, that are concatenated together to form the whole MESSAGE file.
2001-11-14 22:04:43 +01:00
MESSAGE_SUBST_SED= ${MESSAGE_SUBST:S/=/}!/:S/$/!g/:S/^/ -e s!\\\${/}
Make the REQ and MESSAGE files specifiable from package Makefiles, as REQ_FILE and MESSAGE_FILE respectively. Modify xpkgwedge to create its MESSAGE_FILE in the WRKDIR.
1998-09-17 17:47:08 +02:00
.endif
Due to the special handling of pkgtools/digest in pkgsrc (it must be installed before the normal dependency checks), on a system without any packages installed, we can run into the problem where digest won't install because it has already been installed as part of installing a build dependency. This can happen if, e.g., PKGSRC_COMPILER is set to "ccache gcc", and you try to build devel/gmake. Change the way that the uptodate-digest target installs pkgtools/digest by ensuring that pkgtools/digest is deinstalled just before installing it.
2004-03-22 05:07:15 +01:00
# If pkgsrc is supposed to ensure that tests are run before installation
# of the package, then the build targets should be "build test", otherwise
# just "build" suffices. _PKGSRC_BUILD_TARGETS is used in the "all",
# "install", and "uptodate-digest" targets.
#
.if !empty(PKGSRC_RUN_TEST:M[yY][eE][sS])
_PKGSRC_BUILD_TARGETS= build test
.else
_PKGSRC_BUILD_TARGETS= build
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: uptodate-digest
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
uptodate-digest:
Test was reversed.
2004-03-13 21:58:06 +01:00
.if !empty(USE_DIGEST:M[yY][eE][sS])
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Change uptodate-digest check from excluding digest on 'digest-* | pkgchk-*)' to excluding when ! -f ${DISTINFO_FILE}. Now packages without distinfo do not need to build digest before installing (such as cpuflags)
2003-06-17 16:33:17 +02:00
if [ -f ${DISTINFO_FILE} -a \( ! -f ${DIGEST} -o ${DIGEST_VERSION} -lt ${DIGEST_REQD} \) ]; then \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
{ cd ${PKGSRCDIR}/pkgtools/digest; \
Change uptodate-digest check from excluding digest on 'digest-* | pkgchk-*)' to excluding when ! -f ${DISTINFO_FILE}. Now packages without distinfo do not need to build digest before installing (such as cpuflags)
2003-06-17 16:33:17 +02:00
${MAKE} clean; \
if [ -f ${DIGEST} ]; then \
${MAKE} ${MAKEFLAGS} deinstall; \
fi; \
Due to the special handling of pkgtools/digest in pkgsrc (it must be installed before the normal dependency checks), on a system without any packages installed, we can run into the problem where digest won't install because it has already been installed as part of installing a build dependency. This can happen if, e.g., PKGSRC_COMPILER is set to "ccache gcc", and you try to build devel/gmake. Change the way that the uptodate-digest target installs pkgtools/digest by ensuring that pkgtools/digest is deinstalled just before installing it.
2004-03-22 05:07:15 +01:00
${MAKE} ${MAKEFLAGS} ${_PKGSRC_BUILD_TARGETS}; \
if [ -f ${DIGEST} ]; then \
${MAKE} ${MAKEFLAGS} deinstall; \
fi; \
Change uptodate-digest check from excluding digest on 'digest-* | pkgchk-*)' to excluding when ! -f ${DISTINFO_FILE}. Now packages without distinfo do not need to build digest before installing (such as cpuflags)
2003-06-17 16:33:17 +02:00
${MAKE} ${MAKEFLAGS} ${DEPENDS_TARGET}; \
${MAKE} ${MAKEFLAGS} clean; } \
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
fi
If USE_DIGEST is "no", then skip the check for an up-to-date digest. XXX These changes should be removed once we have a working XXX pkgsrc/bootstrap.
2004-03-13 21:54:16 +01:00
.else
@${DO_NADA}
.endif
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
.ifndef PKG_ARGS_COMMON
Revert previous. There are extra quotes in COMMENT files created by make package on NetBSD. Need to find another way to fix this.
2002-10-28 19:53:37 +01:00
PKG_ARGS_COMMON= -v -c -${COMMENT:Q}" " -d ${DESCR} -f ${PLIST}
Change COMMENT handling: COMMENTs are now a variable in the Makefile instead of a pkg/COMMENT file. The COMMENT var should be in the maintainer block after the homepage. Modify bsd.pkg.mk, pkglint, url2pkg, and port2pkg (last one untested) for the new behaviour. Document new state in Packages.txt. This should save lots of inodes, and lots of time when untarring/updating. Idea by Alistair Crooks. For the time being, accept pkg/COMMENT instead of a COMMENT var to avoid a flag day.
2001-02-16 14:06:17 +01:00
PKG_ARGS_COMMON+= -l -b ${BUILD_VERSION_FILE} -B ${BUILD_INFO_FILE}
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
PKG_ARGS_COMMON+= -s ${SIZE_PKG_FILE} -S ${SIZE_ALL_FILE}
Undo implicit recursive package registration. The buildlink[23] system is already taking care of this where it's truly needed. OK'd by agc.
2004-04-02 21:09:18 +02:00
PKG_ARGS_COMMON+= -P "`${MAKE} ${MAKEFLAGS} run-depends-list | ${SORT} -u`"
Don't write conflicts into the binary packages unless they're "overwrite" packages. All pkgviews packages are safe to install into the same tree, and we rely on more dynamic file collision detection through linkfarm(1) rather than just using CONFLICTS.
2003-11-25 13:01:31 +01:00
. if defined(CONFLICTS) && (${PKG_INSTALLATION_TYPE} == "overwrite")
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
PKG_ARGS_COMMON+= -C "${CONFLICTS}"
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. ifdef INSTALL_FILE
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
PKG_ARGS_COMMON+= -i ${INSTALL_FILE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. ifdef DEINSTALL_FILE
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
PKG_ARGS_COMMON+= -k ${DEINSTALL_FILE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. ifdef MESSAGE
Add automatic ${VARIABLE} handling for MESSAGE files. Convert most MESSAGE files to new syntax (${VARIABLE} gets replaced, not @VARIABLE@, nor @@VARIABLE@@). By default, substitutions are done for LOCALBASE, PKGNAME, PREFIX, X11BASE, X11PREFIX; additional patterns can be added via MESSAGE_SUBST. Clean up some packages while I'm there; add RCS tags to most MESSAGEs. Remove some uninteresting MESSAGEs.
2001-01-29 12:34:21 +01:00
PKG_ARGS_COMMON+= -D ${MESSAGE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. ifndef NO_MTREE
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
PKG_ARGS_COMMON+= -m ${MTREE_FILE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Add the PKG_PRESERVE capability to bsd.pkg.mk. A package which has PKG_PRESERVE definied in its Makefile will not be able to be deleted, and the capability is carried into binary packages.
2003-01-10 12:57:09 +01:00
. ifdef PKG_PRESERVE
PKG_ARGS_COMMON+= -n ${PRESERVE_FILE}
. endif
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
PKG_ARGS_INSTALL= -p ${PREFIX} ${PKG_ARGS_COMMON}
Don't use regexp substitution where plain substitution is intended, it might yield some surprising results. Pointed out by Marc Espie <espie@schutzenberger.liafa.jussieu.fr>
2001-03-30 00:08:21 +02:00
PKG_ARGS_BINPKG= -p ${PREFIX:S/^${DESTDIR}//} -L ${PREFIX} ${PKG_ARGS_COMMON}
As per discussion on tech-pkg, correct behavior of `make package' in the case where DESTDIR is set. Previously, if DESTDIR was set when a pkg was installed, $DESTDIR/var/db/pkg/+CONTENT would get `@cwd $DESTDIR/$PREFIX', as would the +CONTENT archived in the binary package. The first is correct, the second is not. Now, @cwd in the +CONTENT recorded in the binary package gets $PREFIX (no $DESTDIR). This allows binary packages built into a DESTDIR to be installed on a real system. The @cwd in the installed PKG_DBDIR ($DESTDIR/var/db/pkg) remains the same. In addition, the full path ($DESTDIR/$PREFIX) is recorded in @src in the binary package's +CONTENT, for reference purposes. This patch is the same as that posted to tech-pkg, except that variable names have been clarified as suggested by hubertf.
2001-01-29 02:56:51 +01:00
.endif # !PKG_ARGS_COMMON
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if ${PKG_INSTALLATION_TYPE} == "pkgviews"
PKG_ARGS_INSTALL+= -U # don't update the pkgdb.byfile.db
PKG_ARGS_BINPKG+= -E # create an empty views file in the binpkg
.endif
Previous commit to bsd.pkg.mk caught some "SMART_MESSAGES" work - document them now, in mk.conf.example, and add a small comment explaining what they do. When SMART_MESSAGES is defined, when compiling packages, the make(1) target is displayed, and also the current stack of packages being built.
2000-03-13 15:24:27 +01:00
# Define SMART_MESSAGES in /etc/mk.conf for messages giving the tree
Typo in comment for SMART_MESSAGES.
2001-01-31 00:18:36 +01:00
# of dependencies for building, and the current target.
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
.ifdef SMART_MESSAGES
_PKGSRC_IN?= ===> ${.TARGET} [${PKGNAME}${_PKGSRC_DEPS}] ===
.else
_PKGSRC_IN?= ===
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Used to print all the '===>' style prompts - override this to turn them off.
ECHO_MSG?= ${ECHO}
Introduce TRUE?=/usr/bin/true, and use it.
1998-02-02 09:10:41 +01:00
# How to do nothing. Override if you, for some strange reason, would rather
# do something.
DO_NADA?= ${TRUE}
Implement a new PASSIVE_FETCH definition, which can be set in the environment, or /etc/mk.conf, and is used to specify -p to ftp(1) when fetching distfiles from behind a filtering firewall. Document PASSIVE_FETCH in mk.conf.example. Suggested by Bill Coldwell.
1998-05-29 11:21:43 +02:00
# If this host is behind a filtering firewall, use passive ftp(1)
.if defined(PASSIVE_FETCH)
FETCH_BEFORE_ARGS += -p
.endif
Remove MASTER_SITE_* definitions from bsd.pkg.mk and put them into the new bsd.sites.mk file. Ok'ed by Hubert Feyrer and Jan Schaumann in tech-pkg@.
2004-06-04 15:45:26 +02:00
# Include popular master sites
.include "../../mk/bsd.sites.mk"
Automatically set MASTER_SITE_LOCAL from MASTER_SITE_BACKUP, and make sure MASTER_SITE_BACKUP doesn't end in //. For this, any setting of MASTER_SITE_LOCAL in /etc/mk.conf must no longer include ${DIST_SUBDIR} - this will be added automatically, if needed.
1999-11-20 22:15:49 +01:00
.if defined(DIST_SUBDIR)
Don't apply ${DIST_SUBDIR} twice. Fixes PR 10646 by Kawaguti Ginga <k.ginga@nbwww.isc.org ???>.
2000-07-26 14:36:49 +02:00
_MASTER_SITE_BACKUP:= ${MASTER_SITE_BACKUP:=${DIST_SUBDIR}/}
. if defined(MASTER_SITE_OVERRIDE)
_MASTER_SITE_OVERRIDE:= ${MASTER_SITE_OVERRIDE:=${DIST_SUBDIR}/}
. endif # MASTER_SITE_OVERRIDE
.else # !DIST_SUBDIR
_MASTER_SITE_BACKUP:= ${MASTER_SITE_BACKUP}
. if defined(MASTER_SITE_OVERRIDE)
_MASTER_SITE_OVERRIDE:= ${MASTER_SITE_OVERRIDE}
. endif # MASTER_SITE_OVERRIDE
Only add DIST_SUBDIR to MASTER_SITE_OVERRIDE if MASTER_SITE_OVERRIDE is actually set. Prevents "Variable MASTER_SITE_OVERRIDE is recursive" message. Found by Ezra Story <ezy@panix.com>
2000-02-25 06:46:27 +01:00
.endif # DIST_SUBDIR
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Where to put distfiles that don't have any other master site
MASTER_SITE_LOCAL?= \
Add "ftp.fi.netbsd.org" to "MASTER_SITE_LOCAL" so that e.g. building the "openssl" package works while "ftp.netbsd.org" is down.
2002-08-11 11:59:43 +02:00
${MASTER_SITE_BACKUP:=LOCAL_PORTS/} \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
ALLFILES?= ${DISTFILES} ${PATCHFILES}
Speed up the evaluation of ${CKSUMFILES} vs. ${IGNOREFILES} a little (and make it cleaner) by replacing the shell expansion with a bmake .for loop.
1999-05-21 21:29:02 +02:00
CKSUMFILES?= ${ALLFILES}
.for __tmp__ in ${IGNOREFILES}
CKSUMFILES:= ${CKSUMFILES:N${__tmp__}}
.endfor
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Fix MD5 checksum test in "do-fetch:" target to work with "DIST_SUBDIR". Patch supplied by Jim Bernard in PR pkg/5962.
1998-08-15 00:10:53 +02:00
# List of all files, with ${DIST_SUBDIR} in front. Used for fetch and checksum.
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if defined(DIST_SUBDIR)
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if ${CKSUMFILES} != ""
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
_CKSUMFILES?= ${CKSUMFILES:S/^/${DIST_SUBDIR}\//}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
When DIST_SUBDIR is set, make sure various download variables don't get changed when they're defined but empty. Although it might sound weird to have DISTFILES empty and DIST_SUBDIR set, it makes sense for wrapper packages such as graphics/Mesa. Fixes PR 23658 from Aaron J. Grier.
2003-12-06 19:47:45 +01:00
. if !empty(DISTFILES)
Fix MD5 checksum test in "do-fetch:" target to work with "DIST_SUBDIR". Patch supplied by Jim Bernard in PR pkg/5962.
1998-08-15 00:10:53 +02:00
_DISTFILES?= ${DISTFILES:S/^/${DIST_SUBDIR}\//}
When DIST_SUBDIR is set, make sure various download variables don't get changed when they're defined but empty. Although it might sound weird to have DISTFILES empty and DIST_SUBDIR set, it makes sense for wrapper packages such as graphics/Mesa. Fixes PR 23658 from Aaron J. Grier.
2003-12-06 19:47:45 +01:00
. else
_DISTFILES?= # empty
. endif
. if defined(IGNOREFILES) && !empty(IGNOREFILES)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
_IGNOREFILES?= ${IGNOREFILES:S/^/${DIST_SUBDIR}\//}
When DIST_SUBDIR is set, make sure various download variables don't get changed when they're defined but empty. Although it might sound weird to have DISTFILES empty and DIST_SUBDIR set, it makes sense for wrapper packages such as graphics/Mesa. Fixes PR 23658 from Aaron J. Grier.
2003-12-06 19:47:45 +01:00
. endif
. if defined(PATCHFILES) && !empty(PATCHFILES)
Fix MD5 checksum test in "do-fetch:" target to work with "DIST_SUBDIR". Patch supplied by Jim Bernard in PR pkg/5962.
1998-08-15 00:10:53 +02:00
_PATCHFILES?= ${PATCHFILES:S/^/${DIST_SUBDIR}\//}
When DIST_SUBDIR is set, make sure various download variables don't get changed when they're defined but empty. Although it might sound weird to have DISTFILES empty and DIST_SUBDIR set, it makes sense for wrapper packages such as graphics/Mesa. Fixes PR 23658 from Aaron J. Grier.
2003-12-06 19:47:45 +01:00
. else
_PATCHFILES?= # empty
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.else
_CKSUMFILES?= ${CKSUMFILES}
Fix MD5 checksum test in "do-fetch:" target to work with "DIST_SUBDIR". Patch supplied by Jim Bernard in PR pkg/5962.
1998-08-15 00:10:53 +02:00
_DISTFILES?= ${DISTFILES}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
_IGNOREFILES?= ${IGNOREFILES}
Fix MD5 checksum test in "do-fetch:" target to work with "DIST_SUBDIR". Patch supplied by Jim Bernard in PR pkg/5962.
1998-08-15 00:10:53 +02:00
_PATCHFILES?= ${PATCHFILES}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Re-work the fetch target(s), as discussed on "tech-pkg". The primary motivation is to make the order that sites are hit more sane, and especially to prevent hitting sites that aren't ever expected to have the sought after file. Now, ${MASTER_SITE_OVERRIDE} is always hit first. If that fails, then the sites designated for that file, then ${MASTER_SITE_BACKUP}, are hit (by default), but with the order subject to ${MASTER_SORT_REGEX} and ${MASTER_SORT_AWK}. The "designated sites" are usually ${MASTER_SITES} for files in ${DISTFILES}, and ${PATCH_SITES} for files in ${PATCH_FILES}. However, defining a variable `${SITES_foo}' in the package "Makefile" overrides that for file "foo". [The use of ${MASTER_SITES_foo} and ${PATCH_SITES_foo}, which is currently only used by a couple of packages for the same purpose, is deprecated, and will be shortly unsupported.] Also eliminate redundancy in the do-fetch and fetch-list-one-pkg targets, by making them use a single, common macro (escaped with `:Q' in the fetch-list-one-pkg case), so "make fetch-list-one-pkg | sh" now does exactly the same thing as "make do-fetch".
2002-01-06 21:03:22 +01:00
_ALLFILES?= ${_DISTFILES} ${_PATCHFILES}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Add _DISTFILES and _PATCHFILES definitions to the list of definitions which get recorded in the build information (accessible with pkg_info -B).
2005-04-18 12:06:47 +02:00
BUILD_DEFS+= _DISTFILES _PATCHFILES
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if defined(GNU_CONFIGURE)
New variable GNU_CONFIGURE_PREFIX whose value is the argument to the --prefix option passed to a GNU configure script. It defaults to ${PREFIX}. This should help simplify the packages out there that avoid setting GNU_CONFIGURE=yes because they need a custom --prefix option.
2002-06-23 05:15:28 +02:00
#
It's GNU_CONFIGURE_PREFIX, not CONFIGURE_PREFIX.
2004-01-31 08:07:33 +01:00
# GNU_CONFIGURE_PREFIX is the argument to the --prefix option passed to the
New variable GNU_CONFIGURE_PREFIX whose value is the argument to the --prefix option passed to a GNU configure script. It defaults to ${PREFIX}. This should help simplify the packages out there that avoid setting GNU_CONFIGURE=yes because they need a custom --prefix option.
2002-06-23 05:15:28 +02:00
# GNU configure script.
#
GNU_CONFIGURE_PREFIX?= ${PREFIX}
Occasionally we do not want the "--host=${MACHINE_GNU_PLATFORM}" argument added if GNU_CONFIGURE is specified - the arla and forthcoming openafs packages are ones that need to be able to let the GNU configure script guess for itself, rather than hardcoding the value. Add a switch called USE_GNU_CONFIGURE_HOST to determine whether the --host argument is provided to the GNU configure script. This is switched on (it has a "yes" value) by default, so the previous behaviour applies, but allows us to specify this value on a package-by-package basis. Documentation changes for this are also forthcoming.
2005-04-16 11:20:18 +02:00
USE_GNU_CONFIGURE_HOST?= yes
. if !empty(USE_GNU_CONFIGURE_HOST:M[yY][eE][sS])
New variable GNU_CONFIGURE_PREFIX whose value is the argument to the --prefix option passed to a GNU configure script. It defaults to ${PREFIX}. This should help simplify the packages out there that avoid setting GNU_CONFIGURE=yes because they need a custom --prefix option.
2002-06-23 05:15:28 +02:00
CONFIGURE_ARGS+= --host=${MACHINE_GNU_PLATFORM}
Occasionally we do not want the "--host=${MACHINE_GNU_PLATFORM}" argument added if GNU_CONFIGURE is specified - the arla and forthcoming openafs packages are ones that need to be able to let the GNU configure script guess for itself, rather than hardcoding the value. Add a switch called USE_GNU_CONFIGURE_HOST to determine whether the --host argument is provided to the GNU configure script. This is switched on (it has a "yes" value) by default, so the previous behaviour applies, but allows us to specify this value on a package-by-package basis. Documentation changes for this are also forthcoming.
2005-04-16 11:20:18 +02:00
. endif
New variable GNU_CONFIGURE_PREFIX whose value is the argument to the --prefix option passed to a GNU configure script. It defaults to ${PREFIX}. This should help simplify the packages out there that avoid setting GNU_CONFIGURE=yes because they need a custom --prefix option.
2002-06-23 05:15:28 +02:00
CONFIGURE_ARGS+= --prefix=${GNU_CONFIGURE_PREFIX}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
HAS_CONFIGURE= yes
Always add --x-{includes,libraries}=${X11BASE}/{include,lib} to CONFIGURE_ARGS if USE_X11 is defined i.e no need to special case xpkgwedge usage. Please note that previously these were added to CONFIGURE_ARGS as soon as xpkgwedge was used even if the package did not actually defined USE_X11 or something implying it.
2004-02-12 16:34:18 +01:00
. if defined(USE_X11)
x11.buildlink.mk is no more so remove checks for X11_BUILDLINK_MK.
2002-10-23 00:57:56 +02:00
CONFIGURE_ARGS+= --x-includes=${X11BASE}/include
Introduce LIBABISUFFIX for platforms which require ABI-dependant library paths.
2004-11-12 21:25:41 +01:00
CONFIGURE_ARGS+= --x-libraries=${X11BASE}/lib${LIBABISUFFIX}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Rename NO_CONFIGURE_INFODIR to CONFIGURE_HAS_INFODIR (which defaults to "yes"). So any packages that have old GNU configure script that does not support --infodir option (while using INFO_FILES) can set CONFIGURE_HAS_INFODIR=NO. This was suggested by agc@ as a clearer definition.
2004-05-08 18:14:55 +02:00
CONFIGURE_HAS_INFODIR?= yes
. if !empty(INFO_FILES) && !empty(CONFIGURE_HAS_INFODIR:M[yY][eE][sS])
If using GNU_CONFIGURE then set --infodir to ${PREFIX}/${INFO_DIR}. I have used this successfully to install over 300 info files using INFO_DIR=share/info. Also add new variable called NO_CONFIGURE_INFODIR (as suggested by Greg Woods) because some autoconf-style configure scripts don't know --infodir. (The alternative is to use HAS_CONFIGURE and add the needed CONFIGURE_ARGS.) Please note that the info tool may not find custom INFO_DIR unless it is configure to know it. (I will update gtexinfo for this, but if using native info you are on your own if you customize this.)
2004-05-08 04:13:00 +02:00
CONFIGURE_ARGS+= --infodir=${PREFIX}/${INFO_DIR}
. endif
if using GNU configure and CONFIG_{GUESS,SUB}_OVERRIDE is not defined, default to overriding ${WRKSRC}/config.{guess,sub} so packages can recognise platforms such as NetBSD/shark.
2004-02-14 01:40:55 +01:00
#
# By default, override config.guess and config.sub for GNU configure
# packages. pkgsrc's updated versions of these scripts allows GNU
# configure to recognise NetBSD ports such as shark.
#
Make CONFIG_{GUESS,SUB}_OVERRIDE slighly more powerful by allowing them to be lists of shell globs relative to ${WRKSRC}.
2004-02-14 16:22:42 +01:00
CONFIG_GUESS_OVERRIDE?= \
config.guess */config.guess */*/config.guess
CONFIG_SUB_OVERRIDE?= \
config.sub */config.sub */*/config.sub
Add override for uncommon "config.rpath" GNU config snippet. Use platform file to indicate that override logic should be turned on. (AFAICT, only Interix is afflicted at the moment.)
2005-01-14 01:10:01 +01:00
CONFIG_RPATH_OVERRIDE?= # set by platform file as needed
Default to overriding config.status after GNU configure has been run. We don't want it to execute during the build phase as it might overwrite post-configure changes we might have made to the generated files. Packages that don't want this behaviour can simply set: CONFIG_STATUS_OVERRIDE= # empty in the package Makefile.
2004-02-14 14:54:28 +01:00
#
# By default, override config.status for GNU configure packages. We
# never want it to execute after the configure phase has ended as it
# might overwrite any post-configure changes we might have made to the
# generated files.
#
CONFIG_STATUS_OVERRIDE?= \
config.status */config.status */*/config.status
Update buildlink-x11 to 0.5. Changes from previous version are the addition of the xmkmf script from the XFree86-4.0.x distribution, which understands passing -D options through to imake. Update the dependency on buildlink-x11 in bsd.pkg.mk for packages using USE_BUILDLINK_ONLY to >=0.5. Also, if USE_BUILDLINK_X11 is set, then refer to ${BUILDLINK_DIR} for location of X headers and libraries.
2001-07-05 06:47:27 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Move the PKG_SYSCONFDIR logic below the definition of PKGNAME. PKGNAME is use to set PKGBASE, which is used in the name of the package-specific config directory override (PKG_SYSCONFDIR.${PKGBASE}). Prior to this, if PKGNAME wasn't explicitly set in the package, then the override would not work.
2001-11-28 20:55:31 +01:00
#
Move description of variables used for config file handling to Packages.txt. Supplements PR 16971 by Julio Merino <jmmv@hispabsd.org>
2002-05-23 01:15:22 +02:00
# Config file related settings - see Packages.txt
Create a new knob: PKG_SYSCONFVAR PKG_SYSCONFVAR is the special suffix used to distinguish any overriding values for a particular package (see next item). It defaults to ${PKGNAME}, but for a collection of related packages that should all have the same PKG_SYSCONFDIR value, it can be set in each of the package Makefiles to a common value. PKG_SYSCONFDIR.${PKG_SYSCONFVAR} overrides the value of ${PKG_SYSCONFDIR} for packages with the same value for PKG_SYSCONFVAR. Package maintainers may want to set PKG_SYSCONFVAR to a common value for related packages, e.g. all of the amanda packages set PKG_SYSCONFVAR=amanda so that the PKG_SYSCONFDIR for all of them may be tweaked by just setting PKG_SYSCONFDIR.amanda in /etc/mk.conf.
2002-02-07 21:34:21 +01:00
#
PKG_SYSCONFVAR?= ${PKGBASE}
Move the PKG_SYSCONFDIR logic below the definition of PKGNAME. PKGNAME is use to set PKGBASE, which is used in the name of the package-specific config directory override (PKG_SYSCONFDIR.${PKGBASE}). Prior to this, if PKGNAME wasn't explicitly set in the package, then the override would not work.
2001-11-28 20:55:31 +01:00
PKG_SYSCONFSUBDIR?= # empty
Set DFLT_PKG_SYSCONDIR to the location that PKG_SYSCONFDIR would have been set to if it hadn't been overridden by PKG_SYSCONFDIR.<pkg>. This can be used in /etc/mk.conf to re-override certain PKG_SYSCONFDIR.<pkg> in /etc/mk.conf from various package Makefiles, e.g. PKG_SYSCONFDIR.openssl= ${DFLT_PKG_SYSCONFDIR} PKG_SYSCONFDIR.tcp_wrappers= ${DFLT_PKG_SYSCONFDIR}
2004-03-29 08:49:04 +02:00
.if ${PKG_INSTALLATION_TYPE} == "overwrite"
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
PKG_SYSCONFDEPOTBASE= # empty
Package Makefiles should refer to PKG_SYSCONFBASEDIR instead of PKG_SYSCONFBASE when they want PKG_SYSCONFDIR stripped of PKG_SYSCONFSUBDIR. This makes PKG_SYSCONFBASE=/etc work with pkgviews by installing all config files into /etc/packages/<pkg> instead of occasionally putting some directly into /etc. Also only create PKG_SYSCONFDIR if we're actually going to copy config files.
2003-09-05 13:34:25 +02:00
PKG_SYSCONFBASEDIR= ${PKG_SYSCONFBASE}
Set DFLT_PKG_SYSCONDIR to the location that PKG_SYSCONFDIR would have been set to if it hadn't been overridden by PKG_SYSCONFDIR.<pkg>. This can be used in /etc/mk.conf to re-override certain PKG_SYSCONFDIR.<pkg> in /etc/mk.conf from various package Makefiles, e.g. PKG_SYSCONFDIR.openssl= ${DFLT_PKG_SYSCONFDIR} PKG_SYSCONFDIR.tcp_wrappers= ${DFLT_PKG_SYSCONFDIR}
2004-03-29 08:49:04 +02:00
.else
. if !empty(PKG_SYSCONFBASE:M${PREFIX}) || \
!empty(PKG_SYSCONFBASE:M${PREFIX}/*)
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
PKG_SYSCONFDEPOTBASE= # empty
Package Makefiles should refer to PKG_SYSCONFBASEDIR instead of PKG_SYSCONFBASE when they want PKG_SYSCONFDIR stripped of PKG_SYSCONFSUBDIR. This makes PKG_SYSCONFBASE=/etc work with pkgviews by installing all config files into /etc/packages/<pkg> instead of occasionally putting some directly into /etc. Also only create PKG_SYSCONFDIR if we're actually going to copy config files.
2003-09-05 13:34:25 +02:00
PKG_SYSCONFBASEDIR= ${PKG_SYSCONFBASE}
Set DFLT_PKG_SYSCONDIR to the location that PKG_SYSCONFDIR would have been set to if it hadn't been overridden by PKG_SYSCONFDIR.<pkg>. This can be used in /etc/mk.conf to re-override certain PKG_SYSCONFDIR.<pkg> in /etc/mk.conf from various package Makefiles, e.g. PKG_SYSCONFDIR.openssl= ${DFLT_PKG_SYSCONFDIR} PKG_SYSCONFDIR.tcp_wrappers= ${DFLT_PKG_SYSCONFDIR}
2004-03-29 08:49:04 +02:00
. if !empty(CONF_DEPENDS)
_PLIST_IGNORE_FILES+= ${PKG_SYSCONFDIR:S,^${PREFIX}/,,}
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
. endif
Move the PKG_SYSCONFDIR logic below the definition of PKGNAME. PKGNAME is use to set PKGBASE, which is used in the name of the package-specific config directory override (PKG_SYSCONFDIR.${PKGBASE}). Prior to this, if PKGNAME wasn't explicitly set in the package, then the override would not work.
2001-11-28 20:55:31 +01:00
. else
Set DFLT_PKG_SYSCONDIR to the location that PKG_SYSCONFDIR would have been set to if it hadn't been overridden by PKG_SYSCONFDIR.<pkg>. This can be used in /etc/mk.conf to re-override certain PKG_SYSCONFDIR.<pkg> in /etc/mk.conf from various package Makefiles, e.g. PKG_SYSCONFDIR.openssl= ${DFLT_PKG_SYSCONFDIR} PKG_SYSCONFDIR.tcp_wrappers= ${DFLT_PKG_SYSCONFDIR}
2004-03-29 08:49:04 +02:00
PKG_SYSCONFDEPOTBASE= ${PKG_SYSCONFBASE}/${DEPOT_SUBDIR}
PKG_SYSCONFBASEDIR= ${PKG_SYSCONFDEPOTBASE}/${PKGNAME}
Move the PKG_SYSCONFDIR logic below the definition of PKGNAME. PKGNAME is use to set PKGBASE, which is used in the name of the package-specific config directory override (PKG_SYSCONFDIR.${PKGBASE}). Prior to this, if PKGNAME wasn't explicitly set in the package, then the override would not work.
2001-11-28 20:55:31 +01:00
. endif
.endif
Set DFLT_PKG_SYSCONDIR to the location that PKG_SYSCONFDIR would have been set to if it hadn't been overridden by PKG_SYSCONFDIR.<pkg>. This can be used in /etc/mk.conf to re-override certain PKG_SYSCONFDIR.<pkg> in /etc/mk.conf from various package Makefiles, e.g. PKG_SYSCONFDIR.openssl= ${DFLT_PKG_SYSCONFDIR} PKG_SYSCONFDIR.tcp_wrappers= ${DFLT_PKG_SYSCONFDIR}
2004-03-29 08:49:04 +02:00
.if empty(PKG_SYSCONFSUBDIR)
DFLT_PKG_SYSCONFDIR:= ${PKG_SYSCONFBASEDIR}
.else
DFLT_PKG_SYSCONFDIR:= ${PKG_SYSCONFBASEDIR}/${PKG_SYSCONFSUBDIR}
.endif
PKG_SYSCONFDIR= ${DFLT_PKG_SYSCONFDIR}
.if defined(PKG_SYSCONFDIR.${PKG_SYSCONFVAR})
PKG_SYSCONFDIR= ${PKG_SYSCONFDIR.${PKG_SYSCONFVAR}}
PKG_SYSCONFBASEDIR= ${PKG_SYSCONFDIR.${PKG_SYSCONFVAR}}
PKG_SYSCONFDEPOTBASE= # empty
.endif
Move the PKG_SYSCONFDIR logic below the definition of PKGNAME. PKGNAME is use to set PKGBASE, which is used in the name of the package-specific config directory override (PKG_SYSCONFDIR.${PKGBASE}). Prior to this, if PKGNAME wasn't explicitly set in the package, then the override would not work.
2001-11-28 20:55:31 +01:00
Visual cleanup from tv-derecurse: Consolidate many MAKE_ENVs and SCRIPTS_ENVs into a common block. (CONFIGURE_ENVs to be done later.) Introduce new variable ALL_ENV which is automatically included into all of MAKE_ENV, SCRIPTS_ENV, and CONFIGURE_ENV; this allows much cleaner addition of the common CC/CXX/CFLAGS/etc. variables needed by all of these.
2005-02-11 17:36:49 +01:00
ALL_ENV+= PKG_SYSCONFDIR=${PKG_SYSCONFDIR:Q}
Introduce a new variable "CONF_DEPENDS" for pkgviews: a package's config directory matches the config directory for the dependency listed in CONF_DEPENDS. Use symlinks to physically point the package's config directory to the dependency's config directory, and handle all of this in the INSTALL/DEINSTALL scripts. Also make the INSTALL/DEINSTALL scripts a bit smarter about not copying files and not removing files if the source and destination file locations point to the same thing.
2003-09-17 04:38:22 +02:00
BUILD_DEFS+= PKG_SYSCONFBASEDIR PKG_SYSCONFDIR
Move the PKG_SYSCONFDIR logic below the definition of PKGNAME. PKGNAME is use to set PKGBASE, which is used in the name of the package-specific config directory override (PKG_SYSCONFDIR.${PKGBASE}). Prior to this, if PKGNAME wasn't explicitly set in the package, then the override would not work.
2001-11-28 20:55:31 +01:00
if NO_BUILD is defined, default to an empty USE_LANGUAGES.
2004-03-31 12:14:22 +02:00
# If NO_BUILD is defined, default to not needing a compiler.
.if defined(NO_BUILD)
USE_LANGUAGES?= # empty
.endif
Include bsd.compiler.mk inside both bsd.pkg.mk and bsd.prefs.mk and define a stack-like mechanism in bsd.prefs.mk to detect when a we're inside bsd.prefs.mk.
2004-02-05 04:37:47 +01:00
# Get the proper dependencies and set the PATH to use the compiler
# named in PKGSRC_COMPILER.
#
* Move pkgsrc/mk/compiler/bsd.compiler.mk to pkgsrc/mk/compiler.mk. Package Makefiles may now directly include compiler.mk. * Don't include compiler.mk within bsd.prefs.mk any longer. It was only included for the purposes of defining CC_VERSION. Packages that want to test the value of CC_VERSION should now first include "../../mk/compiler.mk". Any GCC_REQD statements in package Makefiles should be set before compiler.mk is included. * Simpllfy pkgsrc/mk/compiler/*.mk files as a result of not needing to be included indirectly by bsd.prefs.mk. We remove the special handling associated with detecting whether the file was included from within bsd.prefs.mk. These files are now much more straightforward to write and understand. * G/C the BSD_PREFS_MK stack mechanism as the only users (compiler/*) no longer need it. * Ensure that directories are prepended to the PATH only from within bsd.pkg.mk.
2004-02-18 14:32:38 +01:00
.include "../../mk/compiler.mk"
Include bsd.compiler.mk inside both bsd.pkg.mk and bsd.prefs.mk and define a stack-like mechanism in bsd.prefs.mk to detect when a we're inside bsd.prefs.mk.
2004-02-05 04:37:47 +01:00
Split replace.mk into two parts, one of which is included by bsd.prefs.mk to provide "TOOL" definitions for tools used by a top-level make process (usually because it uses them in a != variable definition). This allows USE_TOOLS to be defined before bsd.prefs.mk is included by a package Makefile, where USE_TOOLS lists the additional (non-default) tools that are required to build the package. Also, drop the fallback to existing "TOOL" definitions because we now have TOOLS_PLATFORM.* for each platform in pkgsr/mk/tools/tools.*.mk.
2005-04-30 06:35:54 +02:00
# These are all of the tools use by pkgsrc Makefiles. This should
# eventually be split up into lists of tools required by different
# phases of a pkgsrc build.
#
USE_TOOLS+= [ awk basename cat chgrp chmod chown cmp cp cut date \
dirname echo egrep env expr false fgrep file find grep \
We need ldconfig (this should probably be restricted only to packages that install shared libraries).
2005-05-02 07:45:15 +02:00
gtar gunzip gzcat gzip head hostname id ldconfig ln ls \
m4 mkdir mtree mv nice pax pwd rm rmdir sed sh shlock \
sort tail tee test touch tr true tsort wc xargs
Split replace.mk into two parts, one of which is included by bsd.prefs.mk to provide "TOOL" definitions for tools used by a top-level make process (usually because it uses them in a != variable definition). This allows USE_TOOLS to be defined before bsd.prefs.mk is included by a package Makefile, where USE_TOOLS lists the additional (non-default) tools that are required to build the package. Also, drop the fallback to existing "TOOL" definitions because we now have TOOLS_PLATFORM.* for each platform in pkgsr/mk/tools/tools.*.mk.
2005-04-30 06:35:54 +02:00
Prepend USE_NEW_TOOLS with an underscore to make it private. Developers should set _USE_NEW_TOOLS=yes in /etc/mk.conf in order to test the new tools framework.
2005-04-22 04:20:22 +02:00
.if !empty(_USE_NEW_TOOLS:M[yY][eE][sS])
Add a new yes/no variable USE_NEW_TOOLS to ease testing the new tools framework. USE_NEW_TOOLS defaults to "no".
2005-04-15 04:04:57 +02:00
.include "../../mk/tools/bsd.tools.mk"
.else
Move the inclusion of tools.mk ahead of bsd.buildlink3.mk so that tools.mk can include buildlink3.mk files. This change also causes work/.tools/bin to appear after work/.buildlink/bin in the PATH, but we never relied on this behaviour, so this side-effect should have no impact.
2004-04-27 21:23:34 +02:00
.include "../../mk/tools.mk"
Add a new yes/no variable USE_NEW_TOOLS to ease testing the new tools framework. USE_NEW_TOOLS defaults to "no".
2005-04-15 04:04:57 +02:00
.endif
Move the inclusion of tools.mk ahead of bsd.buildlink3.mk so that tools.mk can include buildlink3.mk files. This change also causes work/.tools/bin to appear after work/.buildlink/bin in the PATH, but we never relied on this behaviour, so this side-effect should have no impact.
2004-04-27 21:23:34 +02:00
Always include wrapper/bl3, period, even if NO_BUILD.
2005-04-05 16:00:33 +02:00
.include "../../mk/wrapper/bsd.wrapper.mk"
Merge pkgsrc/mk from the buildlink2 branch back into the main trunk.
2002-08-22 10:21:26 +02:00
Process RECOMMENDED *after* including bsd.buildlink3.mk; else it has no real effect at all.
2004-08-24 02:27:13 +02:00
.if defined(RECOMMENDED)
. if !empty(IGNORE_RECOMMENDED:M[nN][oO])
DEPENDS+= ${RECOMMENDED}
. else
BUILD_DEFS+= IGNORE_RECOMMENDED
. endif
.endif
Move the "dependency reduction" code from buildlink3 to bsd.pkg.mk so that all of pkgsrc can benefit from removing redundant dependencies. The code is encapsulated in a new file reduce-depends.mk which is included by bsd.pkg.mk after all dependencies have been specified.
2004-10-05 17:28:50 +02:00
# Remove some redundant dependencies from the DEPENDS list.
.include "../../mk/reduce-depends.mk"
DEPENDS:= ${REDUCED_DEPENDS}
Merge down deferred PREPEND_PATH handling from the tv-derecurse branch. Also merge in deferred CLASSPATH handling, which can now be moved back to java-vm.mk for cleanliness.
2005-01-24 20:32:33 +01:00
# Find out the PREFIX of dependencies where the PREFIX is needed at build time.
.if defined(EVAL_PREFIX)
. for def in ${EVAL_PREFIX}
. if !defined(${def:C/=.*$//})
${def:C/=.*$//}_DEFAULT?=${LOCALBASE}
fix bug introduced in 1.1569: use ${GREP} instead of 'grep' and not all grep(1)s support grepping for empty string. `${GREP} .' instead.
2005-02-20 07:02:06 +01:00
_${def:C/=.*$//}_CMD= ${PKG_INFO} -qp ${def:C/^.*=//} 2>/dev/null | ${AWK} '{ print $$2; exit }' | ${GREP} . || ${ECHO} ${${def:C/=.*$//}_DEFAULT}
Merge down deferred PREPEND_PATH handling from the tv-derecurse branch. Also merge in deferred CLASSPATH handling, which can now be moved back to java-vm.mk for cleanliness.
2005-01-24 20:32:33 +01:00
${def:C/=.*$//}= ${_${def:C/=.*$//}_CMD:sh}
MAKEFLAGS+= ${def:C/=.*//}=${_${def:C/=.*$//}_CMD:sh}
. endif
. endfor
.endif
.if !defined(_PATH_ORIG)
_PATH_ORIG:= ${PATH}
MAKEFLAGS+= _PATH_ORIG=${_PATH_ORIG:Q}
Fix handling of PREPEND_PATH in dependency-recursed builds.
2005-01-27 12:21:46 +01:00
.endif
Merge down deferred PREPEND_PATH handling from the tv-derecurse branch. Also merge in deferred CLASSPATH handling, which can now be moved back to java-vm.mk for cleanliness.
2005-01-24 20:32:33 +01:00
Fix handling of PREPEND_PATH in dependency-recursed builds.
2005-01-27 12:21:46 +01:00
.if !empty(PREPEND_PATH)
Merge down deferred PREPEND_PATH handling from the tv-derecurse branch. Also merge in deferred CLASSPATH handling, which can now be moved back to java-vm.mk for cleanliness.
2005-01-24 20:32:33 +01:00
# This is very Special. Because PREPEND_PATH is set with += in reverse order,
# the awk expression reverses the order again (since bootstrap bmake doesn't
# yet support the :[-1..1] construct).
_PATH_CMD= ${ECHO} `${ECHO} ${PREPEND_PATH:Q} | ${AWK} '{ORS=":";for (i=NF;i>0;i--) print $$i}'`${_PATH_ORIG}
PATH= ${_PATH_CMD:sh} # DOES NOT use :=, to defer evaluation
.endif
* Move pkgsrc/mk/compiler/bsd.compiler.mk to pkgsrc/mk/compiler.mk. Package Makefiles may now directly include compiler.mk. * Don't include compiler.mk within bsd.prefs.mk any longer. It was only included for the purposes of defining CC_VERSION. Packages that want to test the value of CC_VERSION should now first include "../../mk/compiler.mk". Any GCC_REQD statements in package Makefiles should be set before compiler.mk is included. * Simpllfy pkgsrc/mk/compiler/*.mk files as a result of not needing to be included indirectly by bsd.prefs.mk. We remove the special handling associated with detecting whether the file was included from within bsd.prefs.mk. These files are now much more straightforward to write and understand. * G/C the BSD_PREFS_MK stack mechanism as the only users (compiler/*) no longer need it. * Ensure that directories are prepended to the PATH only from within bsd.pkg.mk.
2004-02-18 14:32:38 +01:00
Rename PATH_ENV to BUILD_ENV, which better expresses what it's for. BUILD_ENV is used to set the environment for the sub-make processes that are invoked for each phase after and including buildlink.
2004-02-21 13:29:23 +01:00
# Add these bits to the environment use when invoking the sub-make
# processes for build-related phases.
#
Move the PREPEND_PATH for LOCALBASE/bin (and optionally X11BASE/bin) to a spot that will come before compiler.mk (in bsd.prefs.mk). Previously, LOCALBASE/bin was appearing earlier in the path than work/.<compiler>/bin, which could cause the Wrong Thing to happen.
2005-01-27 05:05:08 +01:00
BUILD_ENV+= PATH=${PATH:Q}
We only prepend a directory to the PATH if we haven't already done so (by checking PREPEND_PATH) and only for those phases of the build that care about the PATH (buildlink or later). We also pass the PATH to those same phases of the build so that executing ${CC} will work correctly from custom {pre,do,post}-* targets that occur at buildlink time or later.
2004-02-06 04:04:50 +01:00
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.MAIN: all
################################################################
This is NetBSD. s/port/package/g
1999-08-10 12:48:23 +02:00
# Many ways to disable a package.
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
#
Revert last change. I only tested this on netbsd-1-4; it doesn't seem to work on netbsd-1-5 at all.
2000-09-07 04:29:40 +02:00
# Ignore packages that can't be resold if building for a CDROM.
Introduce a new category, "crypto", with a special property: If the user sets MKCRYPTO=no, packages in this category won't be fetched, installed, built, or packaged. Also, binary package users forbidden, by law, from using strong cryptography would presumably find the list on the category's automatically generated web page useful for ensuring compliance.
2000-09-06 22:28:32 +02:00
#
Correct a factual error in the comment "ways to disable a package": we actually check for "X", and not just Motif. Also, clean up word wrap, and rearrange the sections to be in the same order as the checks they refer to.
2000-09-07 17:26:52 +02:00
# Don't build a package if it's restricted and we don't want to
# get into that.
#
Add a new mechanism for handling "crypto" packages, as discussed on tech-pkg. Setting either USE_SSL or CRYPTO in the package's Makefile will permit the package to build or not, depending on the setting of MKCRYPTO (set to yes by default).
2000-09-09 20:45:16 +02:00
# Don't build any package that utilizes strong cryptography, for
# when the law of the land forbids it.
#
Correct a factual error in the comment "ways to disable a package": we actually check for "X", and not just Motif. Also, clean up word wrap, and rearrange the sections to be in the same order as the checks they refer to.
2000-09-07 17:26:52 +02:00
# Don't attempt to build packages against X if we don't have X.
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
#
This is NetBSD. s/port/package/g
1999-08-10 12:48:23 +02:00
# Don't build a package if it's broken.
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
################################################################
* Garbage collect IGNORE -> SKIP migration. * {NOT,ONLY}_FOR_PLATHOME mismatch is not an error. Set PKG_SKIP_REASON for those cases. This makes bulk builds happier.
2002-12-27 07:53:42 +01:00
.if !defined(NO_SKIP)
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if (defined(NO_BIN_ON_CDROM) && defined(FOR_CDROM))
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
PKG_FAIL_REASON+= "${PKGNAME} may not be placed in binary form on a CDROM:" \
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
" "${NO_BIN_ON_CDROM:Q}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. if (defined(NO_SRC_ON_CDROM) && defined(FOR_CDROM))
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
PKG_FAIL_REASON+= "${PKGNAME} may not be placed in source form on a CDROM:" \
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
" "${NO_SRC_ON_CDROM:Q}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. if (defined(RESTRICTED) && defined(NO_RESTRICTED))
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
PKG_FAIL_REASON+= "${PKGNAME} is restricted:" \
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
" "${RESTRICTED:Q}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. if !(${MKCRYPTO} == "YES" || ${MKCRYPTO} == yes)
Deprecate USE_SSL.
2002-07-02 17:25:49 +02:00
. if defined(CRYPTO)
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
PKG_FAIL_REASON+= "${PKGNAME} may not be built, because it utilizes strong cryptography"
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Add a new mechanism for handling "crypto" packages, as discussed on tech-pkg. Setting either USE_SSL or CRYPTO in the package's Makefile will permit the package to build or not, depending on the setting of MKCRYPTO (set to yes by default).
2000-09-09 20:45:16 +02:00
. endif
Use X11_TYPE check instead of USE_PKGSRC_XFREE86. It's safe to use the value of X11_TYPE here since it's defined in bsd.buildlink3.mk which is included before this check. This fixes breakage for packages that set USE_X11 due to USE_PKGSRC_XFREE86 not being defined before it's used.
2004-03-29 18:01:39 +02:00
. if defined(USE_X11) && !exists(${X11BASE})
. if ${X11_TYPE} == "native"
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
PKG_FAIL_REASON+= "${PKGNAME} uses X11, but ${X11BASE} not found"
Use X11_TYPE check instead of USE_PKGSRC_XFREE86. It's safe to use the value of X11_TYPE here since it's defined in bsd.buildlink3.mk which is included before this check. This fixes breakage for packages that set USE_X11 due to USE_PKGSRC_XFREE86 not being defined before it's used.
2004-03-29 18:01:39 +02:00
. else
${_PKG_SILENT}${_PKG_DEBUG}${MKDIR} ${X11BASE}
${_PKG_SILENT}${_PKG_DEBUG}${CHOWN} ${ROOT_USER}:${ROOT_GROUP} ${X11BASE}
${_PKG_SILENT}${_PKG_DEBUG}${CHMOD} ${PKGDIRMODE} ${X11BASE}
. endif
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. if defined(BROKEN)
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
PKG_FAIL_REASON+= "${PKGNAME} is marked as broken:" ${BROKEN:Q}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(LICENSE)
. ifdef ACCEPTABLE_LICENSES
. for _lic in ${ACCEPTABLE_LICENSES}
. if ${LICENSE} == "${_lic}"
Add an ACCEPTABLE_LICENCES definition to /etc/mk.conf. This is a whitespace-delimited list, and details which types of licences are allowed. Package Makefiles with non-standard licences (shareware, non-commercial-use only) should have a LICENCE= <type> added. Normal package Makefiles remain unchanged. Add logic to bsd.pkg.mk to check, if LICENCE is defined, that the package is acceptable.
1998-07-06 17:10:48 +02:00
_ACCEPTABLE= yes
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif # LICENSE == _lic
. endfor # _lic
. endif # ACCEPTABLE_LICENSES
. ifndef _ACCEPTABLE
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
PKG_FAIL_REASON+= "${PKGNAME} has an unacceptable license: ${LICENSE}." \
Update the wording a bit when a license is not accepted, to better indicate the proper process: first read the license, then accept, and indicate so in /etc/mk.conf: yui% make ===> graphviz-1.12 has an unacceptable license: graphviz-license. ===> To view the license, enter "/usr/bin/make show-license". ===> To indicate acceptance, add this line to your /etc/mk.conf: ===> ACCEPTABLE_LICENSES+=graphviz-license *** Error code 1
2004-10-01 02:23:18 +02:00
" To view the license, enter \"${MAKE} show-license\"." \
" To indicate acceptance, add this line to your /etc/mk.conf:" \
" ACCEPTABLE_LICENSES+=${LICENSE}"
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif # _ACCEPTABLE
. endif # LICENSE
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
Move all settings of IGNORE into one spot. No need for the arch-checking code (ONLY_FOR_PLATFORM, NOT_FOR_PLATFORM) to be located elsewhere. Inspired by OpenBSD.
2000-05-31 00:36:53 +02:00
# Define __PLATFORM_OK only if the OS matches the pkg's allowed list.
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(ONLY_FOR_PLATFORM) && !empty(ONLY_FOR_PLATFORM)
. for __tmp__ in ${ONLY_FOR_PLATFORM}
. if ${MACHINE_PLATFORM:M${__tmp__}} != ""
Move all settings of IGNORE into one spot. No need for the arch-checking code (ONLY_FOR_PLATFORM, NOT_FOR_PLATFORM) to be located elsewhere. Inspired by OpenBSD.
2000-05-31 00:36:53 +02:00
__PLATFORM_OK?= yes
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif # MACHINE_PLATFORM
. endfor # __tmp__
. else # !ONLY_FOR_PLATFORM
__PLATFORM_OK?= yes
. endif # ONLY_FOR_PLATFORM
. for __tmp__ in ${NOT_FOR_PLATFORM}
. if ${MACHINE_PLATFORM:M${__tmp__}} != ""
. undef __PLATFORM_OK
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
. endif # MACHINE_PLATFORM
. endfor # __tmp__
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if !defined(__PLATFORM_OK)
* Garbage collect IGNORE -> SKIP migration. * {NOT,ONLY}_FOR_PLATHOME mismatch is not an error. Set PKG_SKIP_REASON for those cases. This makes bulk builds happier.
2002-12-27 07:53:42 +01:00
PKG_SKIP_REASON+= "${PKGNAME} is not available for ${MACHINE_PLATFORM}"
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif # !__PLATFORM_OK
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
#
# Now print some error messages that we know we should ignore the pkg
#
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
. if defined(PKG_FAIL_REASON) || defined(PKG_SKIP_REASON)
do not check for ACCEPTABLE_LICENCES when deinstalling. closes PR pkg/13211
2001-06-19 13:54:56 +02:00
fetch checksum extract patch configure all build install package \
Do not attempt to update a package (deleting the package and all its dependencies) if IGNORE is set. This puts the 'update' target in line with 'install', 'package', and other targets that use the same checks.
2002-04-06 01:26:49 +02:00
update install-depends:
* Garbage collect IGNORE -> SKIP migration. * {NOT,ONLY}_FOR_PLATHOME mismatch is not an error. Set PKG_SKIP_REASON for those cases. This makes bulk builds happier.
2002-12-27 07:53:42 +01:00
. if defined(SKIP_SILENT)
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
@${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
@for str in ${PKG_FAIL_REASON} ${PKG_SKIP_REASON}; do \
${ECHO} "${_PKGSRC_IN}> $$str"; \
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
done
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
. if defined(PKG_FAIL_REASON)
@${FALSE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
* Garbage collect IGNORE -> SKIP migration. * {NOT,ONLY}_FOR_PLATHOME mismatch is not an error. Set PKG_SKIP_REASON for those cases. This makes bulk builds happier.
2002-12-27 07:53:42 +01:00
. endif # SKIP
.endif # !NO_SKIP
Re-move .if!defined(__ARCH_OK) to fix use of NO_IGNORE; document some .endif's
1998-01-25 15:46:10 +01:00
+ Add and document the PKG_RELATIVE_SYMLINKS option, which makes sure that a symbolic link in a package is relative to ${PREFIX}, if it is under ${PREFIX}, using the -l argument to pkg_create. + Build a +BUILD_INFO file with appropriate information from the package build. Thanks to Christoph Badura for pointers to the appropriate information.
1998-10-20 18:01:03 +02:00
# Add these defs to the ones dumped into +BUILD_DEFS
* Make sure PKGPATH gets added to BUILD_DEFS (I thought it was there for a long time. Oh well.) * Only replace the value of PATH for "PATH", not any variable whose name starts with PATH (like PKGPATH :-) Hinted by Jason R. Mastaler <jason@mastaler.com> on tech-pkg.
2000-11-27 23:24:49 +01:00
BUILD_DEFS+= PKGPATH
+ Add and document the PKG_RELATIVE_SYMLINKS option, which makes sure that a symbolic link in a package is relative to ${PREFIX}, if it is under ${PREFIX}, using the -l argument to pkg_create. + Build a +BUILD_INFO file with appropriate information from the package build. Thanks to Christoph Badura for pointers to the appropriate information.
1998-10-20 18:01:03 +02:00
BUILD_DEFS+= OPSYS OS_VERSION MACHINE_ARCH MACHINE_GNU_ARCH
make sure FFLAGS ends up in MAKE_ENV as well as BUILD_DEFS. This allows you to do make FFLAGS='-my -f -flags' and at least the pkgs which obey FFLAGS use them.
2002-01-22 19:15:20 +01:00
BUILD_DEFS+= CPPFLAGS CFLAGS FFLAGS LDFLAGS
+ Add and document the PKG_RELATIVE_SYMLINKS option, which makes sure that a symbolic link in a package is relative to ${PREFIX}, if it is under ${PREFIX}, using the -l argument to pkg_create. + Build a +BUILD_INFO file with appropriate information from the package build. Thanks to Christoph Badura for pointers to the appropriate information.
1998-10-20 18:01:03 +02:00
BUILD_DEFS+= CONFIGURE_ENV CONFIGURE_ARGS
Add 'RESTRICTED' to ${BUILD_DEFS}, so that any restrictions are recorded in the binary package.
2000-07-23 20:42:48 +02:00
BUILD_DEFS+= OBJECT_FMT LICENSE RESTRICTED
* Replace MIRROR_DISTFILES with NO_SRC_ON_FTP * Replace NO_CDROM with NO_SRC_ON_CDROM and NO_BIN_ON_CDROM * Add NO_{SRC,BIN}_ON_{FTP,CDROM} to BUILD_DEFS (in case someone wants to re-distribute a binary pkg, so he has a chance to know about resitrictions) * Adjust 'mirror-distfiles' target for NO_SRC_ON_FTP
2000-08-19 00:43:19 +02:00
BUILD_DEFS+= NO_SRC_ON_FTP NO_SRC_ON_CDROM
BUILD_DEFS+= NO_BIN_ON_FTP NO_BIN_ON_CDROM
+ Add and document the PKG_RELATIVE_SYMLINKS option, which makes sure that a symbolic link in a package is relative to ${PREFIX}, if it is under ${PREFIX}, using the -l argument to pkg_create. + Build a +BUILD_INFO file with appropriate information from the package build. Thanks to Christoph Badura for pointers to the appropriate information.
1998-10-20 18:01:03 +02:00
add OSVERSION_SPECIFIC to BUILD_DEFS if defined so it ends up in the binary package. This way we can look at a binary package and determine what versions of the OS it may run on.
2001-07-07 19:10:36 +02:00
.if defined(OSVERSION_SPECIFIC)
BUILD_DEFS+= OSVERSION_SPECIFIC
.endif # OSVERSION_SPECIFIC
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: all
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(all)
Due to the special handling of pkgtools/digest in pkgsrc (it must be installed before the normal dependency checks), on a system without any packages installed, we can run into the problem where digest won't install because it has already been installed as part of installing a build dependency. This can happen if, e.g., PKGSRC_COMPILER is set to "ccache gcc", and you try to build devel/gmake. Change the way that the uptodate-digest target installs pkgtools/digest by ensuring that pkgtools/digest is deinstalled just before installing it.
2004-03-22 05:07:15 +01:00
all: ${_PKGSRC_BUILD_TARGETS}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
.if !defined(DEPENDS_TARGET)
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if make(package)
Set DEPENDS_TARGET=package during a "make package", and propagate DEPENDS_TARGET to builds for dependencies. This results in "make package" creating packages for all missing dependencies as well, instead of merely installing them on the system. This is a more consistent and intuitive behaviour. Running "make install" will install the dependencies just like before (no binary packages will be built).
2000-04-24 05:40:50 +02:00
DEPENDS_TARGET= package
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. elif make(update)
Default DEPENDS_TARGET to replace if UPDATE_TARGET == replace. This avoids multiple rebuilding attempts for prerequisite packages.
2003-05-04 04:59:15 +02:00
. if defined(UPDATE_TARGET) && ${UPDATE_TARGET} == "replace"
DEPENDS_TARGET= ${UPDATE_TARGET}
. else
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
DEPENDS_TARGET= update
Default DEPENDS_TARGET to replace if UPDATE_TARGET == replace. This avoids multiple rebuilding attempts for prerequisite packages.
2003-05-04 04:59:15 +02:00
. endif
bin-install: set DEPENDS_TARGET=bin-install, so depends are still installed via binaries (if possible), even if there is no binary pkg available for the current pkg. Implementation still allows overriding DEPENDS_TARGET=build (etc.) to force installation of dependencies by compiling if desired. Patch submitted by Marko Schuetz <MarkoSchuetz@web.de> in private mail.
2002-04-25 23:58:55 +02:00
. elif make(bin-install)
DEPENDS_TARGET= bin-install
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Use "reinstall" as default for DEPENDS_TARGET, to use any pkg already built and possible deinstalled as dependency. Addresses PR 10629 by Erich.T.Enke@wheaton.edu. (If that's not enough, we can still introduce a "make clean ; make install"- type target, but for now i'd prefer not to was the cycles)
2000-07-20 00:57:31 +02:00
DEPENDS_TARGET= reinstall
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
- Add OpenBSD and FreeBSD Ids in the header - Changes from FreeBSD's V1.164: * Warn if mtree-file not found * On 'make reinstall' gets DEPENDS_TARGET set to 'reinstall' instead of 'install' - Changes from OpenBSD's V1.14: * Documentation and implementation for ONLY_FOR_ARCHS * Documentation and implementation for WRKOBJDIR * Documentation on HAVE_MOTIF and MOTIF_STATIC: set in /etc/mk.conf, not /etc/make.conf * Set NOMANCOMPRESS=no to still compress man-pages (should probably be set to 'yes') * DEF_UMASK=022 * add ${MACHINE} suffix to WRKDIR if OBJMACHINE is set * .include <bsd.own.mk>
1997-09-28 02:44:15 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
.if !defined(UPDATE_TARGET)
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if ${DEPENDS_TARGET} == "update"
. if make(package)
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
UPDATE_TARGET= package
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
UPDATE_TARGET= install
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. else
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
UPDATE_TARGET= ${DEPENDS_TARGET}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
.endif
Set UPDATE_RUNNING=YES when deinstalling due to a update or replace target. Use this to trigger appropriate args to pkg_delete so PKG_PRESERVE packages can be upgraded.
2004-11-02 01:03:09 +01:00
UPDATE_RUNNING?= NO
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
################################################################
# The following are used to create easy dummy targets for
# disabling some bit of default target behavior you don't want.
# They still check to see if the target exists, and if so don't
# do anything, since you might want to set this globally for a
This is NetBSD. s/port/package/g
1999-08-10 12:48:23 +02:00
# group of packages in a Makefile.inc, but still be able to
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# override from an individual Makefile.
################################################################
# Disable checksum
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: checksum
Kill the odd "duplicate checksum target" warning by rearranging a test condition. This only happened when we set PKG_*_REASON and have extracted the package.
2003-09-13 16:40:40 +02:00
.if (defined(NO_CHECKSUM) || exists(${EXTRACT_COOKIE})) && !target(checksum)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
checksum: fetch
@${DO_NADA}
.endif
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
# Disable wrapper
.PHONY: wrapper
USE_BUILDLINK3 is no longer optional (and cannot be turned off). Per mail to tech-pkg: ===== * USE_BUILDLINK3=YES will be unconditional. (In fact, USE_BUILDLINK3 will be ignored altogether by mk/; but see below.) * NO_BUILDLINK and NO_WRAPPER will be ignored by mk/. If a build happens, these phases will happen. * The existing NO_BUILD will imply the previous NO_BUILDLINK and NO_WRAPPER. If no build happens, those phases are not needed. * NO_TOOLS will be ignored by mk/. The tools phase, which provides much more than just the C compiler, will always happen regardless of package. This will make metapackage builds only slightly slower, in trade for far less user error.
2005-03-24 18:46:00 +01:00
.if defined(NO_BUILD) && !target(wrapper)
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
wrapper: tools
${_PKG_SILENT}${_PKG_DEBUG}${TOUCH} ${TOUCH_FLAGS} ${WRAPPER_COOKIE}
Merge pkgsrc/mk from the buildlink2 branch back into the main trunk.
2002-08-22 10:21:26 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Disable configure
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: configure
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if defined(NO_CONFIGURE) && !target(configure)
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
configure: wrapper
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${TOUCH} ${TOUCH_FLAGS} ${CONFIGURE_COOKIE}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
# Disable build
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: build
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if defined(NO_BUILD) && !target(build)
build: configure
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${TOUCH} ${TOUCH_FLAGS} ${BUILD_COOKIE}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
# Disable install
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: install
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if defined(NO_INSTALL) && !target(install)
install: build
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${TOUCH} ${TOUCH_FLAGS} ${INSTALL_COOKIE}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
# Disable package
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: package
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if defined(NO_PACKAGE) && !target(package)
package:
* Garbage collect IGNORE -> SKIP migration. * {NOT,ONLY}_FOR_PLATHOME mismatch is not an error. Set PKG_SKIP_REASON for those cases. This makes bulk builds happier.
2002-12-27 07:53:42 +01:00
. if defined(SKIP_SILENT)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
@${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> ${PKGNAME} may not be packaged: ${NO_PACKAGE}."
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
################################################################
# More standard targets start here.
#
# These are the body of the build/install framework. If you are
# not happy with the default actions, and you can't solve it by
# adding pre-* or post-* targets/scripts, override these.
################################################################
2nd round for PKG_RESUME_TRANSFERS: * Add FETCH_OUTPUT_ARGS (new option, defaults to "-o" with NetBSD's ftp(1)) * Use FETCH_OUTPUT_ARGS to move the file transfer to a temporary name on ${DISTDIR}/${DIST_SUBDIR} with extension ".temp" * If temporary file matches the checksum recorded in distinfo, move it to the original name (removing temp file) For example, if you want to use PKG_RESUME_TRANSFERS with wget (pkgsrc/net/wget), the following vars should be defined in mk.conf: FETCH_CMD=wget FETCH_RESUME_ARGS=-c FETCH_OUTPUT_ARGS=-O No need to set these vars when using defaults (NetBSD's ftp(1))
2004-10-26 23:14:59 +02:00
###
### _RESUME_TRANSFER:
###
### Macro to resume a previous transfer, needs to have defined
### the following options in mk.conf:
###
### PKG_RESUME_TRANSFERS
### FETCH_RESUME_ARGS (if FETCH_CMD != default)
### FETCH_OUTPUT_ARGS (if FETCH_CMD != default)
###
### For example if you want to use wget (pkgsrc/net/wget):
###
### FETCH_CMD=wget
### FETCH_RESUME_ARGS=-c
### FETCH_OUTPUT_ARGS=-O
###
### How does it work?
###
### FETCH_CMD downloads the file and saves it temporally into $$bfile.temp
### if the checksum match the correct one, $$bfile.temp is renamed to
### the original name.
###
Some changes for PKG_RESUME_TRANSFERS: o If origfile is already downloaded (size ok) and there's a temp file with incomplete size, don't try to resume the transfer on the temp file, remove it. o Clean the code a bit.
2005-03-22 23:20:21 +01:00
_RESUME_TRANSFER= \
ofile="${DISTDIR}/${DIST_SUBDIR}/$$bfile"; \
tfile="${DISTDIR}/${DIST_SUBDIR}/$$bfile.temp"; \
Fix some more troubles with my modification for resume transfers: o Use ${WC} -c instead of using ${LS} -l, which works with symlinks. o Only assign -R to FETCH_RESUME_ARGS when FETCH_CMD is the default value (NetBSD's ftp). o Fix for www/navigator package, which uses DIST_SUBDIR. o Replace another [ with ${TEST}. o Print a warning when FETCH_CMD != ftp and FETCH_RESUME_ARGS is empty (resume will not be possible). Thanks to cube@ and tron@ for reporting/helping with the problems.
2004-07-27 12:34:53 +02:00
tsize=`${AWK} '/^Size/ && $$2 == '"\"($$file)\""' { print $$4 }' ${DISTINFO_FILE}` || ${TRUE}; \
Some changes for PKG_RESUME_TRANSFERS: o If origfile is already downloaded (size ok) and there's a temp file with incomplete size, don't try to resume the transfer on the temp file, remove it. o Clean the code a bit.
2005-03-22 23:20:21 +01:00
osize=`${WC} -c < $$ofile`; \
\
case "$$tsize" in \
"") ${ECHO_MSG} "No size in distinfo file (${DISTINFO_FILE})"; \
break;; \
esac; \
\
if [ "$$osize" -eq "$$tsize" ]; then \
if [ -f "$$tfile" ]; then \
${RM} $$tfile; \
fi; \
need_fetch=no; \
break; \
elif [ "$$osize" -lt "$$tsize" -a ! -f "$$tfile" ]; then \
${CP} $$ofile $$tfile; \
dsize=`${WC} -c < $$tfile`; \
need_fetch=yes; \
elif [ -f "$$tfile" -a "$$dsize" -gt "$$ossize" ]; then \
dsize=`${WC} -c < $$tfile`; \
need_fetch=yes; \
else \
dsize=`${WC} -c < $$tfile`; \
need_fetch=yes; \
fi; \
if [ "$$need_fetch" = "no" ]; then \
break; \
elif [ -f "$$tfile" -a "$$dsize" -eq "$$tsize" ]; then \
${MV} $$tfile $$ofile; \
break; \
elif [ -n "$$ftp_proxy" -o -n "$$http_proxy" ]; then \
Add support to resume transfers, by default it will be disabled if $http_proxy or $ftp_proxy are defined (for the moment, NetBSD's ftp(1) doesn't support this). While I'm here, move FETCH_CMD from bsd.pkg.mk to bsd.pkg.defaults.mk, and add the new variable FETCH_RESUME_ARGS (-R by default with ftp(1) from NetBSD/pkgsrc). If you want to use wget to resume transfers, you should use something like: FETCH_CMD=wget FETCH_RESUME_ARGS=-c
2004-07-27 05:59:26 +02:00
${ECHO_MSG} "===> Resume is not supported by ftp(1) using http/ftp proxies."; \
Some changes for PKG_RESUME_TRANSFERS: o If origfile is already downloaded (size ok) and there's a temp file with incomplete size, don't try to resume the transfer on the temp file, remove it. o Clean the code a bit.
2005-03-22 23:20:21 +01:00
break; \
elif [ "$$need_fetch" = "yes" -a "$$dsize" -lt "$$tsize" ]; then \
if [ "${FETCH_CMD:T}" != "ftp" -a -z "${FETCH_RESUME_ARGS}" ]; then \
${ECHO_MSG} "=> Resume transfers are not supported, FETCH_RESUME_ARGS is empty."; \
break; \
else \
for res_site in $$sites; do \
if [ -z "${FETCH_OUTPUT_ARGS}" ]; then \
${ECHO_MSG} "=> FETCH_OUTPUT_ARGS has to be defined."; \
break; \
fi; \
${ECHO_MSG} "=> $$bfile not completed, resuming:"; \
${ECHO_MSG} "=> Downloaded: $$dsize Total: $$tsize."; \
${ECHO_MSG}; \
cd ${DISTDIR}; \
${FETCH_CMD} ${FETCH_BEFORE_ARGS} ${FETCH_RESUME_ARGS} \
${FETCH_OUTPUT_ARGS} $${bfile}.temp $${res_site}$${bfile}; \
if [ $$? -eq 0 ]; then \
ndsize=`${WC} -c < $$tfile`; \
if [ "$$tsize" -eq "$$ndsize" ]; then \
${MV} $$tfile $$ofile; \
fi; \
break; \
fi; \
done; \
fi; \
elif [ "$$dsize" -gt "$$tsize" ]; then \
${ECHO_MSG} "==> Downloaded file larger than the recorded size."; \
break; \
Add support to resume transfers, by default it will be disabled if $http_proxy or $ftp_proxy are defined (for the moment, NetBSD's ftp(1) doesn't support this). While I'm here, move FETCH_CMD from bsd.pkg.mk to bsd.pkg.defaults.mk, and add the new variable FETCH_RESUME_ARGS (-R by default with ftp(1) from NetBSD/pkgsrc). If you want to use wget to resume transfers, you should use something like: FETCH_CMD=wget FETCH_RESUME_ARGS=-c
2004-07-27 05:59:26 +02:00
fi
Fix some problems with REFETCH mods: 1. don't hardcode true, echo and [ 2. Cope with distfiles which are located in a DIST_SUBDIR 3. Don't always assume there will be a size field in the ${DISTINFO_FILE} XXX The refetch mods make an assumption that an exit code of 0 from a ${FETCH_CMD} means that the transfer has been completed successfully, and this used not to be the case. I haven't tested this assumption recently.
2004-07-27 11:57:57 +02:00
Re-work the fetch target(s), as discussed on "tech-pkg". The primary motivation is to make the order that sites are hit more sane, and especially to prevent hitting sites that aren't ever expected to have the sought after file. Now, ${MASTER_SITE_OVERRIDE} is always hit first. If that fails, then the sites designated for that file, then ${MASTER_SITE_BACKUP}, are hit (by default), but with the order subject to ${MASTER_SORT_REGEX} and ${MASTER_SORT_AWK}. The "designated sites" are usually ${MASTER_SITES} for files in ${DISTFILES}, and ${PATCH_SITES} for files in ${PATCH_FILES}. However, defining a variable `${SITES_foo}' in the package "Makefile" overrides that for file "foo". [The use of ${MASTER_SITES_foo} and ${PATCH_SITES_foo}, which is currently only used by a couple of packages for the same purpose, is deprecated, and will be shortly unsupported.] Also eliminate redundancy in the do-fetch and fetch-list-one-pkg targets, by making them use a single, common macro (escaped with `:Q' in the fetch-list-one-pkg case), so "make fetch-list-one-pkg | sh" now does exactly the same thing as "make do-fetch".
2002-01-06 21:03:22 +01:00
#
# Define the elementary fetch macros.
#
Combine fetch loops for distribution and patch files to reduce duplicate code. Patch supplied by Jim Bernard in PR pkg/5977.
1998-09-05 21:45:25 +02:00
_FETCH_FILE= \
Don't try to fetch a file if there's a symbolic link in ${DISTDIR}.
1999-07-28 12:58:30 +02:00
if [ ! -f $$file -a ! -f $$bfile -a ! -h $$bfile ]; then \
In informational messages to the user, use '=>' in preference to '>>', so that cut-n-paste into send-pr will work correctly. Fixes PR 9173 from ITOH Yasufumi.
2000-03-10 17:07:37 +01:00
${ECHO_MSG} "=> $$bfile doesn't seem to exist on this system."; \
Check if $DISTDIR is writable before downloading something (and wonder if the "permission denied" comes from the distfile being unreadable)
2000-05-31 03:07:14 +02:00
if [ ! -w ${_DISTDIR}/. ]; then \
${ECHO_MSG} "=> Can't download to ${_DISTDIR} (permission denied?)."; \
exit 1; \
fi; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
for site in $$sites; do \
In informational messages to the user, use '=>' in preference to '>>', so that cut-n-paste into send-pr will work correctly. Fixes PR 9173 from ITOH Yasufumi.
2000-03-10 17:07:37 +01:00
${ECHO_MSG} "=> Attempting to fetch $$bfile from $${site}."; \
Fix for PR 15761 from Jonathan Perkin <sketch@rd.bbc.co.uk> - only print the number of bytes about to be downloaded, as gleaned from the distinfo file, if a distinfo file exists.
2002-03-01 14:16:45 +01:00
if [ -f ${DISTINFO_FILE} ]; then \
${AWK} 'NF == 5 && $$1 == "Size" && $$2 == "('$$bfile')" { printf("=> [%s %s]\n", $$4, $$5) }' ${DISTINFO_FILE}; \
fi; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
if ${FETCH_CMD} ${FETCH_BEFORE_ARGS} $${site}$${bfile} ${FETCH_AFTER_ARGS}; then \
Deprecate DIGEST_FILE and PATCH_SUM_FILE (replaced by DIGEST_FILE, and not used anymore in pkgsrc).
2001-06-09 14:15:59 +02:00
if [ -n "${FAILOVER_FETCH}" -a -f ${DISTINFO_FILE} -a -f ${_DISTDIR}/$$bfile ]; then \
When FAILOVER_FETCH is set, only use the first digest algorithm found for a dist file or patch, and let the "checksum" target do the full digest integrity checks. Should fix a problem reported by John Klos on tech-pkg, which I was a bit dim in analysing.
2005-03-02 12:11:36 +01:00
alg=`${AWK} 'NF == 4 && $$2 == "('$$file')" && $$3 == "=" {print $$1; exit}' ${DISTINFO_FILE}`; \
Provide a default digest algorithm in the case where FAILOVER_FETCH is set, and a distfile which isn't in files/md5 is attempted to be downloaded. Fixes a problem pointed out by Dan Mcmahill.
2001-03-26 19:08:19 +02:00
if [ -z "$$alg" ]; then \
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
alg=${PATCH_DIGEST_ALGORITHM};\
Provide a default digest algorithm in the case where FAILOVER_FETCH is set, and a distfile which isn't in files/md5 is attempted to be downloaded. Fixes a problem pointed out by Dan Mcmahill.
2001-03-26 19:08:19 +02:00
fi; \
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
CKSUM=`${DIGEST} $$alg < ${_DISTDIR}/$$bfile`; \
Don't fail if there are multiple digest values for the same file with the same algorithm in the distinfo file - pointed out by Robert Elz in PR 29973. Additional testing by wiz - thanks!
2005-04-16 11:26:22 +02:00
CKSUM2=`${AWK} '$$1 == "'$$alg'" && $$2 == "('$$file')" {print $$4; exit}' <${DISTINFO_FILE}`; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
if [ "$$CKSUM" = "$$CKSUM2" -o "$$CKSUM2" = "IGNORE" ]; then \
Add support for distfile-specific master sites, as requested in pkg/7471. Syntax: MASTER_SITES_completefilename= http://specific.master/site and similarly for PATCH_SITES. Convert print/ghostscript-nox11 and x11/kterm to take advantage of this.
2001-12-02 22:29:20 +01:00
break; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
else \
In informational messages to the user, use '=>' in preference to '>>', so that cut-n-paste into send-pr will work correctly. Fixes PR 9173 from ITOH Yasufumi.
2000-03-10 17:07:37 +01:00
${ECHO_MSG} "=> Checksum failure - trying next site."; \
Amend the fetch: target, so that the distfile's integrity is checked with the additional checksum only if FAILOVER_FETCH is set (this does not affect the checksum: target at all). Document FAILOVER_FETCH.
1998-09-01 15:15:29 +02:00
fi; \
Some master-sites (like the second one of the libslang package) seem to return some "please see the new location"-pages instead of the requested archive, which our ftp(1) doesn't take as an error. Now, add a check here to see if ftp(1) actually got the requested file or something else. Detected by Reinoud Zandijk <zandijk@cs.utwente.nl>, talked over with Frederick Bruckman <fb@enteract.com>.
1999-07-21 14:16:27 +02:00
elif [ ! -f ${_DISTDIR}/$$bfile ]; then \
In informational messages to the user, use '=>' in preference to '>>', so that cut-n-paste into send-pr will work correctly. Fixes PR 9173 from ITOH Yasufumi.
2000-03-10 17:07:37 +01:00
${ECHO_MSG} "=> FTP didn't fetch expected file, trying next site." ; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
else \
Add support for distfile-specific master sites, as requested in pkg/7471. Syntax: MASTER_SITES_completefilename= http://specific.master/site and similarly for PATCH_SITES. Convert print/ghostscript-nox11 and x11/kterm to take advantage of this.
2001-12-02 22:29:20 +01:00
break; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
fi; \
fi \
done; \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
if [ ! -f ${_DISTDIR}/$$bfile ]; then \
Add support for distfile-specific master sites, as requested in pkg/7471. Syntax: MASTER_SITES_completefilename= http://specific.master/site and similarly for PATCH_SITES. Convert print/ghostscript-nox11 and x11/kterm to take advantage of this.
2001-12-02 22:29:20 +01:00
${ECHO_MSG} "=> Couldn't fetch $$bfile - please try to retrieve this";\
${ECHO_MSG} "=> file manually into ${_DISTDIR} and try again."; \
exit 1; \
fi; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
fi
Combine fetch loops for distribution and patch files to reduce duplicate code. Patch supplied by Jim Bernard in PR pkg/5977.
1998-09-05 21:45:25 +02:00
Actually apply any patch-local-* local patch files if there are any.
1999-07-27 17:01:30 +02:00
_CHECK_DIST_PATH= \
if [ "X${DIST_PATH}" != "X" ]; then \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
for d in "" ${DIST_PATH:S/:/ /g}; do \
Actually apply any patch-local-* local patch files if there are any.
1999-07-27 17:01:30 +02:00
if [ "X$$d" = "X" -o "X$$d" = "X${DISTDIR}" ]; then continue; fi; \
if [ -f $$d/${DIST_SUBDIR}/$$bfile ]; then \
${ECHO} "Using $$d/${DIST_SUBDIR}/$$bfile"; \
${RM} -f $$bfile; \
${LN} -s $$d/${DIST_SUBDIR}/$$bfile $$bfile; \
break; \
fi; \
done; \
fi
Add possibility to sort the master download site list according to local preferences (domain suffixes or full regular expressions) by defining MASTER_SORT or MASTER_SORT_REGEX appropriately.
1999-07-13 20:09:19 +02:00
#
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
# Set up ORDERED_SITES to work out the exact list of sites for every file,
# using the dynamic sites script, or sorting according to the master site
# list or the patterns in MASTER_SORT or MASTER_SORT_REGEX as appropriate.
# No actual sorting is done until ORDERED_SITES is expanded.
Add possibility to sort the master download site list according to local preferences (domain suffixes or full regular expressions) by defining MASTER_SORT or MASTER_SORT_REGEX appropriately.
1999-07-13 20:09:19 +02:00
#
Some tweaks to "fetch" and "fetch-list" re: DYNAMIC_MASTER_SITES: - - For "fetch", move the assignment of the "getsite.sh" script to a more natural place, to get the full benefit of, and consistent handling for, setting MASTER_SORT and MASTER_SORT_REGEX - - For "fetch-list", embed the "getsite.sh" script into the generated script, so that the calculation can be properly done at run-time. ("fetch-list" also respects MASTER_SORT and MASTER_SORT_REGEX.)
2002-03-10 18:15:29 +01:00
.if defined(MASTER_SORT) || defined(MASTER_SORT_REGEX)
Add possibility to sort the master download site list according to local preferences (domain suffixes or full regular expressions) by defining MASTER_SORT or MASTER_SORT_REGEX appropriately.
1999-07-13 20:09:19 +02:00
MASTER_SORT?=
MASTER_SORT_REGEX?=
MASTER_SORT_REGEX+= ${MASTER_SORT:S/./\\./g:C/.*/:\/\/[^\/]*&\//}
Fix awk warnings in MASTER_SORT* code. From Stoned Elipot (seb@).
2004-01-06 15:43:02 +01:00
MASTER_SORT_AWK= BEGIN { RS = " "; ORS = " "; IGNORECASE = 1 ; gl = "${MASTER_SORT_REGEX:S/\\/\\\\/g}"; }
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
. for srt in ${MASTER_SORT_REGEX}
Fix awk warnings in MASTER_SORT* code. From Stoned Elipot (seb@).
2004-01-06 15:43:02 +01:00
MASTER_SORT_AWK+= /${srt:C/\//\\\//g}/ { good["${srt:S/\\/\\\\/g}"] = good["${srt:S/\\/\\\\/g}"] " " $$0 ; next; }
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
. endfor
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
MASTER_SORT_AWK+= { rest = rest " " $$0; } END { n=split(gl, gla); for(i=1;i<=n;i++) { print good[gla[i]]; } print rest; }
Re-work the fetch target(s), as discussed on "tech-pkg". The primary motivation is to make the order that sites are hit more sane, and especially to prevent hitting sites that aren't ever expected to have the sought after file. Now, ${MASTER_SITE_OVERRIDE} is always hit first. If that fails, then the sites designated for that file, then ${MASTER_SITE_BACKUP}, are hit (by default), but with the order subject to ${MASTER_SORT_REGEX} and ${MASTER_SORT_AWK}. The "designated sites" are usually ${MASTER_SITES} for files in ${DISTFILES}, and ${PATCH_SITES} for files in ${PATCH_FILES}. However, defining a variable `${SITES_foo}' in the package "Makefile" overrides that for file "foo". [The use of ${MASTER_SITES_foo} and ${PATCH_SITES_foo}, which is currently only used by a couple of packages for the same purpose, is deprecated, and will be shortly unsupported.] Also eliminate redundancy in the do-fetch and fetch-list-one-pkg targets, by making them use a single, common macro (escaped with `:Q' in the fetch-list-one-pkg case), so "make fetch-list-one-pkg | sh" now does exactly the same thing as "make do-fetch".
2002-01-06 21:03:22 +01:00
SORT_SITES_CMD= ${ECHO} $$unsorted_sites | ${AWK} '${MASTER_SORT_AWK}'
Fix awk warnings in MASTER_SORT* code. From Stoned Elipot (seb@).
2004-01-06 15:43:02 +01:00
ORDERED_SITES= ${_MASTER_SITE_OVERRIDE} `${SORT_SITES_CMD:S/\\/\\\\/g:C/"/\"/g}`
Re-work the fetch target(s), as discussed on "tech-pkg". The primary motivation is to make the order that sites are hit more sane, and especially to prevent hitting sites that aren't ever expected to have the sought after file. Now, ${MASTER_SITE_OVERRIDE} is always hit first. If that fails, then the sites designated for that file, then ${MASTER_SITE_BACKUP}, are hit (by default), but with the order subject to ${MASTER_SORT_REGEX} and ${MASTER_SORT_AWK}. The "designated sites" are usually ${MASTER_SITES} for files in ${DISTFILES}, and ${PATCH_SITES} for files in ${PATCH_FILES}. However, defining a variable `${SITES_foo}' in the package "Makefile" overrides that for file "foo". [The use of ${MASTER_SITES_foo} and ${PATCH_SITES_foo}, which is currently only used by a couple of packages for the same purpose, is deprecated, and will be shortly unsupported.] Also eliminate redundancy in the do-fetch and fetch-list-one-pkg targets, by making them use a single, common macro (escaped with `:Q' in the fetch-list-one-pkg case), so "make fetch-list-one-pkg | sh" now does exactly the same thing as "make do-fetch".
2002-01-06 21:03:22 +01:00
.else
Put back the DIST_SUBDIR handling in the _FETCH_ALLFILES macro.
2002-01-07 00:01:31 +01:00
ORDERED_SITES= ${_MASTER_SITE_OVERRIDE} $$unsorted_sites
Re-work the fetch target(s), as discussed on "tech-pkg". The primary motivation is to make the order that sites are hit more sane, and especially to prevent hitting sites that aren't ever expected to have the sought after file. Now, ${MASTER_SITE_OVERRIDE} is always hit first. If that fails, then the sites designated for that file, then ${MASTER_SITE_BACKUP}, are hit (by default), but with the order subject to ${MASTER_SORT_REGEX} and ${MASTER_SORT_AWK}. The "designated sites" are usually ${MASTER_SITES} for files in ${DISTFILES}, and ${PATCH_SITES} for files in ${PATCH_FILES}. However, defining a variable `${SITES_foo}' in the package "Makefile" overrides that for file "foo". [The use of ${MASTER_SITES_foo} and ${PATCH_SITES_foo}, which is currently only used by a couple of packages for the same purpose, is deprecated, and will be shortly unsupported.] Also eliminate redundancy in the do-fetch and fetch-list-one-pkg targets, by making them use a single, common macro (escaped with `:Q' in the fetch-list-one-pkg case), so "make fetch-list-one-pkg | sh" now does exactly the same thing as "make do-fetch".
2002-01-06 21:03:22 +01:00
.endif
Add possibility to sort the master download site list according to local preferences (domain suffixes or full regular expressions) by defining MASTER_SORT or MASTER_SORT_REGEX appropriately.
1999-07-13 20:09:19 +02:00
Re-work the fetch target(s), as discussed on "tech-pkg". The primary motivation is to make the order that sites are hit more sane, and especially to prevent hitting sites that aren't ever expected to have the sought after file. Now, ${MASTER_SITE_OVERRIDE} is always hit first. If that fails, then the sites designated for that file, then ${MASTER_SITE_BACKUP}, are hit (by default), but with the order subject to ${MASTER_SORT_REGEX} and ${MASTER_SORT_AWK}. The "designated sites" are usually ${MASTER_SITES} for files in ${DISTFILES}, and ${PATCH_SITES} for files in ${PATCH_FILES}. However, defining a variable `${SITES_foo}' in the package "Makefile" overrides that for file "foo". [The use of ${MASTER_SITES_foo} and ${PATCH_SITES_foo}, which is currently only used by a couple of packages for the same purpose, is deprecated, and will be shortly unsupported.] Also eliminate redundancy in the do-fetch and fetch-list-one-pkg targets, by making them use a single, common macro (escaped with `:Q' in the fetch-list-one-pkg case), so "make fetch-list-one-pkg | sh" now does exactly the same thing as "make do-fetch".
2002-01-06 21:03:22 +01:00
#
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
# Associate each file to fetch with the correct site(s).
Re-work the fetch target(s), as discussed on "tech-pkg". The primary motivation is to make the order that sites are hit more sane, and especially to prevent hitting sites that aren't ever expected to have the sought after file. Now, ${MASTER_SITE_OVERRIDE} is always hit first. If that fails, then the sites designated for that file, then ${MASTER_SITE_BACKUP}, are hit (by default), but with the order subject to ${MASTER_SORT_REGEX} and ${MASTER_SORT_AWK}. The "designated sites" are usually ${MASTER_SITES} for files in ${DISTFILES}, and ${PATCH_SITES} for files in ${PATCH_FILES}. However, defining a variable `${SITES_foo}' in the package "Makefile" overrides that for file "foo". [The use of ${MASTER_SITES_foo} and ${PATCH_SITES_foo}, which is currently only used by a couple of packages for the same purpose, is deprecated, and will be shortly unsupported.] Also eliminate redundancy in the do-fetch and fetch-list-one-pkg targets, by making them use a single, common macro (escaped with `:Q' in the fetch-list-one-pkg case), so "make fetch-list-one-pkg | sh" now does exactly the same thing as "make do-fetch".
2002-01-06 21:03:22 +01:00
#
Some tweaks to "fetch" and "fetch-list" re: DYNAMIC_MASTER_SITES: - - For "fetch", move the assignment of the "getsite.sh" script to a more natural place, to get the full benefit of, and consistent handling for, setting MASTER_SORT and MASTER_SORT_REGEX - - For "fetch-list", embed the "getsite.sh" script into the generated script, so that the calculation can be properly done at run-time. ("fetch-list" also respects MASTER_SORT and MASTER_SORT_REGEX.)
2002-03-10 18:15:29 +01:00
.if defined(DYNAMIC_MASTER_SITES)
. for fetchfile in ${_ALLFILES}
Correctly fetch ALLFILES containing '=' characters
2002-09-29 03:38:32 +02:00
SITES_${fetchfile:T:S/=/--/}?= `${SH} ${FILESDIR}/getsite.sh ${fetchfile:T}`
Some tweaks to "fetch" and "fetch-list" re: DYNAMIC_MASTER_SITES: - - For "fetch", move the assignment of the "getsite.sh" script to a more natural place, to get the full benefit of, and consistent handling for, setting MASTER_SORT and MASTER_SORT_REGEX - - For "fetch-list", embed the "getsite.sh" script into the generated script, so that the calculation can be properly done at run-time. ("fetch-list" also respects MASTER_SORT and MASTER_SORT_REGEX.)
2002-03-10 18:15:29 +01:00
. endfor
.endif
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
.if !empty(_DISTFILES)
. for fetchfile in ${_DISTFILES}
Correctly fetch ALLFILES containing '=' characters
2002-09-29 03:38:32 +02:00
SITES_${fetchfile:T:S/=/--/}?= ${MASTER_SITES}
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
. endfor
.endif
.if !empty(_PATCHFILES)
. for fetchfile in ${_PATCHFILES}
Correctly fetch ALLFILES containing '=' characters
2002-09-29 03:38:32 +02:00
SITES_${fetchfile:T:S/=/--/}?= ${PATCH_SITES}
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
. endfor
.endif
In the do-fetch target, break out the check for distfiles if we're in batch mode, into a separate target. Call that target if we're in batch mode. Modify the check for distfiles code to do a loop for all the distfiles, and only print the message once at the end if the files aren't present, and BATCH is defined. Only check for distfiles' explicit presence if BATCH is defined.
2002-07-03 18:30:18 +02:00
# This code is only called in a batch case, to check for the presence of
# the distfiles
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: batch-check-distfiles
In the do-fetch target, break out the check for distfiles if we're in batch mode, into a separate target. Call that target if we're in batch mode. Modify the check for distfiles code to do a loop for all the distfiles, and only print the message once at the end if the files aren't present, and BATCH is defined. Only check for distfiles' explicit presence if BATCH is defined.
2002-07-03 18:30:18 +02:00
batch-check-distfiles:
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
In the do-fetch target, break out the check for distfiles if we're in batch mode, into a separate target. Call that target if we're in batch mode. Modify the check for distfiles code to do a loop for all the distfiles, and only print the message once at the end if the files aren't present, and BATCH is defined. Only check for distfiles' explicit presence if BATCH is defined.
2002-07-03 18:30:18 +02:00
gotfiles=yes; \
for file in "" ${_ALLFILES}; do \
case "$$file" in \
"") continue ;; \
*) if [ ! -f ${DISTDIR}/$$file ]; then \
gotfiles=no; \
fi ;; \
esac; \
done; \
case "$$gotfiles" in \
no) ${ECHO} "*** This package requires user intervention to download the distfiles"; \
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
${ECHO} "*** Please fetch the distfiles manually and place them in"; \
${ECHO} "*** ${DISTDIR}"; \
[ ! -z "${MASTER_SITES}" ] && \
${ECHO} "*** The distfiles are available from ${MASTER_SITES}"; \
[ ! -z "${HOMEPAGE}" ] && \
${ECHO} "*** See ${HOMEPAGE} for more details"; \
${ECHO}; \
Add a missing semicolon in the INTERACTIVE_STAGE=fetch bulk build logic.
2002-07-24 19:44:16 +02:00
${TOUCH} ${INTERACTIVE_COOKIE}; \
In the do-fetch target, break out the check for distfiles if we're in batch mode, into a separate target. Call that target if we're in batch mode. Modify the check for distfiles code to do a loop for all the distfiles, and only print the message once at the end if the files aren't present, and BATCH is defined. Only check for distfiles' explicit presence if BATCH is defined.
2002-07-03 18:30:18 +02:00
${FALSE} ;; \
esac
Add a check at fetch time to see if there are any known vulnerabilities in a package - should keep some admins' blood pressure a bit lower. Inspired by PR 19479. Output where there is a vulnerability looks like: [13:30:05] agc@sys1 ...pkgsrc/net/samba 28 > make fetch ===> do-fetch [samba-2.2.8] ===> Checking for vulnerabilities in samba-2.2.8 *** WARNING - remote-root-access vulnerability in samba-2.2.8 - see http://lists.samba.org/pipermail/samba-announce/2003-April/000065.html for more information *** *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba whilst output where there is no vulnerability looks like: [13:30:28] agc@sys1 ...pkgsrc/net/ethereal 30 > make fetch ===> do-fetch [ethereal-0.9.11] ===> Checking for vulnerabilities in ethereal-0.9.11 => ethereal-0.9.11.tar.gz doesn't seem to exist on this system. => Attempting to fetch ethereal-0.9.11.tar.gz from ftp://ftp.sunet.se/pub/network/monitoring/ethereal/. => [4457475 bytes] Trying 194.71.11.40... Connected to ftp.sunet.se. 220-####################################################################
2003-04-17 14:36:54 +02:00
# check for any vulnerabilities in the package
# Please do not modify the leading "@" here
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: check-vulnerable
Add a check at fetch time to see if there are any known vulnerabilities in a package - should keep some admins' blood pressure a bit lower. Inspired by PR 19479. Output where there is a vulnerability looks like: [13:30:05] agc@sys1 ...pkgsrc/net/samba 28 > make fetch ===> do-fetch [samba-2.2.8] ===> Checking for vulnerabilities in samba-2.2.8 *** WARNING - remote-root-access vulnerability in samba-2.2.8 - see http://lists.samba.org/pipermail/samba-announce/2003-April/000065.html for more information *** *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba whilst output where there is no vulnerability looks like: [13:30:28] agc@sys1 ...pkgsrc/net/ethereal 30 > make fetch ===> do-fetch [ethereal-0.9.11] ===> Checking for vulnerabilities in ethereal-0.9.11 => ethereal-0.9.11.tar.gz doesn't seem to exist on this system. => Attempting to fetch ethereal-0.9.11.tar.gz from ftp://ftp.sunet.se/pub/network/monitoring/ethereal/. => [4457475 bytes] Trying 194.71.11.40... Connected to ftp.sunet.se. 220-####################################################################
2003-04-17 14:36:54 +02:00
check-vulnerable:
Fix for PR 28230 (bsd.pkg.mk ignores /etc/audit-packages.conf) by sourcing any config file, if it's defined. With thanks to Julio M. Merino Vidal for the first pass at a fix, and for keeping me honest and up-to-date with variable checks in portable shell constructs.
2005-03-31 23:02:28 +02:00
@if [ ! -z "${PKG_SYSCONFDIR.audit-packages}" -a -f ${PKG_SYSCONFDIR.audit-packages}/audit-packages.conf ]; then \
. ${PKG_SYSCONFDIR.audit-packages}/audit-packages.conf; \
elif [ ! -z "${PKG_SYSCONFDIR}" -a -f ${PKG_SYSCONFDIR}/audit-packages.conf ]; then \
. ${PKG_SYSCONFDIR}/audit-packages.conf; \
fi; \
Fix error introduced in rev. 1.1606: "$${PKGVULNDIR}" should be "${PKGVULNDIR}" (one $, not two). This was causing the vulnerabilities file to be looked for in the wrong place. This fix makes the check-vulnerable target actually parse the vulnerabilities file and look for vulnerable installed packages again.
2005-04-13 16:07:30 +02:00
if [ -f ${PKGVULNDIR}/pkg-vulnerabilities ]; then \
Add a check at fetch time to see if there are any known vulnerabilities in a package - should keep some admins' blood pressure a bit lower. Inspired by PR 19479. Output where there is a vulnerability looks like: [13:30:05] agc@sys1 ...pkgsrc/net/samba 28 > make fetch ===> do-fetch [samba-2.2.8] ===> Checking for vulnerabilities in samba-2.2.8 *** WARNING - remote-root-access vulnerability in samba-2.2.8 - see http://lists.samba.org/pipermail/samba-announce/2003-April/000065.html for more information *** *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba whilst output where there is no vulnerability looks like: [13:30:28] agc@sys1 ...pkgsrc/net/ethereal 30 > make fetch ===> do-fetch [ethereal-0.9.11] ===> Checking for vulnerabilities in ethereal-0.9.11 => ethereal-0.9.11.tar.gz doesn't seem to exist on this system. => Attempting to fetch ethereal-0.9.11.tar.gz from ftp://ftp.sunet.se/pub/network/monitoring/ethereal/. => [4457475 bytes] Trying 194.71.11.40... Connected to ftp.sunet.se. 220-####################################################################
2003-04-17 14:36:54 +02:00
${SETENV} PKGNAME="${PKGNAME}" \
Speed-ups for the check-vulnerable target, from Jeremy Reed in PR 21393.
2003-04-29 22:31:52 +02:00
PKGBASE="${PKGBASE}" \
${AWK} '/^$$/ { next } \
/^#.*/ { next } \
Put back the speed-up in the vulnerability checking when building a package, but disable this if the package name in the pkg-vulnerabilities file contains a meta-character. This speeds up the check-vulnerable target from: [9:28:06] agc@sys3 ...pkgsrc/misc/libutf 11 > time make check-vulnerable 1.821u 1.988s 0:02.57 147.8% 0+0k 0+0io 0pf+0w [9:28:17] agc@sys3 ...pkgsrc/misc/libutf 12 > to [9:28:27] agc@sys3 ...pkgsrc/misc/libutf 13 > time make check-vulnerable 0.273u 0.233s 0:00.33 151.5% 0+0k 0+0io 0pf+0w [9:28:30] agc@sys3 ...pkgsrc/misc/libutf 14 > (with caches filled, on a fairly fast machine - P4 2.8 GHz, 2GB RAM).
2005-03-13 10:30:31 +01:00
$$1 !~ ENVIRON["PKGBASE"] && $$1 !~ /\{/ { next } \
The package vulnerability list is now called "pkg-vulnerabilities".
2003-09-03 17:05:43 +02:00
{ s = sprintf("${PKG_ADMIN} pmatch \"%s\" %s && ${ECHO} \"*** WARNING - %s vulnerability in %s - see %s for more information ***\"", $$1, ENVIRON["PKGNAME"], $$2, ENVIRON["PKGNAME"], $$3); system(s); }' < ${PKGVULNDIR}/pkg-vulnerabilities || ${FALSE}; \
Add a check at fetch time to see if there are any known vulnerabilities in a package - should keep some admins' blood pressure a bit lower. Inspired by PR 19479. Output where there is a vulnerability looks like: [13:30:05] agc@sys1 ...pkgsrc/net/samba 28 > make fetch ===> do-fetch [samba-2.2.8] ===> Checking for vulnerabilities in samba-2.2.8 *** WARNING - remote-root-access vulnerability in samba-2.2.8 - see http://lists.samba.org/pipermail/samba-announce/2003-April/000065.html for more information *** *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba whilst output where there is no vulnerability looks like: [13:30:28] agc@sys1 ...pkgsrc/net/ethereal 30 > make fetch ===> do-fetch [ethereal-0.9.11] ===> Checking for vulnerabilities in ethereal-0.9.11 => ethereal-0.9.11.tar.gz doesn't seem to exist on this system. => Attempting to fetch ethereal-0.9.11.tar.gz from ftp://ftp.sunet.se/pub/network/monitoring/ethereal/. => [4457475 bytes] Trying 194.71.11.40... Connected to ftp.sunet.se. 220-####################################################################
2003-04-17 14:36:54 +02:00
fi
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-fetch
In the do-fetch target, break out the check for distfiles if we're in batch mode, into a separate target. Call that target if we're in batch mode. Modify the check for distfiles code to do a loop for all the distfiles, and only print the message once at the end if the files aren't present, and BATCH is defined. Only check for distfiles' explicit presence if BATCH is defined.
2002-07-03 18:30:18 +02:00
.if !target(do-fetch)
do-fetch:
Add an ALLOW_VULNERABLE_PACKAGES switch, defaulting to undefined, to allow vulnerable packages to be built (in extremis, of course). Prompted by Thomas Klausner.
2003-04-17 15:04:56 +02:00
. if !defined(ALLOW_VULNERABLE_PACKAGES)
Add a check at fetch time to see if there are any known vulnerabilities in a package - should keep some admins' blood pressure a bit lower. Inspired by PR 19479. Output where there is a vulnerability looks like: [13:30:05] agc@sys1 ...pkgsrc/net/samba 28 > make fetch ===> do-fetch [samba-2.2.8] ===> Checking for vulnerabilities in samba-2.2.8 *** WARNING - remote-root-access vulnerability in samba-2.2.8 - see http://lists.samba.org/pipermail/samba-announce/2003-April/000065.html for more information *** *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba *** Error code 1 Stop. make: stopped in /usr/pkgsrc/net/samba whilst output where there is no vulnerability looks like: [13:30:28] agc@sys1 ...pkgsrc/net/ethereal 30 > make fetch ===> do-fetch [ethereal-0.9.11] ===> Checking for vulnerabilities in ethereal-0.9.11 => ethereal-0.9.11.tar.gz doesn't seem to exist on this system. => Attempting to fetch ethereal-0.9.11.tar.gz from ftp://ftp.sunet.se/pub/network/monitoring/ethereal/. => [4457475 bytes] Trying 194.71.11.40... Connected to ftp.sunet.se. 220-####################################################################
2003-04-17 14:36:54 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
if [ -f ${PKGVULNDIR}/pkg-vulnerabilities ]; then \
Check for the existence of the vulnerabilities file before telling people that we're checking for vulnerabilities. If the vulnerabilities file can't be found, display a warning. Addresses PR 21572 from Robert Elz.
2003-05-18 22:52:28 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> Checking for vulnerabilities in ${PKGNAME}"; \
vul=`${MAKE} ${MAKEFLAGS} check-vulnerable`; \
case "$$vul" in \
"") ;; \
*) ${ECHO} "$$vul"; \
${ECHO} "or define ALLOW_VULNERABLE_PACKAGES if this package is absolutely essential"; \
${FALSE} ;; \
esac; \
else \
The package vulnerability list is now called "pkg-vulnerabilities".
2003-09-03 17:05:43 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> *** No ${PKGVULNDIR}/pkg-vulnerabilities file found,"; \
Improve message when vulnerability file is not there.
2003-06-02 01:56:50 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> *** skipping vulnerability checks. To fix, install"; \
more wording tweak for missing vulnerability file.
2003-06-02 01:59:07 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> *** the pkgsrc/security/audit-packages package and run"; \
${ECHO_MSG} "${_PKGSRC_IN}> *** '${LOCALBASE}/sbin/download-vulnerability-list'."; \
Check for the existence of the vulnerabilities file before telling people that we're checking for vulnerabilities. If the vulnerabilities file can't be found, display a warning. Addresses PR 21572 from Robert Elz.
2003-05-18 22:52:28 +02:00
fi
Add an ALLOW_VULNERABLE_PACKAGES switch, defaulting to undefined, to allow vulnerable packages to be built (in extremis, of course). Prompted by Thomas Klausner.
2003-04-17 15:04:56 +02:00
. endif
In the do-fetch target, break out the check for distfiles if we're in batch mode, into a separate target. Call that target if we're in batch mode. Modify the check for distfiles code to do a loop for all the distfiles, and only print the message once at the end if the files aren't present, and BATCH is defined. Only check for distfiles' explicit presence if BATCH is defined.
2002-07-03 18:30:18 +02:00
. if !empty(_ALLFILES)
${_PKG_SILENT}${_PKG_DEBUG} \
${TEST} -d ${_DISTDIR} || ${MKDIR} ${_DISTDIR}
. if ${INTERACTIVE_STAGE:Mfetch} == "fetch" && defined(BATCH)
${_PKG_SILENT}${_PKG_DEBUG} \
${MAKE} ${MAKEFLAGS} batch-check-distfiles
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
. else
. for fetchfile in ${_ALLFILES}
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
. if defined(_FETCH_MESSAGE)
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Implement DYNAMIC_MASTER_SITES If you need to download from a dynamic URL you can set DYNAMIC_MASTER_SITES and a 'make fetch' will call files/getsite.sh with the name of each file to download as an argument, expecting it to output the URL of the directory from which to download it. graphics/ns-cult3d is an example of this usage.
2002-02-13 21:05:02 +01:00
file="${fetchfile}"; \
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
if [ ! -f ${DISTDIR}/$$file ]; then \
${_FETCH_MESSAGE}; \
fi
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
. else
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
cd ${_DISTDIR}; \
Add support for distfile-specific master sites, as requested in pkg/7471. Syntax: MASTER_SITES_completefilename= http://specific.master/site and similarly for PATCH_SITES. Convert print/ghostscript-nox11 and x11/kterm to take advantage of this.
2001-12-02 22:29:20 +01:00
file="${fetchfile}"; \
bfile="${fetchfile:T}"; \
Correctly fetch ALLFILES containing '=' characters
2002-09-29 03:38:32 +02:00
unsorted_sites="${SITES_${fetchfile:T:S/=/--/}} ${_MASTER_SITE_BACKUP}"; \
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
sites="${ORDERED_SITES}"; \
Add support for distfile-specific master sites, as requested in pkg/7471. Syntax: MASTER_SITES_completefilename= http://specific.master/site and similarly for PATCH_SITES. Convert print/ghostscript-nox11 and x11/kterm to take advantage of this.
2001-12-02 22:29:20 +01:00
${_CHECK_DIST_PATH}; \
Make the option to resume transfer optional, as suggested by Robert Elz in PR pkg/26448 (at least for now). If you want to use it, PKG_RESUME_TRANSFERS must be set to [Yy][Ee][Ss]. While here, no need to use ${AWK} to get the size with ${WC} -c, ${WC} -c < file is enough. Thanks to Robert Elz for suggestions, still I have to fix the problems reported in the PR.
2004-07-27 17:56:19 +02:00
if ${TEST} "${PKG_RESUME_TRANSFERS:M[Yy][Ee][Ss]}" ; then \
${_FETCH_FILE}; ${_RESUME_TRANSFER}; \
else \
${_FETCH_FILE}; \
fi
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
. endif # defined(_FETCH_MESSAGE)
. endfor
. endif # INTERACTIVE_STAGE == fetch
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
. endif # !empty(_ALLFILES)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
add 'show-depends-dir' target. This shows the directories of the "top level" depends for the package. Both build and run depends are shown. This is a non-recursive target. Ie, only explicitly depends are shown. Example: bondage 109 # cd /usr/pkgsrc/cad/geda && make show-depends-dirs cad/gschem cad/gnetlist cad/gsymcheck cad/geda-utils cad/geda-docs This target is useful for collecting dependency tree information for bulk builds.
2001-01-22 05:00:54 +01:00
# show both build and run depends directories (non-recursively)
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-depends-dirs
add 'show-depends-dir' target. This shows the directories of the "top level" depends for the package. Both build and run depends are shown. This is a non-recursive target. Ie, only explicitly depends are shown. Example: bondage 109 # cd /usr/pkgsrc/cad/geda && make show-depends-dirs cad/gschem cad/gnetlist cad/gsymcheck cad/geda-utils cad/geda-docs This target is useful for collecting dependency tree information for bulk builds.
2001-01-22 05:00:54 +01:00
.if !target(show-depends-dirs)
show-depends-dirs:
Don't echo the debug message for the targets whose output is used by other targets. OK'd by jlam. This should close pkg/24377.
2004-10-03 06:24:20 +02:00
@dlist=""; \
Make BUILD_DIR always point to a physical path by invoking /bin/pwd instead of relying on the shell's builtin pwd. This makes BUILD_DIR consistently point to the correct directory regardless of the definition of SHELL. This fixes elusive some buildlink2 errors due to the fact that BUILDLINK_DIR is derived from BUILD_DIR and some paths were incorrectly being translated.
2003-01-15 21:55:38 +01:00
thisdir=`${PWD_CMD}`; \
Fix problem with Solaris' /bin/sh not accepting an empty list in a for loop noted in pkg/13894 by Stoned Elipot <Stoned.Elipot@script.jussieu.fr>. The suggested fix was applied.
2001-09-14 20:54:43 +02:00
for reldir in "" ${DEPENDS:C/^[^:]*://:C/:.*$//} ${BUILD_DEPENDS:C/^[^:]*://:C/:.*$//} ;\
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
do \
if [ "X$$reldir" = "X" ]; then continue; fi; \
cd $$thisdir/$$reldir; \
Make BUILD_DIR always point to a physical path by invoking /bin/pwd instead of relying on the shell's builtin pwd. This makes BUILD_DIR consistently point to the correct directory regardless of the definition of SHELL. This fixes elusive some buildlink2 errors due to the fact that BUILDLINK_DIR is derived from BUILD_DIR and some paths were incorrectly being translated.
2003-01-15 21:55:38 +01:00
WD=`${PWD_CMD}`; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
d=`dirname $$WD`; \
absdir=`basename $$d`/`basename $$WD`; \
dlist="$$dlist $$absdir"; \
done; \
cd $$thisdir; \
add 'show-depends-dir' target. This shows the directories of the "top level" depends for the package. Both build and run depends are shown. This is a non-recursive target. Ie, only explicitly depends are shown. Example: bondage 109 # cd /usr/pkgsrc/cad/geda && make show-depends-dirs cad/gschem cad/gnetlist cad/gsymcheck cad/geda-utils cad/geda-docs This target is useful for collecting dependency tree information for bulk builds.
2001-01-22 05:00:54 +01:00
${ECHO} "$$dlist"
.endif
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
# Show all build and run depends, reverse-breadth first, with options.
.if make(show-all-depends-dirs) || make(show-all-depends-dirs-excl) || make (show-root-dirs)
# "awk" macro to recurse over the dependencies efficiently, never running in
# the same same directory twice. You may set the following options via "-v":
#
# NonSelf = 1 to not print own directory;
# RootsOnly = 1 to print only root directories (i.e. directories
# of packages with no dependencies), including possibly
# own directory
#
_RECURSE_DEPENDS_DIRS= \
function append_dirs(dir) { \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
command = "cd ../../" dir " && ${MAKE} show-depends-dirs"; \
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
command | getline tmp_dirs; \
close(command); \
if (tmp_dirs ~ /^$$/) \
root_dirs[p++] = dir; \
for (i = 1; i <= split(tmp_dirs, tmp_r); i++) \
if (!(tmp_r[i] in hash_all_dirs)) { \
all_dirs[n++] = tmp_r[i]; \
hash_all_dirs[tmp_r[i]] = 1 \
} \
} \
BEGIN { \
Make BUILD_DIR always point to a physical path by invoking /bin/pwd instead of relying on the shell's builtin pwd. This makes BUILD_DIR consistently point to the correct directory regardless of the definition of SHELL. This fixes elusive some buildlink2 errors due to the fact that BUILDLINK_DIR is derived from BUILD_DIR and some paths were incorrectly being translated.
2003-01-15 21:55:38 +01:00
command = "${PWD_CMD}"; \
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
command | getline start_dir; \
close(command); \
i = split(start_dir, tmp_r, /\//); \
all_dirs[n++] = tmp_r[i-1] "/" tmp_r[i]; \
for (; m < n; ) \
append_dirs(all_dirs[m++]); \
if (RootsOnly) { \
printf("%s", root_dirs[--p]); \
for (; p > 0; ) \
printf(" %s", root_dirs[--p]) \
} \
else { \
if (m > NonSelf) \
printf("%s", all_dirs[--m]); \
for (; m > NonSelf; ) \
printf(" %s", all_dirs[--m]) \
} \
print \
}
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-all-depends-dirs
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.if make(show-all-depends-dirs)
show-all-depends-dirs:
Don't echo the debug message for the targets whose output is used by other targets. OK'd by jlam. This should close pkg/24377.
2004-10-03 06:24:20 +02:00
@${AWK} '${_RECURSE_DEPENDS_DIRS}'
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-all-depends-dirs-excl
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.if make(show-all-depends-dirs-excl)
show-all-depends-dirs-excl:
Don't echo the debug message for the targets whose output is used by other targets. OK'd by jlam. This should close pkg/24377.
2004-10-03 06:24:20 +02:00
@${AWK} -v NonSelf=1 '${_RECURSE_DEPENDS_DIRS}'
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-root-dirs
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.if make(show-root-dirs)
show-root-dirs:
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${AWK} -v RootsOnly=1 '${_RECURSE_DEPENDS_DIRS}'
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.endif
.endif # make(show-{all-depends-dirs{,-excl},root-dirs})
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-distfiles
Add a "show-distfiles" target, used to show the distfiles and patchfiles (including any subdir) for a package. If ${IGNORE} is set, for any reason, do not print anything. The output is relative to ${DISTDIR}. This target can be used to identify out of date distfiles and patchfiles in ${DISTDIR}, in conjunction with suitable filters.
1999-06-28 13:42:28 +02:00
.if !target(show-distfiles)
show-distfiles:
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
. if defined(PKG_FAIL_REASON)
Add a "show-distfiles" target, used to show the distfiles and patchfiles (including any subdir) for a package. If ${IGNORE} is set, for any reason, do not print anything. The output is relative to ${DISTDIR}. This target can be used to identify out of date distfiles and patchfiles in ${DISTDIR}, in conjunction with suitable filters.
1999-06-28 13:42:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Add a "show-distfiles" target, used to show the distfiles and patchfiles (including any subdir) for a package. If ${IGNORE} is set, for any reason, do not print anything. The output is relative to ${DISTDIR}. This target can be used to identify out of date distfiles and patchfiles in ${DISTDIR}, in conjunction with suitable filters.
1999-06-28 13:42:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
for file in "" ${_CKSUMFILES}; do \
if [ "X$$file" = "X" ]; then continue; fi; \
${ECHO} $$file; \
done
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Add a "show-distfiles" target, used to show the distfiles and patchfiles (including any subdir) for a package. If ${IGNORE} is set, for any reason, do not print anything. The output is relative to ${DISTDIR}. This target can be used to identify out of date distfiles and patchfiles in ${DISTDIR}, in conjunction with suitable filters.
1999-06-28 13:42:28 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-downlevel
Add a show-downlevel target, which shows output of the form: sudo package: sudo-1.5.7p4 installed, pkgsrc version sudo-1.5.9p1 for installed packages whose version does not match that in pkgsrc. No output is produced if the package is not installed. This allows an easy overview of the packages installed on a machine which need to be upgraded.
1999-07-02 14:54:28 +02:00
.if !target(show-downlevel)
show-downlevel:
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
. if defined(PKG_FAIL_REASON)
Add a show-downlevel target, which shows output of the form: sudo package: sudo-1.5.7p4 installed, pkgsrc version sudo-1.5.9p1 for installed packages whose version does not match that in pkgsrc. No output is produced if the package is not installed. This allows an easy overview of the packages installed on a machine which need to be upgraded.
1999-07-02 14:54:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Add a show-downlevel target, which shows output of the form: sudo package: sudo-1.5.7p4 installed, pkgsrc version sudo-1.5.9p1 for installed packages whose version does not match that in pkgsrc. No output is produced if the package is not installed. This allows an easy overview of the packages installed on a machine which need to be upgraded.
1999-07-02 14:54:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Fix cases where we were calling "${PKG_BEST_EXISTS} <pkgbase>". <pkgbase> isn't actually a pkgpattern; what we really want is <pkgwildcard>. This should fix breakage with the "update" and "replace" targets introduced in revisions 1.1275 and 1.1278.
2003-09-13 20:26:15 +02:00
found="`${PKG_BEST_EXISTS} \"${PKGWILDCARD}\" || ${TRUE}`"; \
Add a show-downlevel target, which shows output of the form: sudo package: sudo-1.5.7p4 installed, pkgsrc version sudo-1.5.9p1 for installed packages whose version does not match that in pkgsrc. No output is produced if the package is not installed. This allows an easy overview of the packages installed on a machine which need to be upgraded.
1999-07-02 14:54:28 +02:00
if [ "X$$found" != "X" -a "X$$found" != "X${PKGNAME}" ]; then \
Add PKGPATH, PKGBASE, PKGVERSION and PKGWILDCARD make(1) variables, so that the substitution hacks are all done in one place. Protect a few things with double quotes just to be safe. Fix `readme' target that was left broken in last commit.
2000-06-03 20:51:34 +02:00
${ECHO} "${PKGBASE} package: $$found installed, pkgsrc version ${PKGNAME}"; \
Add support for two environment variables to help update all your installed packages and only rebuild everything once: STOP_DOWNLEVEL_AFTER_FIRST if set makes "make show-downlevel" stop a bit earlier (since with the other one below you will be only interested in the first downlevel pkg found). REBUILD_DOWNLEVEL_DEPENDS if set forces the pattern match for dependencies to fail whenever the installed pkg is not the version in the makefile (i.e. "make show-downlevel" would print a version mismatch for the dependecy). This causes all downlevel dependencies (and everything depending on them) to be rebuild. Enhancements, like making this all work with make command line flags and settings in /etc/mk.conf are welcome. A way to stop "make show-downlevel" through all upper levels of recursion imediately would be very usefull too.
2001-09-10 22:03:17 +02:00
if [ "X$$STOP_DOWNLEVEL_AFTER_FIRST" != "X" ]; then \
TYpoe: 'stoping' -> 'stopping'
2002-07-22 04:33:55 +02:00
${ECHO} "stopping after first downlevel pkg found"; \
Add support for two environment variables to help update all your installed packages and only rebuild everything once: STOP_DOWNLEVEL_AFTER_FIRST if set makes "make show-downlevel" stop a bit earlier (since with the other one below you will be only interested in the first downlevel pkg found). REBUILD_DOWNLEVEL_DEPENDS if set forces the pattern match for dependencies to fail whenever the installed pkg is not the version in the makefile (i.e. "make show-downlevel" would print a version mismatch for the dependecy). This causes all downlevel dependencies (and everything depending on them) to be rebuild. Enhancements, like making this all work with make command line flags and settings in /etc/mk.conf are welcome. A way to stop "make show-downlevel" through all upper levels of recursion imediately would be very usefull too.
2001-09-10 22:03:17 +02:00
exit 1; \
fi; \
Add a show-downlevel target, which shows output of the form: sudo package: sudo-1.5.7p4 installed, pkgsrc version sudo-1.5.9p1 for installed packages whose version does not match that in pkgsrc. No output is produced if the package is not installed. This allows an easy overview of the packages installed on a machine which need to be upgraded.
1999-07-02 14:54:28 +02:00
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Add a show-downlevel target, which shows output of the form: sudo package: sudo-1.5.7p4 installed, pkgsrc version sudo-1.5.9p1 for installed packages whose version does not match that in pkgsrc. No output is produced if the package is not installed. This allows an easy overview of the packages installed on a machine which need to be upgraded.
1999-07-02 14:54:28 +02:00
.endif
Define a new target, "show-pkgsrc-dir", which prints the directory from which an installed package can be re-installed. This can be used to build up a list of host specific packages, which is useful, for example, in re-building all packages on a machine for a.out to ELF transition.
2000-05-11 13:23:20 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-installed-depends
Add 'show-installed-depends' - neat implementation thanks to Hubert.
2001-10-01 00:10:33 +02:00
.if !target(show-installed-depends)
show-installed-depends:
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
. if !empty(DEPENDS)
Use the usual _PKG_SILENT and _PKG_DEBUG definitions in the show-installed-depends target.
2001-10-05 00:01:54 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
for i in ${DEPENDS:C/:.*$//:Q:S/\ / /g} ; do \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
echo "$$i =>" `${PKG_BEST_EXISTS} "$$i"` ; \
Add 'show-installed-depends' - neat implementation thanks to Hubert.
2001-10-01 00:10:33 +02:00
done
use ${PKG_INFO} rather than pkg_info in show-installed-depends target, so that the correct setting of PKG_DBDIR is used, and the correct pkg_info binary too, presumably. Make show-installed-depends work on Solaris (untested) by only defining the target if DEPENDS is defined.
2001-10-04 23:48:12 +02:00
. endif
Add 'show-installed-depends' - neat implementation thanks to Hubert.
2001-10-01 00:10:33 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-needs-update
Add a ``show-needs-update'' target that checks each dependency to see if it's up to date. For dependencies (including the current package) that are not up to date (or are not installed), a one line note is printed thusly: % cd audio/xmradio/ % make show-needs-update audio/mad => mad-0.14.2b => needs update to mad-0.14.2bnb1 audio/xmradio => (none) => needs install of xmradio-1.2
2003-06-14 20:52:46 +02:00
.if !target(show-needs-update)
show-needs-update:
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
. if !empty(DEPENDS)
Add a ``show-needs-update'' target that checks each dependency to see if it's up to date. For dependencies (including the current package) that are not up to date (or are not installed), a one line note is printed thusly: % cd audio/xmradio/ % make show-needs-update audio/mad => mad-0.14.2b => needs update to mad-0.14.2bnb1 audio/xmradio => (none) => needs install of xmradio-1.2
2003-06-14 20:52:46 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
for i in `${MAKE} show-all-depends-dirs`; do \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
cd ${PKGSRCDIR}/$$i; \
show-needs-update target: run ${MAKE}, not "make".
2004-01-25 17:12:40 +01:00
want=`${MAKE} show-var VARNAME=PKGNAME`; \
wild=`${MAKE} show-var VARNAME=PKGWILDCARD`; \
Fix cases where we were calling "${PKG_BEST_EXISTS} <pkgbase>". <pkgbase> isn't actually a pkgpattern; what we really want is <pkgwildcard>. This should fix breakage with the "update" and "replace" targets introduced in revisions 1.1275 and 1.1278.
2003-09-13 20:26:15 +02:00
have=`${PKG_BEST_EXISTS} "$$wild" || ${TRUE}`; \
Add a ``show-needs-update'' target that checks each dependency to see if it's up to date. For dependencies (including the current package) that are not up to date (or are not installed), a one line note is printed thusly: % cd audio/xmradio/ % make show-needs-update audio/mad => mad-0.14.2b => needs update to mad-0.14.2bnb1 audio/xmradio => (none) => needs install of xmradio-1.2
2003-06-14 20:52:46 +02:00
if [ -z "$$have" ]; then \
Fix cases where we were calling "${PKG_BEST_EXISTS} <pkgbase>". <pkgbase> isn't actually a pkgpattern; what we really want is <pkgwildcard>. This should fix breakage with the "update" and "replace" targets introduced in revisions 1.1275 and 1.1278.
2003-09-13 20:26:15 +02:00
${ECHO} "$$i => (none) => needs install of $$want"; \
Add a ``show-needs-update'' target that checks each dependency to see if it's up to date. For dependencies (including the current package) that are not up to date (or are not installed), a one line note is printed thusly: % cd audio/xmradio/ % make show-needs-update audio/mad => mad-0.14.2b => needs update to mad-0.14.2bnb1 audio/xmradio => (none) => needs install of xmradio-1.2
2003-06-14 20:52:46 +02:00
elif [ "$$have" != "$$want" ]; then \
Fix cases where we were calling "${PKG_BEST_EXISTS} <pkgbase>". <pkgbase> isn't actually a pkgpattern; what we really want is <pkgwildcard>. This should fix breakage with the "update" and "replace" targets introduced in revisions 1.1275 and 1.1278.
2003-09-13 20:26:15 +02:00
${ECHO} "$$i => $$have => needs update to $$want"; \
Add a ``show-needs-update'' target that checks each dependency to see if it's up to date. For dependencies (including the current package) that are not up to date (or are not installed), a one line note is printed thusly: % cd audio/xmradio/ % make show-needs-update audio/mad => mad-0.14.2b => needs update to mad-0.14.2bnb1 audio/xmradio => (none) => needs install of xmradio-1.2
2003-06-14 20:52:46 +02:00
fi; \
done
. endif
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-pkgsrc-dir
Define a new target, "show-pkgsrc-dir", which prints the directory from which an installed package can be re-installed. This can be used to build up a list of host specific packages, which is useful, for example, in re-building all packages on a machine for a.out to ELF transition.
2000-05-11 13:23:20 +02:00
.if !target(show-pkgsrc-dir)
show-pkgsrc-dir:
Replace IGNORE into PKG_FAIL_REASON and PKG_SKIP_REASON, so pkgsrc can stop builds when a dependency is broken, yet continue builds when a dependency is merely skipped (usually because it duplicates functionality in the base system). Thus IGNORE_FAIL, the oft-misunderstood and naught-documented option, goes away. This addresses many people's complaints on tech-pkg@ and other lists, as well as PR pkg/18157.
2002-12-07 03:37:56 +01:00
. if defined(PKG_FAIL_REASON)
Define a new target, "show-pkgsrc-dir", which prints the directory from which an installed package can be re-installed. This can be used to build up a list of host specific packages, which is useful, for example, in re-building all packages on a machine for a.out to ELF transition.
2000-05-11 13:23:20 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Define a new target, "show-pkgsrc-dir", which prints the directory from which an installed package can be re-installed. This can be used to build up a list of host specific packages, which is useful, for example, in re-building all packages on a machine for a.out to ELF transition.
2000-05-11 13:23:20 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
found="`${PKG_BEST_EXISTS} \"${PKGWILDCARD}\" || ${TRUE}`"; \
Define a new target, "show-pkgsrc-dir", which prints the directory from which an installed package can be re-installed. This can be used to build up a list of host specific packages, which is useful, for example, in re-building all packages on a machine for a.out to ELF transition.
2000-05-11 13:23:20 +02:00
if [ "X$$found" != "X" ]; then \
Add PKGPATH, PKGBASE, PKGVERSION and PKGWILDCARD make(1) variables, so that the substitution hacks are all done in one place. Protect a few things with double quotes just to be safe. Fix `readme' target that was left broken in last commit.
2000-06-03 20:51:34 +02:00
${ECHO} ${PKGPATH}; \
Define a new target, "show-pkgsrc-dir", which prints the directory from which an installed package can be re-installed. This can be used to build up a list of host specific packages, which is useful, for example, in re-building all packages on a machine for a.out to ELF transition.
2000-05-11 13:23:20 +02:00
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Define a new target, "show-pkgsrc-dir", which prints the directory from which an installed package can be re-installed. This can be used to build up a list of host specific packages, which is useful, for example, in re-building all packages on a machine for a.out to ELF transition.
2000-05-11 13:23:20 +02:00
.endif
Add a show-downlevel target, which shows output of the form: sudo package: sudo-1.5.7p4 installed, pkgsrc version sudo-1.5.9p1 for installed packages whose version does not match that in pkgsrc. No output is produced if the package is not installed. This allows an easy overview of the packages installed on a machine which need to be upgraded.
1999-07-02 14:54:28 +02:00
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
# Extract
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
# pkgsrc coarse-grained locking definitions and targets
.if ${PKGSRC_LOCKTYPE} == "none"
_ACQUIRE_LOCK= ${_PKG_SILENT}${_PKG_DEBUG}${DO_NADA}
_RELEASE_LOCK= ${_PKG_SILENT}${_PKG_DEBUG}${DO_NADA}
.else
LOCKFILE= ${WRKDIR}/.lockfile
_ACQUIRE_LOCK= \
${_PKG_SILENT}${_PKG_DEBUG} \
When setting up the lockfile, don't store whitespace printed by ps(1), so that % make ===> do-fetch [digest-20021220] ===> Checking for vulnerabilities in digest-20021220 => Lock acquired on behalf of process 356 => Lock released on behalf of process 356 becomes % make ===> do-fetch [digest-20021220] ===> Checking for vulnerabilities in digest-20021220 => Lock acquired on behalf of process 814 => Lock released on behalf of process 814
2003-06-14 18:53:00 +02:00
ppid=`${PS} -p $$$$ -o ppid | ${AWK} 'NR == 2 { print $$1 }'`; \
catch locking failure case where we couldn't find the parent's process ID. from Roland Illig.
2004-07-04 00:11:56 +02:00
if ${TEST} "$$ppid" = ""; then \
${ECHO} "No parent process ID found."; \
${FALSE}; \
fi; \
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
while true; do \
Don't use the -nt operator for ${TEST}. It is not defined in POSIX, and not provided by the Solaris 9 /bin/sh. Use the find(1) -newer expression instead, as it expresses exactly what we want, compared to the "ls -t" hack.
2005-02-12 01:27:41 +01:00
: "Remove lock files older than the last reboot"; \
if ${TEST} -f /var/run/dmesg.boot -a -f ${LOCKFILE}; then \
Some older ${FIND}s do not have an implicit "-print"; make sure to specify it explicitly (as with other ${FIND} usage here).
2005-02-15 17:21:41 +01:00
rebooted=`${FIND} /var/run/dmesg.boot -newer ${LOCKFILE} -print`; \
Corrected a misspelling of a variable. I used $rebooted in a shell code, which had better been $$rebooted (Makefile quoting). Approved by jlam.
2005-02-13 04:49:12 +01:00
if ${TEST} x"$$rebooted" != x; then \
Don't use the -nt operator for ${TEST}. It is not defined in POSIX, and not provided by the Solaris 9 /bin/sh. Use the find(1) -newer expression instead, as it expresses exactly what we want, compared to the "ls -t" hack.
2005-02-12 01:27:41 +01:00
${ECHO} "=> Removing stale ${LOCKFILE}"; \
${RM} ${LOCKFILE}; \
fi; \
Check that the lockfile isn't stale from being left over across a system reboot by testing whether or not /var/run/dmesg.boot is newer than it or not. Fix suggested by agc@NetBSD.org and simonb@NetBSD.org.
2003-09-09 13:00:30 +02:00
fi; \
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
${SHLOCK} -f ${LOCKFILE} -p $$ppid && break; \
${ECHO} "=> Lock is held by pid `cat ${LOCKFILE}`"; \
case "${PKGSRC_LOCKTYPE}" in \
once) exit 1 ;; \
sleep) sleep ${PKGSRC_SLEEPSECS} ;; \
esac \
done; \
Wrap "Lock acquired ..." and "Lock released ..." messages with PKG_VERBOSE check
2003-02-18 11:36:44 +01:00
if [ "${PKG_VERBOSE}" != "" ]; then \
${ECHO_MSG} "=> Lock acquired on behalf of process $$ppid"; \
fi
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
_RELEASE_LOCK= \
${_PKG_SILENT}${_PKG_DEBUG} \
Wrap "Lock acquired ..." and "Lock released ..." messages with PKG_VERBOSE check
2003-02-18 11:36:44 +01:00
if [ "${PKG_VERBOSE}" != "" ]; then \
${ECHO_MSG} "=> Lock released on behalf of process `${CAT} ${LOCKFILE}`"; \
fi; \
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
${RM} ${LOCKFILE}
.endif # PKGSRC_LOCKTYPE
Add a MIRROR_DISTFILE variable; defaults to "yes", and should be set to "no" in a package's Makefile when the authors have placed restrictions on the redistribution of the distfiles. (Thanks to Jason Thorpe for the idea). Add a `mirror' target, which will do a "make fetch" if the package's distfile is allowed to be redistributed. This is for use on sites which mirror package distfiles. (Thanks to Charles Hannum for the idea).
1998-01-22 11:20:48 +01:00
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
${WRKDIR}:
.if !defined(KEEP_WRKDIR)
. if ${PKGSRC_LOCKTYPE} == "sleep" || ${PKGSRC_LOCKTYPE} == "once"
. if !exists(${LOCKFILE})
${_PKG_SILENT}${_PKG_DEBUG}${RM} -rf ${WRKDIR}
. endif
. endif
.endif
${_PKG_SILENT}${_PKG_DEBUG}${MKDIR} ${WRKDIR}
Move the creation of ${PKG_DB_TMPDIR} to when we create ${WRKDIR}. This ensures that it's always created with the same user/group/mode as ${WRKDIR}, so if a non-root user thinks he has permissions to remove ${WRKDIR}, then that expectation will remain true for ${PKG_DB_TMPDIR}.
2005-01-27 19:32:20 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${MKDIR} ${PKG_DB_TMPDIR}
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
.ifdef WRKOBJDIR
. if ${PKGSRC_LOCKTYPE} == "sleep" || ${PKGSRC_LOCKTYPE} == "once"
. if !exists(${LOCKFILE})
${_PKG_SILENT}${_PKG_DEBUG} \
${RM} -f ${WRKDIR_BASENAME} || ${TRUE}
. endif
. endif
${_PKG_SILENT}${_PKG_DEBUG} \
if ${LN} -s ${WRKDIR} ${WRKDIR_BASENAME} 2>/dev/null; then \
${ECHO} "${WRKDIR_BASENAME} -> ${WRKDIR}"; \
fi
.endif # WRKOBJDIR
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Move the creation of ${PKG_DB_TMPDIR} to when we create ${WRKDIR}. This ensures that it's always created with the same user/group/mode as ${WRKDIR}, so if a non-root user thinks he has permissions to remove ${WRKDIR}, then that expectation will remain true for ${PKG_DB_TMPDIR}.
2005-01-27 19:32:20 +01:00
Fix spelling of "suffixes". From Todd Vierling in PR 24400.
2004-02-13 20:02:07 +01:00
_EXTRACT_SUFFIXES= .tar.gz .tgz .tar.bz2 .tbz .tar.Z .tar _tar.gz
_EXTRACT_SUFFIXES+= .shar.gz .shar.bz2 .shar.Z .shar
_EXTRACT_SUFFIXES+= .zip
_EXTRACT_SUFFIXES+= .lha .lzh
_EXTRACT_SUFFIXES+= .Z .bz2 .gz
_EXTRACT_SUFFIXES+= .zoo
Implement EXTRACT_CMD.bin
2004-07-06 13:28:55 +02:00
_EXTRACT_SUFFIXES+= .bin
Fix spelling of "suffixes". From Todd Vierling in PR 24400.
2004-02-13 20:02:07 +01:00
_EXTRACT_SUFFIXES+= .rar
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
# If the distfile has a tar.bz2 suffix, use bzcat in preference to gzcat,
# pulling in the "bzip2" package if necessary. [Note: this is only for
# the benefit of pre-1.5 NetBSD systems. "gzcat" on newer systems happily
# decodes bzip2.] Do likewise for ".zip" and ".lha" distfiles.
#
Make sure that the decompression/extraction programs needed are automatically DEPENDed if EXTRACT_SUFX is set to a known suffix but EXTRACT_ONLY is set to an empty value. This should fix the problems for audio/guspatches and textproc/iso8879 in the 2002-01-28 i386 bulk-build results.
2002-01-28 20:08:18 +01:00
.if !empty(EXTRACT_ONLY:M*.bz2) || !empty(EXTRACT_ONLY:M*.tbz) || \
!empty(EXTRACT_SUFX:M*.bz2) || !empty(EXTRACT_SUFX:M*.tbz)
Teach the new tools framework about USE_TOOLS+=bzcat.
2005-05-02 04:50:33 +02:00
. if !empty(_USE_NEW_TOOLS:M[yY][eE][sS])
USE_TOOLS+= bzcat
. elif exists(/usr/bin/bzcat)
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
BZCAT= /usr/bin/bzcat <
. else
BUILD_DEPENDS+= bzip2>=0.9.0b:../../archivers/bzip2
BZCAT= ${LOCALBASE}/bin/bzcat
. endif
.endif
Make sure that the decompression/extraction programs needed are automatically DEPENDed if EXTRACT_SUFX is set to a known suffix but EXTRACT_ONLY is set to an empty value. This should fix the problems for audio/guspatches and textproc/iso8879 in the 2002-01-28 i386 bulk-build results.
2002-01-28 20:08:18 +01:00
.if !empty(EXTRACT_ONLY:M*.zip) || !empty(EXTRACT_SUFX:M*.zip)
We've changed DEPENDS=foo-* into DEPENDS=foo-[0-9]* everywhere else but here, so correct that oversight.
2002-01-28 20:00:05 +01:00
BUILD_DEPENDS+= unzip-[0-9]*:../../archivers/unzip
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
.endif
Make sure that the decompression/extraction programs needed are automatically DEPENDed if EXTRACT_SUFX is set to a known suffix but EXTRACT_ONLY is set to an empty value. This should fix the problems for audio/guspatches and textproc/iso8879 in the 2002-01-28 i386 bulk-build results.
2002-01-28 20:08:18 +01:00
.if !empty(EXTRACT_ONLY:M*.lzh) || !empty(EXTRACT_ONLY:M*.lha) || \
!empty(EXTRACT_SUFX:M*.lzh) || !empty(EXTRACT_SUFX:M*.lha)
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
BUILD_DEPENDS+= lha>=114.9:../../archivers/lha
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
.endif
Re-order nested .if statements for brevity.
2005-05-02 05:09:04 +02:00
.if !empty(EXTRACT_ONLY:M*.gz) || !empty(EXTRACT_ONLY:M*.tgz) || \
!empty(EXTRACT_SUFX:M*.gz) || !empty(EXTRACT_SUFX:M*.tgz)
. if !empty(_USE_NEW_TOOLS:M[yY][eE][sS])
Split replace.mk into two parts, one of which is included by bsd.prefs.mk to provide "TOOL" definitions for tools used by a top-level make process (usually because it uses them in a != variable definition). This allows USE_TOOLS to be defined before bsd.prefs.mk is included by a package Makefile, where USE_TOOLS lists the additional (non-default) tools that are required to build the package. Also, drop the fallback to existing "TOOL" definitions because we now have TOOLS_PLATFORM.* for each platform in pkgsr/mk/tools/tools.*.mk.
2005-04-30 06:35:54 +02:00
USE_TOOLS+= gzcat
Re-order nested .if statements for brevity.
2005-05-02 05:09:04 +02:00
. elif !defined(GZCAT)
Bump gzip-base to 1.2.4b for security fix.
2004-10-12 20:29:25 +02:00
BUILD_DEPENDS+= gzip-base>=1.2.4b:../../archivers/gzip-base
on Solaris, depend on gzip-base if GZCAT is not set, and neither /usr/bin/gzcat nor ${LOCALBASE}/bin/gzip exist. required for Solaris systems which do not include gzcat in the base system. sanity checked by Stoned Elipot <seb@netbsd.org>.
2002-07-11 17:52:35 +02:00
GZCAT= ${LOCALBASE}/bin/zcat
. endif
.endif
Add an extract method for zoo archives
2002-12-23 02:59:43 +01:00
.if !empty(EXTRACT_ONLY:M*.zoo) || !empty(EXTRACT_SUFX:M*.zoo)
BUILD_DEPENDS+= unzoo-[0-9]*:../../archivers/unzoo
.endif
Enable EXTRACT_{ONLY,SUFX}= .rar. Ok'ed by jlam
2004-02-12 22:29:28 +01:00
.if !empty(EXTRACT_ONLY:M*.rar) || !empty(EXTRACT_SUFX:M*.rar)
BUILD_DEPENDS+= unrar>=3.3.4:../../archivers/unrar
.endif
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
DECOMPRESS_CMD.tar.gz?= ${GZCAT}
DECOMPRESS_CMD.tgz?= ${DECOMPRESS_CMD.tar.gz}
DECOMPRESS_CMD.tar.bz2?= ${BZCAT}
DECOMPRESS_CMD.tbz?= ${DECOMPRESS_CMD.tar.bz2}
Explicitly recognize .tar.Z files so that they don't fall through into .Z handling.
2002-02-25 06:07:45 +01:00
DECOMPRESS_CMD.tar.Z?= ${GZCAT}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
DECOMPRESS_CMD.tar?= ${CAT}
Recognize how to automatically extract .shar.gz, .shar.bz2, .shar.Z, and .shar distfiles.
2002-01-27 19:31:47 +01:00
DECOMPRESS_CMD.shar.gz?= ${GZCAT}
DECOMPRESS_CMD.shar.bz2?= ${BZCAT}
DECOMPRESS_CMD.shar.Z?= ${GZCAT}
DECOMPRESS_CMD.shar?= ${CAT}
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
DECOMPRESS_CMD.Z?= ${GZCAT}
DECOMPRESS_CMD.bz2?= ${BZCAT}
DECOMPRESS_CMD.gz?= ${GZCAT}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
DECOMPRESS_CMD?= ${GZCAT}
Fix spelling of "suffixes". From Todd Vierling in PR 24400.
2004-02-13 20:02:07 +01:00
.for __suffix__ in ${_EXTRACT_SUFFIXES}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
. if !defined(DECOMPRESS_CMD${__suffix__})
DECOMPRESS_CMD${__suffix__}?= ${DECOMPRESS_CMD}
. endif
.endfor
# If this is empty, then everything gets extracted.
EXTRACT_ELEMENTS?= # empty
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
DOWNLOADED_DISTFILE= $${extract_file}
Rewrite the extraction commands to be more intuitive. In the vast majority of cases, nothing has changed (i.e. .tgz, .tar.gz, and .tar.bz2). EXTRACT_USING_PAX can be set as before. For custom extractions, instead of using EXTRACT_BEFORE_ARGS, EXTRACT_AFTER_ARGS and EXTRACT_CMD, simply set EXTRACT_CMD to be the command needed to decompress and extract the lements from the archive. ${DOWNLOADED_DISTFILE} can be used to reference the distfile(s). e.g. for compressed shars, where previously there was: EXTRACT_CMD= ${GZCAT} EXTRACT_BEFORE_ARGS= EXTRACT_AFTER_ARGS= |sh now use: EXTRACT_CMD= ${GZCAT} ${DOWNLOADED_DISTFILE} | ${SH}
1999-04-01 16:07:52 +02:00
Add public variables EXTRACT_CMD_OPTS.{zip,lha,lhz,zoo} that may be set in package Makefiles to customize the options passed to the extraction commands when the defaults don't suffice. Patch from Michal Pasternak with additions by me.
2003-11-23 00:47:47 +01:00
EXTRACT_CMD.zip?= ${LOCALBASE}/bin/unzip ${EXTRACT_CMD_OPTS.zip} $${extract_file}
EXTRACT_CMD_OPTS.zip?= -Laqo
EXTRACT_CMD.lha?= ${LOCALBASE}/bin/lha ${EXTRACT_CMD_OPTS.lha} $${extract_file}
EXTRACT_CMD_OPTS.lha?= xq
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
EXTRACT_CMD.lzh?= ${EXTRACT_CMD.lha}
Add public variables EXTRACT_CMD_OPTS.{zip,lha,lhz,zoo} that may be set in package Makefiles to customize the options passed to the extraction commands when the defaults don't suffice. Patch from Michal Pasternak with additions by me.
2003-11-23 00:47:47 +01:00
EXTRACT_CMD_OPTS.lzh?= ${EXTRACT_CMD_OPTS.lha}
EXTRACT_CMD.zoo?= ${LOCALBASE}/bin/unzoo ${EXTRACT_CMD_OPTS.zoo} $${extract_file}
EXTRACT_CMD_OPTS.zoo?= -x
Enable EXTRACT_{ONLY,SUFX}= .rar. Ok'ed by jlam
2004-02-12 22:29:28 +01:00
EXTRACT_CMD.rar?= ${LOCALBASE}/bin/unrar ${EXTRACT_CMD_OPTS.rar} $${extract_file}
EXTRACT_CMD_OPTS.rar?= x -inul
Add EXTRACT_ENV.bin, allowing things to be passed into the environment of a ".bin" extraction (to be used shortly by lang/sun-jre15). This isn't a generic EXTRACT_ENV, as the rest of the extraction tools don't currently need such a knob. The most common use would be to set things in the env of a binary-only distribution to make for silent or otherwise twiddled package extraction.
2004-10-09 05:47:13 +02:00
EXTRACT_ENV.bin?= # empty
EXTRACT_CMD.bin?= ${ECHO} yes | ${SETENV} ${EXTRACT_ENV.bin} $${extract_file} ${EXTRACT_CMD_OPTS.bin} >/dev/null
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
.for __suffix__ in .gz .bz2 .Z
EXTRACT_CMD${__suffix__}?= ${DECOMPRESS_CMD${__suffix__}} $${extract_file} > `${BASENAME} $${extract_file} ${__suffix__}`
.endfor
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
Recognize how to automatically extract .shar.gz, .shar.bz2, .shar.Z, and .shar distfiles.
2002-01-27 19:31:47 +01:00
.for __suffix__ in .shar.gz .shar.bz2 .shar.Z .shar
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
EXTRACT_CMD${__suffix__}?= ${DECOMPRESS_CMD${__suffix__}} $${extract_file} | ${SH}
Recognize how to automatically extract .shar.gz, .shar.bz2, .shar.Z, and .shar distfiles.
2002-01-27 19:31:47 +01:00
.endfor
Get rid of dd workaround for old paxes (last seen in NetBSD before 1.5; bootstrap pax doesn't need it either). Closes PR 18814 by Greg A. Woods.
2004-03-31 13:19:09 +02:00
# If EXTRACT_USING_PAX is defined, use pax in preference to (GNU) tar.
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
#
.if defined(EXTRACT_USING_PAX)
Get rid of dd workaround for old paxes (last seen in NetBSD before 1.5; bootstrap pax doesn't need it either). Closes PR 18814 by Greg A. Woods.
2004-03-31 13:19:09 +02:00
_DFLT_EXTRACT_CMD?= ${DECOMPRESS_CMD} $${extract_file} | ${PAX} -O -r ${EXTRACT_ELEMENTS}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
.else
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
_DFLT_EXTRACT_CMD?= ${DECOMPRESS_CMD} $${extract_file} | ${GTAR} -xf - ${EXTRACT_ELEMENTS}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
.endif
Fix spelling of "suffixes". From Todd Vierling in PR 24400.
2004-02-13 20:02:07 +01:00
.for __suffix__ in ${_EXTRACT_SUFFIXES}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
. if !defined(EXTRACT_CMD${__suffix__})
. if defined(EXTRACT_USING_PAX)
Get rid of dd workaround for old paxes (last seen in NetBSD before 1.5; bootstrap pax doesn't need it either). Closes PR 18814 by Greg A. Woods.
2004-03-31 13:19:09 +02:00
EXTRACT_CMD${__suffix__}?= ${DECOMPRESS_CMD${__suffix__}} $${extract_file} | ${PAX} -O -r ${EXTRACT_ELEMENTS}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
. else
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
EXTRACT_CMD${__suffix__}?= ${DECOMPRESS_CMD${__suffix__}} $${extract_file} | ${GTAR} -xf - ${EXTRACT_ELEMENTS}
Teach bsd.pkg.mk how to extract all files in ${EXTRACT_ONLY} that end in suffices listed in ${_EXTRACT_SUFFICES}. Currently, _EXTRACT_SUFFICES has .tar.bz2, .tbz, .tar.gz, .tgz, .tar, .zip, .lzh, and .lha. If the file doesn't end in any of these suffices, then we use the old code to do the extraction. This rids us of the mild annoyance where if you have several distfiles, some ending in .tar.bz2, some in .tar.gz, and others in .zip, then you have to create a post-extract target to extract all of the ones not ending in the suffix that you designate in EXTRACT_SUFX. I've tested this with some representative packages: archivers/gtar-base, print/cups, www/surfraw, and print/ghostscript-nox11.
2001-12-20 00:14:01 +01:00
. endif
. endif
.endfor
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
# _SHELL_EXTRACT is a "subroutine" for extracting an archive. It extracts
# the contents of archive named by the shell variable "extract_file" based
# on the file extension of the archive.
Optimize the code to extract distfiles based on the extract suffix by putting them all into one large case statement instead of successive if statements, and by removing the creation and constant testing of a temp file to mark when the distfile is extracted.
2002-01-25 08:16:16 +01:00
#
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
_SHELL_EXTRACT= case $${extract_file} in
Fix spelling of "suffixes". From Todd Vierling in PR 24400.
2004-02-13 20:02:07 +01:00
.for __suffix__ in ${_EXTRACT_SUFFIXES}
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
_SHELL_EXTRACT+= *${__suffix__}) ${EXTRACT_CMD${__suffix__}} ;;
Optimize the code to extract distfiles based on the extract suffix by putting them all into one large case statement instead of successive if statements, and by removing the creation and constant testing of a temp file to mark when the distfile is extracted.
2002-01-25 08:16:16 +01:00
.endfor
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
_SHELL_EXTRACT+= *) ${_DFLT_EXTRACT_CMD} ;;
_SHELL_EXTRACT+= esac
Optimize the code to extract distfiles based on the extract suffix by putting them all into one large case statement instead of successive if statements, and by removing the creation and constant testing of a temp file to mark when the distfile is extracted.
2002-01-25 08:16:16 +01:00
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
EXTRACT_CMD?= ${_SHELL_EXTRACT}
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-extract
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
.if !target(do-extract)
do-extract: ${WRKDIR}
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
. for __file__ in ${EXTRACT_ONLY}
Rearrange code so that if EXTRACT_CMD is defined, then it is used to extract all of the files listed in ${EXTRACT_ONLY}. This is so that a package author doesn't have to grub through bsd.pkg.mk in order to find how to override the extraction method. Problem noted by Antti Kantee in private email.
2002-01-15 01:32:12 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
* Turn _SHELL_EXTRACT into a "subroutine" for extracting an archive. It extracts the contents of archive named by the shell variable "extract_file" based on the file extension of the archive. The contents are dumped into the current directory. * Unify the EXTRACT_CMD handling. We now define the default EXTRACT_CMD to be ${_SHELL_EXTRACT} (see above). Otherwise, if it's already defined in the package Makefile, then that EXTRACT_CMD is used to extract all of the archives. * Rearrange the locking code so that it's localized instead of interspersed in the extraction code. * Add EXTRACT_CMD definitions for files that are only compressed (foo.c.gz) that decompress the file into ${WRKDIR}.
2002-02-25 05:43:55 +01:00
extract_file="${_DISTDIR}/${__file__}"; export extract_file; \
cd ${WRKDIR}; ${EXTRACT_CMD}
. endfor
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
# Patch
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
# LOCALPATCHES contains the location of local patches to packages
# that are maintained in a directory tree reflecting the same
# hierarchy as the pkgsrc tree, i.e. local patches for www/apache
# would be found as ${LOCALPATCHES}/www/apache/*.
#
only add ${LOCALPATCHES}/${PKGPATH}/* to the list of patches to apply if LOCALPATCHES is defined. fixes problem on Solaris w/ automounter handling /net, where any package from the 'net' category would cause a blocking read on /net/pkgname looking for 'pkgname' as a host exporting NFS file systems (saves a few useless stat() calls, too).
2003-04-15 21:51:21 +02:00
.if defined(LOCALPATCHES)
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
_DFLT_LOCALPATCHFILES= ${LOCALPATCHES}/${PKGPATH}/*
_LOCALPATCHFILES= ${_DFLT_LOCALPATCHFILES}
only add ${LOCALPATCHES}/${PKGPATH}/* to the list of patches to apply if LOCALPATCHES is defined. fixes problem on Solaris w/ automounter handling /net, where any package from the 'net' category would cause a blocking read on /net/pkgname looking for 'pkgname' as a host exporting NFS file systems (saves a few useless stat() calls, too).
2003-04-15 21:51:21 +02:00
.endif
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-patch
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(do-patch)
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
do-patch: uptodate-digest
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(PATCHFILES)
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> Applying distribution patches for ${PKGNAME}"
Add some new variables to control how "distribution" patches are extracted and applied to the package sources. The existing options are: PATCH_DIST_STRIP - option that sets the pathname strip count. PATCH_DIST_ARGS - the list of arguments to pass to patch(1). The new options are: PATCH_DIST_CAT is the command that outputs the patch to stdout. For each of these variables, there is a patch-specific variant that may be set, i.e. PATCH_DIST_STRIP.<patch>, PATCH_DIST_ARGS.<patch>, PATCH_DIST_CAT.<patch>.
2004-08-07 17:58:59 +02:00
. for i in ${PATCHFILES}
Standardise on the format of continuation lines. No functional change.
2002-05-13 13:43:05 +02:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${_DISTDIR}; \
Add some new variables to control how "distribution" patches are extracted and applied to the package sources. The existing options are: PATCH_DIST_STRIP - option that sets the pathname strip count. PATCH_DIST_ARGS - the list of arguments to pass to patch(1). The new options are: PATCH_DIST_CAT is the command that outputs the patch to stdout. For each of these variables, there is a patch-specific variant that may be set, i.e. PATCH_DIST_STRIP.<patch>, PATCH_DIST_ARGS.<patch>, PATCH_DIST_CAT.<patch>.
2004-08-07 17:58:59 +02:00
if [ ${PATCH_DEBUG_TMP} = yes ]; then \
${ECHO_MSG} "${_PKGSRC_IN}> Applying distribution patch ${i}"; \
fi; \
${PATCH_DIST_CAT.${i:S/=/--/}} | \
${PATCH} ${PATCH_DIST_ARGS.${i:S/=/--/}} \
Add a knob PKGSRC_SHOW_PATCH_ERRORMSG to control whether the newly verbose message displayed on patch failure is in fact displayed on patch failure. It defaults to "no" if PKG_DEVELOPER is defined, "yes" otherwise. While here, rename PKG_PATCH_FAIL to PKGSRC_PATCH_FAIL.
2004-08-25 06:09:10 +02:00
|| { ${ECHO} "Patch ${i} failed"; ${PKGSRC_PATCH_FAIL}; }
Add some new variables to control how "distribution" patches are extracted and applied to the package sources. The existing options are: PATCH_DIST_STRIP - option that sets the pathname strip count. PATCH_DIST_ARGS - the list of arguments to pass to patch(1). The new options are: PATCH_DIST_CAT is the command that outputs the patch to stdout. For each of these variables, there is a patch-specific variant that may be set, i.e. PATCH_DIST_STRIP.<patch>, PATCH_DIST_ARGS.<patch>, PATCH_DIST_CAT.<patch>.
2004-08-07 17:58:59 +02:00
. endfor
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
patchlist=""; \
if [ -d ${PATCHDIR} ]; then \
Add a cdrom-readme target, a clone of the readme target, for ease of use. The URLs in the generated README.html files can be specified by overriding the CDROM_PKG_URL_HOST and CDROM_PKG_URL_DIR definitions. Document the targets, and clean up some English, in Packages.txt
1998-09-23 15:09:32 +02:00
if [ "`${ECHO} ${PATCHDIR}/patch-*`" = "${PATCHDIR}/patch-*" ]; then \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Ignoring empty patch directory"; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
if [ -d ${PATCHDIR}/CVS ]; then \
Change the message "Perhaps you forgot the -P flag to cvs co or update?" to "Perhaps you forgot the -P flag to 'cvs checkout' or 'cvs update'?"
2002-01-21 22:09:41 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Perhaps you forgot the -P flag to 'cvs checkout' or 'cvs update'?"; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
fi; \
else \
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
patchlist=`${ECHO} ${PATCHDIR}/patch-*`; \
fi; \
fi; \
if [ "X${_LOCALPATCHFILES}" = "X${_DFLT_LOCALPATCHFILES}" ]; then \
localpatchfiles="`${ECHO} ${_LOCALPATCHFILES}`"; \
if [ "$${localpatchfiles}" != "${_LOCALPATCHFILES}" ]; then \
patchlist="$${patchlist} $${localpatchfiles}"; \
fi; \
else \
patchlist=`${ECHO} $${patchlist} ${_LOCALPATCHFILES}`; \
fi; \
if [ -n "$${patchlist}" ]; then \
'Applying pkgsrc patches' it is.
2002-11-26 16:19:50 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Applying pkgsrc patches for ${PKGNAME}" ; \
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
fail=""; \
for i in $${patchlist}; do \
if [ ! -f "$$i" ]; then \
${ECHO_MSG} "${_PKGSRC_IN}> $$i is not a valid patch file - skipping"; \
continue; \
fi; \
case $$i in \
*.orig|*.rej|*~) \
${ECHO_MSG} "${_PKGSRC_IN}> Ignoring patchfile $$i"; \
continue; \
;; \
${PATCHDIR}/patch-local-*) \
;; \
${PATCHDIR}/patch-*) \
if [ -f ${DISTINFO_FILE} ]; then \
filename=`expr $$i : '.*/\(.*\)'`; \
algsum=`${AWK} 'NF == 4 && $$2 == "('$$filename')" && $$3 == "=" {print $$1 " " $$4}' ${DISTINFO_FILE} || ${TRUE}`; \
if [ "X$$algsum" != "X" ]; then \
alg=`${ECHO} $$algsum | ${AWK} '{ print $$1 }'`; \
recorded=`${ECHO} $$algsum | ${AWK} '{ print $$2 }'`; \
calcsum=`${SED} -e '/\$$NetBSD.*/d' $$i | ${DIGEST} $$alg`; \
if [ ${PATCH_DEBUG_TMP} = yes ]; then \
${ECHO_MSG} "=> Verifying $$filename (using digest algorithm $$alg)"; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
fi; \
fi; \
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
if [ "X$$algsum" = "X" -o "X$$recorded" = "X" ]; then \
${ECHO_MSG} "**************************************"; \
${ECHO_MSG} "Ignoring unknown patch file: $$i"; \
${ECHO_MSG} "**************************************"; \
continue; \
fi; \
if [ "X$$calcsum" != "X$$recorded" ]; then \
${ECHO_MSG} "**************************************"; \
${ECHO_MSG} "Patch file $$i has been modified"; \
${ECHO_MSG} "**************************************"; \
fail="$$fail $$filename"; \
continue; \
fi; \
When there is no distinfo file, do not apply patches we would not normally apply: treat them as if the distinfo file was empty. This solves the reason why pkgtools/pkg_install/distinfo was resurrected, and allows us to remove it again, since it had the side effect of introducing a dependency on digest, which is not acceptable when updating pkg_install: it made it impossible to use the pkgsrc tree, as reported by several people.
2003-11-20 16:07:53 +01:00
else \
${ECHO_MSG} "**************************************"; \
${ECHO_MSG} "Ignoring unknown patch file: $$i"; \
${ECHO_MSG} "**************************************"; \
Forgot a \ to make line continue. Sorry for breakage.
2003-11-20 17:59:02 +01:00
continue; \
Actually apply any patch-local-* local patch files if there are any.
1999-07-27 17:01:30 +02:00
fi; \
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
;; \
esac; \
if [ ${PATCH_DEBUG_TMP} = yes ]; then \
'Applying pkgsrc patches' it is.
2002-11-26 16:19:50 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Applying pkgsrc patch $$i" ; \
Re-visit the patch-sum modifications: (a) if the patch file does not exist in the patch-sum file, then print a warning to the user, and ignore the patch file. In particular, do not exit with an error. (b) if the patch file's calculated checksum does not match the one recorded in the patch-sum file, print a warning to the user, and try the other patch files. Once all patch files have been attempted, exit with an error. Also, don't bother to check for patch-local-* files inside the checksum logic - they're already ignored at the head of the loop.
1999-07-26 18:46:43 +02:00
fi; \
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
fuzz=""; \
${PATCH} -v > /dev/null 2>&1 && fuzz="${PATCH_FUZZ_FACTOR}"; \
${PATCH} $$fuzz ${PATCH_ARGS} < $$i || \
Add a knob PKGSRC_SHOW_PATCH_ERRORMSG to control whether the newly verbose message displayed on patch failure is in fact displayed on patch failure. It defaults to "no" if PKG_DEVELOPER is defined, "yes" otherwise. While here, rename PKG_PATCH_FAIL to PKGSRC_PATCH_FAIL.
2004-08-25 06:09:10 +02:00
{ ${ECHO} Patch $$i failed ; ${PKGSRC_PATCH_FAIL}; }; \
Re-write do-patch target to fix broken behaviour where if patches/ didn't exist, then the local patches weren't applied.
2002-04-25 02:02:41 +02:00
done; \
if [ "X$$fail" != "X" ]; then \
${ECHO_MSG} "Patching failed due to modified patch file(s): $$fail"; \
Add a knob PKGSRC_SHOW_PATCH_ERRORMSG to control whether the newly verbose message displayed on patch failure is in fact displayed on patch failure. It defaults to "no" if PKG_DEVELOPER is defined, "yes" otherwise. While here, rename PKG_PATCH_FAIL to PKGSRC_PATCH_FAIL.
2004-08-25 06:09:10 +02:00
${PKGSRC_PATCH_FAIL}; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
fi; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fi
Move the REPLACE_PERL code into a separate configure pre-requisite target "replace-perl", instead of keeping it as part of the do-patch target. Now, a "make patch" will result in just the patched sources.
2001-11-30 04:48:14 +01:00
.endif
# Configure
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Move code from end of the do-configure target that performed the LIBTOOL_OVERRIDE steps into a separate do-libtool-override target. Create a new variable _CONFIGURE_POSTREQ that currently lists do-libtool-override. Also add some documentation for the two _CONFIGURE_* variables: _CONFIGURE_PREREQ is a list of targets to run after pre-configure but before do-configure. These targets typically edit the files used by the do-configure target. _CONFIGURE_POSTREQ is a list of targets to run after do-configure but before post-configure. These targets typically edit the files generated by the do-configure target that are used during the build phase.
2001-11-30 06:03:06 +01:00
# _CONFIGURE_PREREQ is a list of targets to run after pre-configure but before
# do-configure. These targets typically edit the files used by the
Fix a typo in a comment
2003-07-08 13:48:37 +02:00
# do-configure target. The targets are run as dependencies of
Fix for _CONFIGURE_PREREQ targets not being run when a package defines a do-configure target. To do this I've introduced a new target, pre-configure-overrides, that depends on ${_CONFIGURE_PREREQ}. It's run after pre-configure but before do-configure. This fixes pkg/22077, and allows benchmarks/benchfft to build correctly on a -current cats box. Reviewed by wiz@
2003-07-08 11:20:30 +02:00
# pre-configure-override.
Move code from end of the do-configure target that performed the LIBTOOL_OVERRIDE steps into a separate do-libtool-override target. Create a new variable _CONFIGURE_POSTREQ that currently lists do-libtool-override. Also add some documentation for the two _CONFIGURE_* variables: _CONFIGURE_PREREQ is a list of targets to run after pre-configure but before do-configure. These targets typically edit the files used by the do-configure target. _CONFIGURE_POSTREQ is a list of targets to run after do-configure but before post-configure. These targets typically edit the files generated by the do-configure target that are used during the build phase.
2001-11-30 06:03:06 +01:00
#
# _CONFIGURE_POSTREQ is a list of targets to run after do-configure but before
# post-configure. These targets typically edit the files generated by
# the do-configure target that are used during the build phase.
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
.if defined(USE_PKGLOCALEDIR)
_PKGLOCALEDIR= ${PREFIX}/${PKGLOCALEDIR}/locale
REPLACE_LOCALEDIR_PATTERNS?= # empty
_REPLACE_LOCALEDIR_PATTERNS= ${REPLACE_LOCALEDIR_PATTERNS}
. if defined(HAS_CONFIGURE) || defined(GNU_CONFIGURE)
Allow more case variations on "Makefile" to be included in USE_PKGLOCALEDIR transformations.
2004-02-13 13:03:06 +01:00
_REPLACE_LOCALEDIR_PATTERNS+= [Mm]akefile.in*
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
. else
Allow more case variations on "Makefile" to be included in USE_PKGLOCALEDIR transformations.
2004-02-13 13:03:06 +01:00
_REPLACE_LOCALEDIR_PATTERNS+= [Mm]akefile*
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
. endif
_REPLACE_LOCALEDIR_PATTERNS_FIND= \
\( ${_REPLACE_LOCALEDIR_PATTERNS:S/$/!/:S/^/-o -name !/:S/!/"/g:S/-o//1} \)
REPLACE_LOCALEDIR?= # empty
_REPLACE_LOCALEDIR= \
${REPLACE_LOCALEDIR} \
`${FIND} . ${_REPLACE_LOCALEDIR_PATTERNS_FIND} -print | ${SED} -e 's|^\./||' | ${GREP} -v '\.orig' | ${SORT} -u`
_CONFIGURE_PREREQ+= subst-pkglocaledir
. if empty(USE_PKGLOCALEDIR:M[nN][oO])
SUBST_CLASSES+= pkglocaledir
. endif
SUBST_MESSAGE.pkglocaledir= "Fixing locale directory references."
SUBST_FILES.pkglocaledir= ${_REPLACE_LOCALEDIR}
SUBST_SED.pkglocaledir= \
-e 's|^\(localedir[ :]*=\).*|\1 ${_PKGLOCALEDIR}|' \
-e 's|^\(gnulocaledir[ :]*=\).*|\1 ${_PKGLOCALEDIR}|' \
In SUBST_SED.pkglocaledir, escape the rest of the line in the pattern. Saves trailing \ properly. Addresses PR pkg/26022 by Akio OBATA.
2004-06-23 13:42:18 +02:00
-e 's|\(-DLOCALEDIR[ ]*=\)[^ ]*\(\.\*\)|\1"\\"${_PKGLOCALEDIR}\\""\2|'
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
.endif
Move the REPLACE_PERL code into a separate configure pre-requisite target "replace-perl", instead of keeping it as part of the do-patch target. Now, a "make patch" will result in just the patched sources.
2001-11-30 04:48:14 +01:00
.if defined(REPLACE_PERL)
Append perl to the list of interpreters to be replaced if ${REPLACE_PERL} is set, rather than a straight assignation.
2003-07-09 18:18:06 +02:00
REPLACE_INTERPRETER+= perl
Make REPLACE_PERL more general, so that it will work with any interpreter, and also with multiple interpreters. Keep REPLACE_PERL for backwards compatibility.
2003-07-09 18:07:29 +02:00
_REPLACE.perl.old= .*/bin/perl
_REPLACE.perl.new= ${PERL5}
_REPLACE_FILES.perl= ${REPLACE_PERL}
.endif
Remove some extraneous commands and switch some targets to be conditionally defined.
2004-01-19 15:54:26 +01:00
.if defined(REPLACE_INTERPRETER)
Make REPLACE_PERL more general, so that it will work with any interpreter, and also with multiple interpreters. Keep REPLACE_PERL for backwards compatibility.
2003-07-09 18:07:29 +02:00
_CONFIGURE_PREREQ+= replace-interpreter
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: replace-interpreter
Make REPLACE_PERL more general, so that it will work with any interpreter, and also with multiple interpreters. Keep REPLACE_PERL for backwards compatibility.
2003-07-09 18:07:29 +02:00
replace-interpreter:
. for lang in ${REPLACE_INTERPRETER}
extend REPLACE_INTERPRETER to accept shell wildcards
2004-03-02 16:25:23 +01:00
. for pattern in ${_REPLACE_FILES.${lang}}
Move the REPLACE_PERL code into a separate configure pre-requisite target "replace-perl", instead of keeping it as part of the do-patch target. Now, a "make patch" will result in just the patched sources.
2001-11-30 04:48:14 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
cd ${WRKSRC}; \
extend REPLACE_INTERPRETER to accept shell wildcards
2004-03-02 16:25:23 +01:00
for f in ${pattern}; do \
Rewrite REPLACE_PERL as shell-loop, so one can use shell wildcards,e.g. REPLACE_PERL= *.cgi *.perl *.pl Code submitted by Amitai Schlair <schmonz@schmonz.com>
2002-03-24 15:06:22 +01:00
if [ -f $${f} ]; then \
extend REPLACE_INTERPRETER to accept shell wildcards
2004-03-02 16:25:23 +01:00
${SED} -e '1s|^#!${_REPLACE.${lang}.old}|#!${_REPLACE.${lang}.new}|' \
Rewrite REPLACE_PERL as shell-loop, so one can use shell wildcards,e.g. REPLACE_PERL= *.cgi *.perl *.pl Code submitted by Amitai Schlair <schmonz@schmonz.com>
2002-03-24 15:06:22 +01:00
$${f} > $${f}.new; \
if [ -x $${f} ]; then \
${CHMOD} a+x $${f}.new; \
fi; \
${MV} -f $${f}.new $${f}; \
fi; \
done
extend REPLACE_INTERPRETER to accept shell wildcards
2004-03-02 16:25:23 +01:00
. endfor
Make REPLACE_PERL more general, so that it will work with any interpreter, and also with multiple interpreters. Keep REPLACE_PERL for backwards compatibility.
2003-07-09 18:07:29 +02:00
. endfor
o move REPLACE_PERL code to more proper location -- to the do-patch target; now it works when NO_CONFIGURE is set
2001-06-27 13:40:25 +02:00
.endif
Add USE_CURSES for packages that need curses functionality that only got added in 1.4Y; on older version of NetBSD, ncurses will be used. Replace previous code that did the same, but wasn't shared.
2000-07-28 12:33:56 +02:00
Remove some extraneous commands and switch some targets to be conditionally defined.
2004-01-19 15:54:26 +01:00
.if defined(USE_LIBTOOL) && defined(LTCONFIG_OVERRIDE)
Rename "do-libtool" target to "do-ltconfig-override" as the new name is more descriptive. There is a separate step to replace the actual libtool script that is controlled by LIBTOOL_OVERRIDE within the "do-configure" target.
2001-11-30 04:19:04 +01:00
_CONFIGURE_PREREQ+= do-ltconfig-override
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-ltconfig-override
Rename "do-libtool" target to "do-ltconfig-override" as the new name is more descriptive. There is a separate step to replace the actual libtool script that is controlled by LIBTOOL_OVERRIDE within the "do-configure" target.
2001-11-30 04:19:04 +01:00
do-ltconfig-override:
The do-configure target has some "stuff" that is supposed to be done after the pre-configure target is called but before the configure script is called, but it's added in a non-extensible way. This "stuff" is the replace-ncurses step and the ltconfig-override step. Move these steps out into their own targets that are named as prerequisites to the do-configure step. The prerequisites are specified in the private variable _CONFIGURE_PREREQ, to which other independent targets may be appended.
2001-10-24 21:42:28 +02:00
. for ltconfig in ${LTCONFIG_OVERRIDE}
${_PKG_SILENT}${_PKG_DEBUG} \
if [ -f ${ltconfig} ]; then \
${RM} -f ${ltconfig}; \
We need the full path to the libtool that needs to be invoked when we use LIBTOOL_OVERRIDE. In the buildlink[23] case, that is supposed to be the one in ${BUILDLINK_DIR}. Create new private variables _LIBTOOL and _SHLIBTOOL to hold these paths.
2004-02-09 02:30:59 +01:00
${ECHO} "${RM} -f libtool; ${LN} -s ${_LIBTOOL} libtool" \
The do-configure target has some "stuff" that is supposed to be done after the pre-configure target is called but before the configure script is called, but it's added in a non-extensible way. This "stuff" is the replace-ncurses step and the ltconfig-override step. Move these steps out into their own targets that are named as prerequisites to the do-configure step. The prerequisites are specified in the private variable _CONFIGURE_PREREQ, to which other independent targets may be appended.
2001-10-24 21:42:28 +02:00
> ${ltconfig}; \
${CHMOD} +x ${ltconfig}; \
Added LTCONFIG_OVERRIDE and LIBTOOL_OVERRIDE definition to substitute pkglibtool for the pkg's own libtool. "ltconfig" is script to make "libtool" script. "ltconfig" is executed by GNU's "configure" script at "do-configure" target. if USE_LIBTOOL and LTCONFIG_OVERRIDE defined, specified "ltconfig" is overrided to symlink pkglibtool to "libtool" instead of making "libtool" before executing "configure". if USE_LIBTOOL and LIBTOOL_OVERRIDE defined, specified "libtool" is removed, and symlinked from pkglibtool after executing "configure" (and "ltconfig").
1999-12-07 09:55:58 +01:00
fi
The do-configure target has some "stuff" that is supposed to be done after the pre-configure target is called but before the configure script is called, but it's added in a non-extensible way. This "stuff" is the replace-ncurses step and the ltconfig-override step. Move these steps out into their own targets that are named as prerequisites to the do-configure step. The prerequisites are specified in the private variable _CONFIGURE_PREREQ, to which other independent targets may be appended.
2001-10-24 21:42:28 +02:00
. endfor
.endif
Add two knobs for packages: CONFIG_GUESS_OVERRIDE and CONFIG_SUB_OVERRIDE. Example: CONFIG_GUESS_OVERRIDE= ${WRKSRC}/config.guess CONFIG_SUB_OVERRIDE= ${WRKSRC}/config.sub Just before the bulk of the "configure" phase, the named files will be replaced with symlinks to their canonical pkgsrc versions in pkgsrc/mk/gnu-config.
2002-07-16 17:02:01 +02:00
_CONFIGURE_PREREQ+= do-config-star-override
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-config-star-override
Add two knobs for packages: CONFIG_GUESS_OVERRIDE and CONFIG_SUB_OVERRIDE. Example: CONFIG_GUESS_OVERRIDE= ${WRKSRC}/config.guess CONFIG_SUB_OVERRIDE= ${WRKSRC}/config.sub Just before the bulk of the "configure" phase, the named files will be replaced with symlinks to their canonical pkgsrc versions in pkgsrc/mk/gnu-config.
2002-07-16 17:02:01 +02:00
do-config-star-override:
Only do the config.* overrides iff GNU_CONFIGURE. (defined() check needed to make older bmakes happy.)
2005-01-14 15:08:06 +01:00
.if defined(GNU_CONFIGURE)
. if !empty(CONFIG_GUESS_OVERRIDE)
. for _pattern_ in ${CONFIG_GUESS_OVERRIDE}
Make CONFIG_{GUESS,SUB}_OVERRIDE slighly more powerful by allowing them to be lists of shell globs relative to ${WRKSRC}.
2004-02-14 16:22:42 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${WRKSRC}; \
for file in ${_pattern_}; do \
if [ -f "$$file" ]; then \
${RM} -f $$file; \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
${LN} -s ${PKGSRCDIR}/mk/gnu-config/config.guess \
Make CONFIG_{GUESS,SUB}_OVERRIDE slighly more powerful by allowing them to be lists of shell globs relative to ${WRKSRC}.
2004-02-14 16:22:42 +01:00
$$file; \
fi; \
Add two knobs for packages: CONFIG_GUESS_OVERRIDE and CONFIG_SUB_OVERRIDE. Example: CONFIG_GUESS_OVERRIDE= ${WRKSRC}/config.guess CONFIG_SUB_OVERRIDE= ${WRKSRC}/config.sub Just before the bulk of the "configure" phase, the named files will be replaced with symlinks to their canonical pkgsrc versions in pkgsrc/mk/gnu-config.
2002-07-16 17:02:01 +02:00
done
Only do the config.* overrides iff GNU_CONFIGURE. (defined() check needed to make older bmakes happy.)
2005-01-14 15:08:06 +01:00
. endfor
. endif
. if !empty(CONFIG_SUB_OVERRIDE)
. for _pattern_ in ${CONFIG_SUB_OVERRIDE}
Make CONFIG_{GUESS,SUB}_OVERRIDE slighly more powerful by allowing them to be lists of shell globs relative to ${WRKSRC}.
2004-02-14 16:22:42 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${WRKSRC}; \
for file in ${_pattern_}; do \
if [ -f "$$file" ]; then \
${RM} -f $$file; \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
${LN} -s ${PKGSRCDIR}/mk/gnu-config/config.sub \
Make CONFIG_{GUESS,SUB}_OVERRIDE slighly more powerful by allowing them to be lists of shell globs relative to ${WRKSRC}.
2004-02-14 16:22:42 +01:00
$$file; \
fi; \
Add two knobs for packages: CONFIG_GUESS_OVERRIDE and CONFIG_SUB_OVERRIDE. Example: CONFIG_GUESS_OVERRIDE= ${WRKSRC}/config.guess CONFIG_SUB_OVERRIDE= ${WRKSRC}/config.sub Just before the bulk of the "configure" phase, the named files will be replaced with symlinks to their canonical pkgsrc versions in pkgsrc/mk/gnu-config.
2002-07-16 17:02:01 +02:00
done
Only do the config.* overrides iff GNU_CONFIGURE. (defined() check needed to make older bmakes happy.)
2005-01-14 15:08:06 +01:00
. endfor
. endif
. if !empty(CONFIG_RPATH_OVERRIDE)
. for _pattern_ in ${CONFIG_RPATH_OVERRIDE}
Add override for uncommon "config.rpath" GNU config snippet. Use platform file to indicate that override logic should be turned on. (AFAICT, only Interix is afflicted at the moment.)
2005-01-14 01:10:01 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${WRKSRC}; \
for file in ${_pattern_}; do \
if [ -f "$$file" ]; then \
${RM} -f $$file; \
${LN} -s ${PKGSRCDIR}/mk/gnu-config/config.rpath \
$$file; \
fi; \
done
Only do the config.* overrides iff GNU_CONFIGURE. (defined() check needed to make older bmakes happy.)
2005-01-14 15:08:06 +01:00
. endfor
. endif
Add two knobs for packages: CONFIG_GUESS_OVERRIDE and CONFIG_SUB_OVERRIDE. Example: CONFIG_GUESS_OVERRIDE= ${WRKSRC}/config.guess CONFIG_SUB_OVERRIDE= ${WRKSRC}/config.sub Just before the bulk of the "configure" phase, the named files will be replaced with symlinks to their canonical pkgsrc versions in pkgsrc/mk/gnu-config.
2002-07-16 17:02:01 +02:00
.endif
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
PKGCONFIG_OVERRIDE_SED= \
Introduce the PKGCONFIG_OVERRIDE_STAGE variable to let a package specify when the pkgconfig substitutions have to be applied; defaults to pre-configure. This is useful because some packages modify the .pc files during the build process (adding -L flags), so they have to wait until the build finishes to add the correct rpaths. Also modify the regular expression in PKGCONFIG_OVERRIDE_SED to match and fix multiple -L flags, instead of just one.
2004-10-25 20:02:20 +02:00
'/^Libs:.*[ ]/s|-L\([ ]*[^ ]*\)|${COMPILER_RPATH_FLAG}\1 -L\1|g'
PKGCONFIG_OVERRIDE_STAGE?= pre-configure
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
.if defined(PKGCONFIG_OVERRIDE) && !empty(PKGCONFIG_OVERRIDE)
Introduce the PKGCONFIG_OVERRIDE_STAGE variable to let a package specify when the pkgconfig substitutions have to be applied; defaults to pre-configure. This is useful because some packages modify the .pc files during the build process (adding -L flags), so they have to wait until the build finishes to add the correct rpaths. Also modify the regular expression in PKGCONFIG_OVERRIDE_SED to match and fix multiple -L flags, instead of just one.
2004-10-25 20:02:20 +02:00
. if ${PKGCONFIG_OVERRIDE_STAGE} == "pre-configure"
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
_CONFIGURE_PREREQ+= subst-pkgconfig
Introduce the PKGCONFIG_OVERRIDE_STAGE variable to let a package specify when the pkgconfig substitutions have to be applied; defaults to pre-configure. This is useful because some packages modify the .pc files during the build process (adding -L flags), so they have to wait until the build finishes to add the correct rpaths. Also modify the regular expression in PKGCONFIG_OVERRIDE_SED to match and fix multiple -L flags, instead of just one.
2004-10-25 20:02:20 +02:00
. elif ${PKGCONFIG_OVERRIDE_STAGE} == "post-configure"
_CONFIGURE_POSTREQ+= subst-pkgconfig
. else
SUBST_STAGE.pkgconfig= ${PKGCONFIG_OVERRIDE_STAGE}
. endif
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
SUBST_CLASSES+= pkgconfig
SUBST_MESSAGE.pkgconfig= "Adding rpaths to pkgconfig files."
SUBST_FILES.pkgconfig= ${PKGCONFIG_OVERRIDE:S/^${WRKSRC}\///}
SUBST_SED.pkgconfig= ${PKGCONFIG_OVERRIDE_SED}
Add PKGCONFIG_OVERRIDE to get rid of tons of unnecessary patches. PKGCONFIG_OVERRIDE may contain a list of ".pc.in" pkgconfig template files that require the addition of -Wl,-R in the pre-configure stage.
2002-10-30 03:25:29 +01:00
.endif
"the the" -> "the"
2004-12-01 10:52:59 +01:00
# By adding this target, it makes sure the above PREREQ's work.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: pre-configure-override
Fix for _CONFIGURE_PREREQ targets not being run when a package defines a do-configure target. To do this I've introduced a new target, pre-configure-overrides, that depends on ${_CONFIGURE_PREREQ}. It's run after pre-configure but before do-configure. This fixes pkg/22077, and allows benchmarks/benchfft to build correctly on a -current cats box. Reviewed by wiz@
2003-07-08 11:20:30 +02:00
pre-configure-override: ${_CONFIGURE_PREREQ}
In pre-config-override, don't trumpet the fact that we're doing nothing.
2003-07-09 16:23:56 +02:00
@${DO_NADA}
Fix for _CONFIGURE_PREREQ targets not being run when a package defines a do-configure target. To do this I've introduced a new target, pre-configure-overrides, that depends on ${_CONFIGURE_PREREQ}. It's run after pre-configure but before do-configure. This fixes pkg/22077, and allows benchmarks/benchfft to build correctly on a -current cats box. Reviewed by wiz@
2003-07-08 11:20:30 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-configure
The do-configure target has some "stuff" that is supposed to be done after the pre-configure target is called but before the configure script is called, but it's added in a non-extensible way. This "stuff" is the replace-ncurses step and the ltconfig-override step. Move these steps out into their own targets that are named as prerequisites to the do-configure step. The prerequisites are specified in the private variable _CONFIGURE_PREREQ, to which other independent targets may be appended.
2001-10-24 21:42:28 +02:00
.if !target(do-configure)
Remove trailing whitespace.
2004-02-12 14:16:02 +01:00
do-configure:
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(HAS_CONFIGURE)
Merge pkgsrc/mk from the buildlink2 branch back into the main trunk.
2002-08-22 10:21:26 +02:00
. for DIR in ${CONFIGURE_DIRS}
openoffice requires the unlimit on configure (for jdk) as well as for build. Add ${_ULIMIT_CMD} to do-configure:
2002-09-02 04:42:26 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${_ULIMIT_CMD}cd ${DIR} && ${SETENV} \
Visual cleanup from tv-derecurse: Consolidate many MAKE_ENVs and SCRIPTS_ENVs into a common block. (CONFIGURE_ENVs to be done later.) Introduce new variable ALL_ENV which is automatically included into all of MAKE_ENV, SCRIPTS_ENV, and CONFIGURE_ENV; this allows much cleaner addition of the common CC/CXX/CFLAGS/etc. variables needed by all of these.
2005-02-11 17:36:49 +01:00
AWK="${AWK}" \
Avoid shell smarts: on NetBSD, type is a sh(1) builtin. If make(1) is invoked, a subsequent "type make" will give "make is a tracked alias for /usr/bin/make". Using awk(1) to find the third part of that can lose. type is also a sh(1) builtin on Solaris, which prints "make is hashed (/usr/bin/make)". However, there is a separate /usr/bin/type, which prints the pathname at the end. Minor reformat in same target - replace spaces by tabs where possible.
1999-03-31 12:59:44 +02:00
INSTALL="`${TYPE} ${INSTALL} | ${AWK} '{ print $$NF }'` -c -o ${BINOWN} -g ${BINGRP}" \
When calling configure, set ac_given_INSTALL as well as INSTALL, to work around a bug in how autoconf 2.13-generated scripts pass INSTALL to sub-configures.
2000-05-02 20:44:37 +02:00
ac_given_INSTALL="`${TYPE} ${INSTALL} | ${AWK} '{ print $$NF }'` -c -o ${BINOWN} -g ${BINGRP}" \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
INSTALL_DATA="${INSTALL_DATA}" \
INSTALL_PROGRAM="${INSTALL_PROGRAM}" \
INSTALL_GAME="${INSTALL_GAME}" \
INSTALL_GAME_DATA="${INSTALL_GAME_DATA}" \
INSTALL_SCRIPT="${INSTALL_SCRIPT}" \
${CONFIGURE_ENV} ${CONFIG_SHELL} \
run GNU configure scripts with ${CONFIG_SHELL}, which defaults to ${SH} when GNU_CONFIGURE is defined. fixes problems on Solaris and BSDI with some GNU configure scripts.
2004-01-25 11:36:15 +01:00
${CONFIGURE_SCRIPT} ${CONFIGURE_ARGS}
Merge pkgsrc/mk from the buildlink2 branch back into the main trunk.
2002-08-22 10:21:26 +02:00
. endfor
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
* Push the imake- and xmkmf-handling into the new tools framework. * Get rid of an explicit check for ${_IMAKE_MAKE} == ${GMAKE} in bsd.pkg.mk to check for whether we need to depend on gmake or not. Instead, we now note in Linux.mk that packages that need imake will also need to use gmake by setting _IMAKE_TOOLS+=gmake. * Push the definition of MAKE_PROGRAM from bsd.pkg.mk into make.mk where it's closer to related code.
2005-05-02 23:10:02 +02:00
. if defined(USE_IMAKE) || !empty(USE_TOOLS:Mimake)
Make CONFIGURE_DIRS have meaning for programs that use imake.
2002-12-29 20:02:06 +01:00
. for DIR in ${CONFIGURE_DIRS}
${_PKG_SILENT}${_PKG_DEBUG}cd ${DIR} && ${SETENV} ${SCRIPTS_ENV} XPROJECTROOT=${X11BASE} ${XMKMF}
. endfor
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Move code from end of the do-configure target that performed the LIBTOOL_OVERRIDE steps into a separate do-libtool-override target. Create a new variable _CONFIGURE_POSTREQ that currently lists do-libtool-override. Also add some documentation for the two _CONFIGURE_* variables: _CONFIGURE_PREREQ is a list of targets to run after pre-configure but before do-configure. These targets typically edit the files used by the do-configure target. _CONFIGURE_POSTREQ is a list of targets to run after do-configure but before post-configure. These targets typically edit the files generated by the do-configure target that are used during the build phase.
2001-11-30 06:03:06 +01:00
.endif
Remove some extraneous commands and switch some targets to be conditionally defined.
2004-01-19 15:54:26 +01:00
.if defined(USE_LIBTOOL) && \
(defined(LIBTOOL_OVERRIDE) || defined(SHLIBTOOL_OVERRIDE))
Move code from end of the do-configure target that performed the LIBTOOL_OVERRIDE steps into a separate do-libtool-override target. Create a new variable _CONFIGURE_POSTREQ that currently lists do-libtool-override. Also add some documentation for the two _CONFIGURE_* variables: _CONFIGURE_PREREQ is a list of targets to run after pre-configure but before do-configure. These targets typically edit the files used by the do-configure target. _CONFIGURE_POSTREQ is a list of targets to run after do-configure but before post-configure. These targets typically edit the files generated by the do-configure target that are used during the build phase.
2001-11-30 06:03:06 +01:00
_CONFIGURE_POSTREQ+= do-libtool-override
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-libtool-override
Move code from end of the do-configure target that performed the LIBTOOL_OVERRIDE steps into a separate do-libtool-override target. Create a new variable _CONFIGURE_POSTREQ that currently lists do-libtool-override. Also add some documentation for the two _CONFIGURE_* variables: _CONFIGURE_PREREQ is a list of targets to run after pre-configure but before do-configure. These targets typically edit the files used by the do-configure target. _CONFIGURE_POSTREQ is a list of targets to run after do-configure but before post-configure. These targets typically edit the files generated by the do-configure target that are used during the build phase.
2001-11-30 06:03:06 +01:00
do-libtool-override:
Install a ${PREFIX}/bin/shlibtool as part of the devel/libtool-base package. shlibtool will not build static libraries. Add handling in bsd.pkg.mk for a variable named "SHLIBTOOL_OVERRIDE" that is analogous to LIBTOOL_OVERRIDE and causes any listed libtool scripts to be replaced with a symlink to shlibtool, and teach buildlink2 about shlibtool. Bump PKGREVISION of devel/libtool* packages to 11.
2003-01-10 09:44:18 +01:00
. if defined(LIBTOOL_OVERRIDE)
Make LIBTOOL_OVERRIDE more powerful by taking a list of shell globs relative to ${WRKSRC}. Also automaticslly set LIBTOOL_OVERRIDE if USE_LIBTOOL is defined, as the vast majority of packages need both if they need USE_LIBTOOL. The default LIBTOOL_OVERRIDE overrides libtool scripts up to a directory depth of 2 within ${WRKSRC}. Note that SHLIBTOOL_OVERRIDE must still be set manually in package Makefiles that need it; however, it's also been changed to be a list of shell globs relative to ${WRKSRC}.
2004-02-14 18:18:49 +01:00
. for _pattern_ in ${LIBTOOL_OVERRIDE}
${_PKG_SILENT}${_PKG_DEBUG}cd ${WRKSRC}; \
for file in ${_pattern_}; do \
if [ -f "$$file" ]; then \
${RM} -f $$file; \
(${ECHO} '#!${CONFIG_SHELL}'; \
${ECHO} 'exec ${_LIBTOOL} "$$@"'; \
) > $$file; \
${CHMOD} +x $$file; \
When overriding libtool in the source directory, replace it with a shell script that invokes the correct libtool instead of using a symlink. This fixes the problem where libtool was invoked by /bin/sh regardless of whether libtool was built to be run by /bin/ksh instead.
2004-02-01 02:09:09 +01:00
fi; \
Make LIBTOOL_OVERRIDE more powerful by taking a list of shell globs relative to ${WRKSRC}. Also automaticslly set LIBTOOL_OVERRIDE if USE_LIBTOOL is defined, as the vast majority of packages need both if they need USE_LIBTOOL. The default LIBTOOL_OVERRIDE overrides libtool scripts up to a directory depth of 2 within ${WRKSRC}. Note that SHLIBTOOL_OVERRIDE must still be set manually in package Makefiles that need it; however, it's also been changed to be a list of shell globs relative to ${WRKSRC}.
2004-02-14 18:18:49 +01:00
done
Install a ${PREFIX}/bin/shlibtool as part of the devel/libtool-base package. shlibtool will not build static libraries. Add handling in bsd.pkg.mk for a variable named "SHLIBTOOL_OVERRIDE" that is analogous to LIBTOOL_OVERRIDE and causes any listed libtool scripts to be replaced with a symlink to shlibtool, and teach buildlink2 about shlibtool. Bump PKGREVISION of devel/libtool* packages to 11.
2003-01-10 09:44:18 +01:00
. endfor
. endif
. if defined(SHLIBTOOL_OVERRIDE)
Make LIBTOOL_OVERRIDE more powerful by taking a list of shell globs relative to ${WRKSRC}. Also automaticslly set LIBTOOL_OVERRIDE if USE_LIBTOOL is defined, as the vast majority of packages need both if they need USE_LIBTOOL. The default LIBTOOL_OVERRIDE overrides libtool scripts up to a directory depth of 2 within ${WRKSRC}. Note that SHLIBTOOL_OVERRIDE must still be set manually in package Makefiles that need it; however, it's also been changed to be a list of shell globs relative to ${WRKSRC}.
2004-02-14 18:18:49 +01:00
. for _pattern_ in ${SHLIBTOOL_OVERRIDE}
${_PKG_SILENT}${_PKG_DEBUG}cd ${WRKSRC}; \
for file in ${_pattern_}; do \
if [ -f "$$file" ]; then \
${RM} -f $$file; \
(${ECHO} '#!${CONFIG_SHELL}'; \
${ECHO} 'exec ${_SHLIBTOOL} "$$@"'; \
) > $$file; \
${CHMOD} +x $$file; \
Also turn shlibtool into a shell script instead of a symlink. This duplicates the change made in rev. 1.1368 for libtool.
2004-02-09 02:52:14 +01:00
fi; \
Make LIBTOOL_OVERRIDE more powerful by taking a list of shell globs relative to ${WRKSRC}. Also automaticslly set LIBTOOL_OVERRIDE if USE_LIBTOOL is defined, as the vast majority of packages need both if they need USE_LIBTOOL. The default LIBTOOL_OVERRIDE overrides libtool scripts up to a directory depth of 2 within ${WRKSRC}. Note that SHLIBTOOL_OVERRIDE must still be set manually in package Makefiles that need it; however, it's also been changed to be a list of shell globs relative to ${WRKSRC}.
2004-02-14 18:18:49 +01:00
done
Install a ${PREFIX}/bin/shlibtool as part of the devel/libtool-base package. shlibtool will not build static libraries. Add handling in bsd.pkg.mk for a variable named "SHLIBTOOL_OVERRIDE" that is analogous to LIBTOOL_OVERRIDE and causes any listed libtool scripts to be replaced with a symlink to shlibtool, and teach buildlink2 about shlibtool. Bump PKGREVISION of devel/libtool* packages to 11.
2003-01-10 09:44:18 +01:00
. endfor
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Allow CONFIG_STATUS_OVERRIDE to override config.status scripts generated by GNU configure scripts at CONFIGURE_POSTREQ time. The new config.status scripts merely return success. This prevents newer Makefiles from re-running the configure script with the wrong shell environment if we've touched some GNU autotool-related files during the patch stage. We might want to always do this, i.e. make this opt-out instead of opt-in. However, we start with opt-in so that no existing packages can break.
2004-01-19 16:04:05 +01:00
.if defined(CONFIG_STATUS_OVERRIDE)
_CONFIGURE_POSTREQ+= do-config-status-override
.PHONY: do-config-status-override
Remove trailing whitespace.
2004-02-12 14:16:02 +01:00
do-config-status-override:
Default to overriding config.status after GNU configure has been run. We don't want it to execute during the build phase as it might overwrite post-configure changes we might have made to the generated files. Packages that don't want this behaviour can simply set: CONFIG_STATUS_OVERRIDE= # empty in the package Makefile.
2004-02-14 14:54:28 +01:00
. for _pattern_ in ${CONFIG_STATUS_OVERRIDE}
Change into ${WRKSRC} first so that shell globs expand properly.
2004-02-14 16:16:31 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${WRKSRC}; \
Default to overriding config.status after GNU configure has been run. We don't want it to execute during the build phase as it might overwrite post-configure changes we might have made to the generated files. Packages that don't want this behaviour can simply set: CONFIG_STATUS_OVERRIDE= # empty in the package Makefile.
2004-02-14 14:54:28 +01:00
for file in ${_pattern_}; do \
Change into ${WRKSRC} first so that shell globs expand properly.
2004-02-14 16:16:31 +01:00
if [ -f "$$file" ]; then \
Be less aggressive about overriding all of the functionality of config.status. Instead, only block "config.status --recheck", which re-executes the GNU configure script. This still allows config.status to overwrite post-configure changes to Makefiles and config.h files, so those changes should (properly) be applied to Makefile.in and config.h.in instead.
2004-02-16 21:25:18 +01:00
${MV} -f $$file $$file.overridden; \
need to escape ) in the awk regex in do-config-status-override target; some pickier awks (e.g nawk) refuse the regex otherwise
2004-02-28 08:12:21 +01:00
${AWK} '/ *-recheck *\| *--recheck.*\)/ { \
Instead of wrapping the config.status script, edit it directly to prevent --recheck from re-executing the configure script by inserting "exit 0" judiciously. This works with configure scripts generated by autoconf>=2.13 and fixes problems with running config.status on Solaris systems where /bin/sh isn't POSIX enough.
2004-02-27 17:50:48 +01:00
print; \
print " exit 0"; \
next; \
} \
{ print } \
' $$file.overridden > $$file; \
Change into ${WRKSRC} first so that shell globs expand properly.
2004-02-14 16:16:31 +01:00
${CHMOD} +x $$file; \
Default to overriding config.status after GNU configure has been run. We don't want it to execute during the build phase as it might overwrite post-configure changes we might have made to the generated files. Packages that don't want this behaviour can simply set: CONFIG_STATUS_OVERRIDE= # empty in the package Makefile.
2004-02-14 14:54:28 +01:00
fi; \
done
Remove trailing whitespace.
2004-02-12 14:16:02 +01:00
. endfor
.endif
Allow CONFIG_STATUS_OVERRIDE to override config.status scripts generated by GNU configure scripts at CONFIGURE_POSTREQ time. The new config.status scripts merely return success. This prevents newer Makefiles from re-running the configure script with the wrong shell environment if we've touched some GNU autotool-related files during the patch stage. We might want to always do this, i.e. make this opt-out instead of opt-in. However, we start with opt-in so that no existing packages can break.
2004-01-19 16:04:05 +01:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: post-configure
Move code from end of the do-configure target that performed the LIBTOOL_OVERRIDE steps into a separate do-libtool-override target. Create a new variable _CONFIGURE_POSTREQ that currently lists do-libtool-override. Also add some documentation for the two _CONFIGURE_* variables: _CONFIGURE_PREREQ is a list of targets to run after pre-configure but before do-configure. These targets typically edit the files used by the do-configure target. _CONFIGURE_POSTREQ is a list of targets to run after do-configure but before post-configure. These targets typically edit the files generated by the do-configure target that are used during the build phase.
2001-11-30 06:03:06 +01:00
post-configure: ${_CONFIGURE_POSTREQ}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Build
Differentiate between MAKE_FLAGS passed to the build, test and install targets by using BUILD_MAKE_FLAGS, TEST_MAKE_FLAGS, and INSTALL_MAKE_FLAGS in the respective targets. All of the new variables default to ${MAKE_FLAGS} to preserve the current behaviour.
2004-01-13 01:40:25 +01:00
BUILD_DIRS?= ${WRKSRC}
BUILD_MAKE_FLAGS?= ${MAKE_FLAGS}
Add BUILD_DIRS and INSTALL_DIRS, which have the same semantics as CONFIGURE_DIRS, and allows running the shell commands in the do-build and do-install targets within the specified directories. Both variables default to ${WRKSRC}.
2002-09-12 12:34:29 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-build
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(do-build)
do-build:
Add BUILD_DIRS and INSTALL_DIRS, which have the same semantics as CONFIGURE_DIRS, and allows running the shell commands in the do-build and do-install targets within the specified directories. Both variables default to ${WRKSRC}.
2002-09-12 12:34:29 +02:00
. for DIR in ${BUILD_DIRS}
Rename ALL_TARGET to BUILD_TARGET for consistency with other *_TARGETs. Suggested by Roland Illig, ok'd by various.
2004-12-03 16:14:50 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${_ULIMIT_CMD}cd ${DIR} && ${SETENV} ${MAKE_ENV} ${MAKE_PROGRAM} ${BUILD_MAKE_FLAGS} -f ${MAKEFILE} ${BUILD_TARGET}
Add BUILD_DIRS and INSTALL_DIRS, which have the same semantics as CONFIGURE_DIRS, and allows running the shell commands in the do-build and do-install targets within the specified directories. Both variables default to ${WRKSRC}.
2002-09-12 12:34:29 +02:00
. endfor
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
#Test
Differentiate between MAKE_FLAGS passed to the build, test and install targets by using BUILD_MAKE_FLAGS, TEST_MAKE_FLAGS, and INSTALL_MAKE_FLAGS in the respective targets. All of the new variables default to ${MAKE_FLAGS} to preserve the current behaviour.
2004-01-13 01:40:25 +01:00
TEST_DIRS?= ${BUILD_DIRS}
TEST_ENV+= ${MAKE_ENV}
TEST_MAKE_FLAGS?= ${MAKE_FLAGS}
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-test
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
.if !target(do-test)
do-test:
Only run the tests if TEST_TARGET is non-empty.
2003-09-13 08:21:36 +02:00
. if defined(TEST_TARGET) && !empty(TEST_TARGET)
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
. for DIR in ${TEST_DIRS}
Differentiate between MAKE_FLAGS passed to the build, test and install targets by using BUILD_MAKE_FLAGS, TEST_MAKE_FLAGS, and INSTALL_MAKE_FLAGS in the respective targets. All of the new variables default to ${MAKE_FLAGS} to preserve the current behaviour.
2004-01-13 01:40:25 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${_ULIMIT_CMD}cd ${DIR} && ${SETENV} ${TEST_ENV} ${MAKE_PROGRAM} ${TEST_MAKE_FLAGS} -f ${MAKEFILE} ${TEST_TARGET}
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
. endfor
. else
@${DO_NADA}
. endif
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Install
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-install
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(do-install)
do-install:
Add BUILD_DIRS and INSTALL_DIRS, which have the same semantics as CONFIGURE_DIRS, and allows running the shell commands in the do-build and do-install targets within the specified directories. Both variables default to ${WRKSRC}.
2002-09-12 12:34:29 +02:00
. for DIR in ${INSTALL_DIRS}
Differentiate between MAKE_FLAGS passed to the build, test and install targets by using BUILD_MAKE_FLAGS, TEST_MAKE_FLAGS, and INSTALL_MAKE_FLAGS in the respective targets. All of the new variables default to ${MAKE_FLAGS} to preserve the current behaviour.
2004-01-13 01:40:25 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${_ULIMIT_CMD}cd ${DIR} && ${SETENV} ${MAKE_ENV} ${MAKE_PROGRAM} ${INSTALL_MAKE_FLAGS} -f ${MAKEFILE} ${INSTALL_TARGET}
Add BUILD_DIRS and INSTALL_DIRS, which have the same semantics as CONFIGURE_DIRS, and allows running the shell commands in the do-build and do-install targets within the specified directories. Both variables default to ${WRKSRC}.
2002-09-12 12:34:29 +02:00
. endfor
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
# Package
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: real-su-package
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
.if !target(real-su-package)
real-su-package: ${PLIST} ${DESCR}
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Building binary package for ${PKGNAME}"; \
Sanity check that the package is actually installed by running pkg_info(1) before trying to create the binary package. This makes: cd pkgsrc/category/foo make install pkg_delete foo make package fail because package "foo" isn't actually installed, so no binary package can be built.
2003-08-24 10:51:11 +02:00
if ${PKG_INFO} -qe ${PKGNAME}; then \
: The package is installed. ; \
else \
${ECHO_MSG} "=> ${PKGNAME} is not installed."; \
exit 1; \
fi; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
if [ ! -d ${PKGREPOSITORY} ]; then \
${MKDIR} ${PKGREPOSITORY}; \
if [ $$? -ne 0 ]; then \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
${ECHO_MSG} "=> Can't create directory ${PKGREPOSITORY}."; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
exit 1; \
fi; \
fi; \
if ${PKG_CREATE} ${PKG_ARGS_BINPKG} ${PKGFILE}; then \
${MAKE} ${MAKEFLAGS} package-links; \
else \
Print out the exitcode of pkg_create if nonzero, as it sometimes exists without an error message. (Example: editors/abiword on NetBSD-1.6.2/i386.) Approved by jlam.
2005-04-13 18:15:59 +02:00
${ECHO} "(pkg_create: exitcode $$?)"; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${MAKE} ${MAKEFLAGS} delete-package; \
exit 1; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fi
I have no idea where it got broken, but put back the creation of .package_done files.
2001-09-25 22:26:37 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${TOUCH} ${TOUCH_FLAGS} ${PACKAGE_COOKIE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(NO_BIN_ON_CDROM)
Introduce NO_{SRC,BIN}_ON_{FTP,CDROM}. Warn about placing files on CD-ROM or ftp during 'make package' or 'make mirror-distfiles'.
2000-07-20 14:33:43 +02:00
@${ECHO_MSG} "${_PKGSRC_IN}> Warning: ${PKGNAME} may not be put on a CD-ROM:"
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
@${ECHO_MSG} "${_PKGSRC_IN}> " ${NO_BIN_ON_CDROM:Q}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. if defined(NO_BIN_ON_FTP)
Introduce NO_{SRC,BIN}_ON_{FTP,CDROM}. Warn about placing files on CD-ROM or ftp during 'make package' or 'make mirror-distfiles'.
2000-07-20 14:33:43 +02:00
@${ECHO_MSG} "${_PKGSRC_IN}> Warning: ${PKGNAME} may not be made available through FTP:"
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
@${ECHO_MSG} "${_PKGSRC_IN}> " ${NO_BIN_ON_FTP:Q}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Add *RECOMMENDED variables as discussed on tech-pkg@ to allow for a more fine-grained distinction between required versions of pre-requisites (DEPENDS) and versions that are recommended for security or library ABI consistency reasons (RECOMMENDED). The contents of ${RECOMMENDED} are added to DEPENDS unless IGNORE_RECOMMENDED is set to YES, in which case a warning will be printed and IGNORE_RECOMMENDED will be added to BUILD_DEFS. Add a corresponding BUILDLINK_RECOMMENDED.<pkg> variable for use with buildlink2 and buildlink3.
2004-01-14 07:57:45 +01:00
. if defined(RECOMMENDED) && !empty(IGNORE_RECOMMENDED:M[yY][eE][sS])
@${ECHO_MSG} "${_PKGSRC_IN}> Warning: dependency recommendations are being ignored!"
@${ECHO_MSG} "${_PKGSRC_IN}> ${PKGNAME} should not be uploaded nor"
@${ECHO_MSG} "${_PKGSRC_IN}> otherwise be used as a binary package!"
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
# Some support rules for real-su-package
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: package-links
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(package-links)
package-links:
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} delete-package-links
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG}for cat in ${CATEGORIES}; do \
if [ ! -d ${PACKAGES}/$$cat ]; then \
Solaris sh(1) doesn't understand the "if ! command; then..." construct. So dumb-down the shell commands used: make them into "command; if [ $? -ne 0 ]; then..." Suggested by David Brownlee.
1999-03-09 12:14:52 +01:00
${MKDIR} ${PACKAGES}/$$cat; \
if [ $$? -ne 0 ]; then \
In informational messages to the user, use '=>' in preference to '>>', so that cut-n-paste into send-pr will work correctly. Fixes PR 9173 from ITOH Yasufumi.
2000-03-10 17:07:37 +01:00
${ECHO_MSG} "=> Can't create directory ${PACKAGES}/$$cat."; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
exit 1; \
fi; \
fi; \
Remove links in "package-links" target before trying to create them to avoid failure if they already exist. We cannot use "ln -f" here because it doesn't work as expected under Solaris.
2000-06-03 22:58:46 +02:00
${RM} -f ${PACKAGES}/$$cat/${PKGNAME}${PKG_SUFX}; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${LN} -s ../${PKGREPOSITORYSUBDIR}/${PKGNAME}${PKG_SUFX} ${PACKAGES}/$$cat; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
done;
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: delete-package-links
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(delete-package-links)
delete-package-links:
Remove restrictions on category names, and make delete-links target safer. From OpenBSD.
2000-05-30 23:35:32 +02:00
${_PKG_SILENT}${_PKG_DEBUG}\
Use $FIND with "-print". Noted by Georg Schwarz in PR pkg/24248
2004-01-26 22:10:31 +01:00
${FIND} ${PACKAGES} -type l -name ${PKGNAME}${PKG_SUFX} -print | ${XARGS} ${RM} -f
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: delete-package
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(delete-package)
delete-package:
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} delete-package-links
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${RM} -f ${PKGFILE}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: real-su-install
Add automatic ${VARIABLE} handling for MESSAGE files. Convert most MESSAGE files to new syntax (${VARIABLE} gets replaced, not @VARIABLE@, nor @@VARIABLE@@). By default, substitutions are done for LOCALBASE, PKGNAME, PREFIX, X11BASE, X11PREFIX; additional patterns can be added via MESSAGE_SUBST. Clean up some packages while I'm there; add RCS tags to most MESSAGEs. Remove some uninteresting MESSAGEs.
2001-01-29 12:34:21 +01:00
real-su-install: ${MESSAGE}
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if !defined(NO_PKG_REGISTER) && !defined(FORCE_PKG_REGISTER) && \
(${PKG_INSTALLATION_TYPE} == "overwrite")
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(CONFLICTS)
Add PKGPATH, PKGBASE, PKGVERSION and PKGWILDCARD make(1) variables, so that the substitution hacks are all done in one place. Protect a few things with double quotes just to be safe. Fix `readme' target that was left broken in last commit.
2000-06-03 20:51:34 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
${RM} -f ${WRKDIR}/.CONFLICTS
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. for conflict in ${CONFLICTS}
Add PKGPATH, PKGBASE, PKGVERSION and PKGWILDCARD make(1) variables, so that the substitution hacks are all done in one place. Protect a few things with double quotes just to be safe. Fix `readme' target that was left broken in last commit.
2000-06-03 20:51:34 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
found="`${PKG_BEST_EXISTS} \"${conflict}\" || ${TRUE}`"; \
+ Remove the generated ${PLIST} file when doing a re-install. + Enable wildcard matching in DEPENDS and CONFLICTS definitions.
1998-08-25 18:26:09 +02:00
if [ X"$$found" != X"" ]; then \
${ECHO} "$$found" >> ${WRKDIR}/.CONFLICTS; \
fi
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
. endfor
Add PKGPATH, PKGBASE, PKGVERSION and PKGWILDCARD make(1) variables, so that the substitution hacks are all done in one place. Protect a few things with double quotes just to be safe. Fix `readme' target that was left broken in last commit.
2000-06-03 20:51:34 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
if [ -s ${WRKDIR}/.CONFLICTS ]; then \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
found=`${SED} -e s'|${_PKG_DBDIR}/||g' ${WRKDIR}/.CONFLICTS | tr '\012' ' '`; \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> ${PKGNAME} conflicts with installed package(s): $$found found."; \
${ECHO_MSG} "*** They install the same files into the same place."; \
${ECHO_MSG} "*** Please remove $$found first with pkg_delete(1)."; \
+ Remove the generated ${PLIST} file when doing a re-install. + Enable wildcard matching in DEPENDS and CONFLICTS definitions.
1998-08-25 18:26:09 +02:00
${RM} -f ${WRKDIR}/.CONFLICTS; \
exit 1; \
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif # CONFLICTS
Add PKGPATH, PKGBASE, PKGVERSION and PKGWILDCARD make(1) variables, so that the substitution hacks are all done in one place. Protect a few things with double quotes just to be safe. Fix `readme' target that was left broken in last commit.
2000-06-03 20:51:34 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
found="`${PKG_BEST_EXISTS} \"${PKGWILDCARD}\" || ${TRUE}`"; \
Do not overwrite (files of) other pkg versions, warn instead.
1999-02-16 04:23:04 +01:00
if [ "$$found" != "" ]; then \
trim whitespace from a message.
2003-12-30 06:07:41 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> $$found is already installed - perhaps an older version?"; \
* make print-PLIST deal better with /emul/linux/proc * clearer description of possible options when a pkg is already installed ===> glib-1.2.10nb4 is already installed - perhaps an older version? *** If so, you may use either of: *** - "pkg_delete glib-1.2.10nb4" and "make reinstall" to upgrade properly *** - "make update" to rebuild the package and all of its dependencies *** - "make replace" to replace only the package without re-linking *** dependencies, risking various problems. Improved wording welcome.
2003-03-19 02:19:30 +01:00
${ECHO_MSG} "*** If so, you may use either of:"; \
Back out bikeshed material added in rev. 1.1371
2004-02-05 18:20:46 +01:00
${ECHO_MSG} "*** - \"pkg_delete $$found\" and \"${MAKE} reinstall\" to upgrade properly"; \
* make print-PLIST deal better with /emul/linux/proc * clearer description of possible options when a pkg is already installed ===> glib-1.2.10nb4 is already installed - perhaps an older version? *** If so, you may use either of: *** - "pkg_delete glib-1.2.10nb4" and "make reinstall" to upgrade properly *** - "make update" to rebuild the package and all of its dependencies *** - "make replace" to replace only the package without re-linking *** dependencies, risking various problems. Improved wording welcome.
2003-03-19 02:19:30 +01:00
${ECHO_MSG} "*** - \"${MAKE} update\" to rebuild the package and all of its dependencies"; \
${ECHO_MSG} "*** - \"${MAKE} replace\" to replace only the package without re-linking"; \
${ECHO_MSG} "*** dependencies, risking various problems."; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
exit 1; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fi
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.endif # !NO_PKG_REGISTER && !NO_FORCE_REGISTER && overwrite
.if ${PKG_INSTALLATION_TYPE} == "pkgviews"
${_PKG_SILENT}${_PKG_DEBUG} \
found="`${PKG_INFO} -e ${PKGNAME} || ${TRUE}`"; \
if [ "$$found" != "" ]; then \
${ECHO_MSG} "${_PKGSRC_IN}> $$found is already installed."; \
exit 1; \
fi
.endif
Fix two problems when using a ksh(1) as /bin/sh (ln -sf /bin/ksh /bin/sh [*]): * umask handlint due to umask producing different number of leading 0s in ksh(1) and sh(1) * dependency handling is different due to "`...`" being interpreted differently ([*] needed on NetBSD 1.5.1_BETA2/cobalt, as /bin/sh keeps on dumping core in ramdom situations.)
2001-05-20 02:54:09 +02:00
${_PKG_SILENT}${_PKG_DEBUG}if [ `${SH} -c umask` -ne ${DEF_UMASK} ]; then \
message whitespace fix
2005-01-22 10:46:55 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Warning: your umask is \"`${SH} -c umask`"\".; \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "If this is not desired, set it to an appropriate value (${DEF_UMASK})"; \
${ECHO_MSG} "and install this package again by \`\`${MAKE} deinstall reinstall''."; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fi
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
.if defined(PKG_DEVELOPER) && (${CHECK_FILES} == "YES")
${_PKG_SILENT}${_PKG_DEBUG}${ECHO_MSG} "${_PKGSRC_IN}> Generating pre-install file lists"
Some older ${FIND}s do not have an implicit "-print"; make sure to specify it explicitly (as with other ${FIND} usage here).
2005-02-15 17:21:41 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${FIND} ${PREFIX} -type f -or -type l -print \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
2>/dev/null ${CHECK_FILES_SKIP_CMD} >${WRKDIR}/.prefix.pre \
|| ${TRUE}
. if ${CHECK_FILES_STRICT} == "YES"
Some older ${FIND}s do not have an implicit "-print"; make sure to specify it explicitly (as with other ${FIND} usage here).
2005-02-15 17:21:41 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${FIND} ${PKG_SYSCONFDIR} -print \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
2>/dev/null ${CHECK_FILES_SKIP_CMD} >${WRKDIR}/.sysconfdir.pre \
|| ${TRUE}
Some older ${FIND}s do not have an implicit "-print"; make sure to specify it explicitly (as with other ${FIND} usage here).
2005-02-15 17:21:41 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${FIND} ${VARBASE} -print \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
2>/dev/null ${CHECK_FILES_SKIP_CMD} >${WRKDIR}/.varbase.pre \
|| ${TRUE}
. endif
.endif
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if defined(INSTALLATION_DIRS) && !empty(INSTALLATION_DIRS)
${_PKG_SILENT}${_PKG_DEBUG}${ECHO_MSG} "${_PKGSRC_IN}> Creating installation directories"
${_PKG_SILENT}${_PKG_DEBUG} \
for dir in ${INSTALLATION_DIRS}; do \
case $$dir in \
/*) ;; \
*bin|*bin/*|*libexec|*libexec/*) \
${INSTALL_PROGRAM_DIR} ${PREFIX}/$$dir ;; \
*man/*) \
${INSTALL_MAN_DIR} ${PREFIX}/$$dir ;; \
*) \
${INSTALL_DATA_DIR} ${PREFIX}/$$dir ;; \
esac; \
done
.endif # INSTALLATION_DIRS
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !defined(NO_MTREE)
Pay attention to the setting of $ROOT_USER when checking whether we're root.
2003-12-03 19:25:05 +01:00
${_PKG_SILENT}${_PKG_DEBUG}if [ `${ID} -u` = `${ID} -u ${ROOT_USER}` ]; then \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
if [ ! -f ${MTREE_FILE} ]; then \
- Add OpenBSD and FreeBSD Ids in the header - Changes from FreeBSD's V1.164: * Warn if mtree-file not found * On 'make reinstall' gets DEPENDS_TARGET set to 'reinstall' instead of 'install' - Changes from OpenBSD's V1.14: * Documentation and implementation for ONLY_FOR_ARCHS * Documentation and implementation for WRKOBJDIR * Documentation on HAVE_MOTIF and MOTIF_STATIC: set in /etc/mk.conf, not /etc/make.conf * Set NOMANCOMPRESS=no to still compress man-pages (should probably be set to 'yes') * DEF_UMASK=022 * add ${MACHINE} suffix to WRKDIR if OBJMACHINE is set * .include <bsd.own.mk>
1997-09-28 02:44:15 +02:00
${ECHO_MSG} "Error: mtree file \"${MTREE_FILE}\" is missing."; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
exit 1; \
else \
if [ ! -d ${PREFIX} ]; then \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
${MKDIR} ${PREFIX}; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
fi; \
Linux bits; now the package system works on Linux too.
1999-06-23 19:06:20 +02:00
${MTREE} ${MTREE_ARGS} ${PREFIX}/; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
fi; \
else \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${ECHO_MSG} "Warning: not superuser, can't run mtree."; \
ROOT_USER is parameterized, so use it in human-readable printed messages as well.
2004-04-26 05:07:46 +02:00
${ECHO_MSG} "Become ${ROOT_USER} and try again to ensure correct permissions."; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fi
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.else
Test for ${PREFIX} existence before mkdir'ing it. This fix NO_MTREE installation on Solaris when ${PREFIX} already exists and is a symlink. Example: # uname -sr SunOS 5.9 # cd /tmp # rm -f foo # ln -s /etc foo # /bin/mkdir -p /tmp/foo mkdir: "/tmp/foo": Exists but is not a directory #
2003-09-02 14:01:14 +02:00
${_PKG_SILENT}${_PKG_DEBUG}[ -d ${PREFIX} ] || ${MKDIR} ${PREFIX}
+ Factor out the real-install target from the rest of the .USE ones. This makes it possible to do the next step. There were so many .if make(install) checks that this arguably should be done anyway. + In interactive installs, if the effective uid is not zero when "make install" is attempted, then ${SU_CMD} is invoked (defaults to "su - root -c", but sudo or priv could be used). This allows running as non-root for much more time, which is a significant benefit. This also allows installation of pre-requisite (DEPENDS) packages while building a package as non-root user. + A hook has been added to run ${PRE_ROOT_CMD} before becoming root, so that root can be given write access to ${WRKDIR} directories, for example, on -noroot mounted NFS filesystems.
1999-03-31 11:04:18 +02:00
.endif # !NO_MTREE
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
.if !defined(NO_PKG_REGISTER)
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} pre-install-fake-pkg
.endif # !NO_PKG_REGISTER
Change PKG_CREATE_USERGROUP and PKG_CONFIG so that their values are merely defaults for the INSTALL/DEINSTALL scripts. The environment variables of the same name take precedence during installation of a binary package.
2002-06-22 04:48:35 +02:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} pre-install-script
MAKE_ENV is the make environment passed down to the build process. It shouldn't be set as the environment for a make process invoked at the package (${.CURDIR}) level during the install phase because it can contain harmful settings.
2002-01-27 20:07:41 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} pre-install
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} do-install
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} post-install
Use ${DO_NADA} to introduce a do-nothing line (as is done elsewhere in pkgsrc); simply # makes older bmake hork up in combination with some shells.
2004-03-11 20:37:20 +01:00
${_PKG_SILENT}${DO_NADA} \
# \
# PLIST must be generated at this late point (instead of \
# depending on it somewhere earlier), because it needs \
# to be created _after_ the {pre,do,post}-install \
# targets are run. \
# \
# We generate _before_ post-install-script is run so \
# that the real config files and rc.d scripts aren't \
# listed in the PLIST. \
#
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} ${PLIST}
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
.if defined(PKG_DEVELOPER) && (${CHECK_FILES} == "YES")
${_PKG_SILENT}${_PKG_DEBUG}${ECHO_MSG} "${_PKGSRC_IN}> Generating post-install file lists"
Some older ${FIND}s do not have an implicit "-print"; make sure to specify it explicitly (as with other ${FIND} usage here).
2005-02-15 17:21:41 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${FIND} ${PREFIX} -type f -or -type l -print \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
2>/dev/null ${CHECK_FILES_SKIP_CMD} >${WRKDIR}/.prefix.post \
|| ${TRUE}
. if ${CHECK_FILES_STRICT} == "YES"
Some older ${FIND}s do not have an implicit "-print"; make sure to specify it explicitly (as with other ${FIND} usage here).
2005-02-15 17:21:41 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${FIND} ${PKG_SYSCONFDIR} -print \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
2>/dev/null ${CHECK_FILES_SKIP_CMD} >${WRKDIR}/.sysconfdir.post\
|| ${TRUE}
Some older ${FIND}s do not have an implicit "-print"; make sure to specify it explicitly (as with other ${FIND} usage here).
2005-02-15 17:21:41 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${FIND} ${VARBASE} -print \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
2>/dev/null ${CHECK_FILES_SKIP_CMD} >${WRKDIR}/.varbase.post \
|| ${TRUE}
. endif
.endif
Change PKG_CREATE_USERGROUP and PKG_CONFIG so that their values are merely defaults for the INSTALL/DEINSTALL scripts. The environment variables of the same name take precedence during installation of a binary package.
2002-06-22 04:48:35 +02:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} post-install-script
Remove a bunch of unneeded ()s. Inspired by OpenBSD.
2000-06-03 01:30:52 +02:00
${_PKG_SILENT}${_PKG_DEBUG}newmanpages=`${EGREP} -h \
Do not include lines like @exec ln -fs dlopen.3.gz %D/emul/linux/usr/man/man3/dlclose.3.gz into the list of manpages to decompress (for automatic manpage compression handling. Bug hit in the emulators/suse_base package, hashed out by tron.
1999-09-21 14:53:14 +02:00
'^([^@/]*/)*man/([^/]*/)?(man[1-9ln]/.*\.[1-9ln]|cat[1-9ln]/.*\.0)(\.gz)?$$' \
Be quiet about finding no gzipped man pages in PLIST.
2000-02-12 18:24:41 +01:00
${PLIST} 2>/dev/null || ${TRUE}`; \
Add automatic manual page handling. This is backwards-compatible with the MANx and CATx definitions in package Makefiles - if neither MANx or CATx is specified, the PLIST will be searched for the package's manual pages. If automatic manual page handling is in operation, and PKG_VERBOSE is set, echo the manual page names.
1998-06-18 13:45:33 +02:00
if [ X"${MANCOMPRESSED}" != X"" -a X"${MANZ}" = X"" ]; then \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> [Automatic manual page handling]"; \
${ECHO_MSG} "${_PKGSRC_IN}> Decompressing manual pages for ${PKGNAME}"; \
Use one gonzo regular expression to recognise man and catman pages.
1998-06-22 11:12:48 +02:00
for manpage in $$newmanpages; do \
Add automatic manual page handling. This is backwards-compatible with the MANx and CATx definitions in package Makefiles - if neither MANx or CATx is specified, the PLIST will be searched for the package's manual pages. If automatic manual page handling is in operation, and PKG_VERBOSE is set, echo the manual page names.
1998-06-18 13:45:33 +02:00
manpage=`${ECHO} $$manpage | ${SED} -e 's|\.gz$$||'`; \
Use -h instead of -L in test(1) when testing for a symbolic link.
1999-08-24 21:03:08 +02:00
if [ -h ${PREFIX}/$$manpage.gz ]; then \
Use "${LS}" instead of "${FILE_CMD}" to find out the target of softlink when (de)compressing manual pages. This fixes problems caused by the output format of NetBSD-current's and shouldn't hurt any other platform.
2003-12-28 22:20:02 +01:00
set - `${LS} -l ${PREFIX}/$$manpage.gz | ${SED} -e 's|\.gz$$||'`; \
Handle decompression of symlinked manual pages.
1998-11-19 23:59:03 +01:00
shift `expr $$# - 1`; \
Improve ELF shared lib handling, so that it now works on Solaris. In the process, it was found that the Solaris ln(1) doesn't work quite the way its manual page claims it does. In particular, "ln -fs a b" will fail if b already exists, rather than deleting the link. Work around this by using an explicit rm(1) prior to calling "ln -s".
1999-03-09 16:31:59 +01:00
${RM} -f ${PREFIX}/$$manpage; \
${LN} -s $${1} ${PREFIX}/$$manpage; \
Handle decompression of symlinked manual pages.
1998-11-19 23:59:03 +01:00
${RM} ${PREFIX}/$$manpage.gz; \
else \
${GUNZIP_CMD} ${PREFIX}/$$manpage.gz; \
fi; \
Add automatic manual page handling. This is backwards-compatible with the MANx and CATx definitions in package Makefiles - if neither MANx or CATx is specified, the PLIST will be searched for the package's manual pages. If automatic manual page handling is in operation, and PKG_VERBOSE is set, echo the manual page names.
1998-06-18 13:45:33 +02:00
if [ X"${PKG_VERBOSE}" != X"" ]; then \
${ECHO_MSG} "$$manpage"; \
fi; \
done; \
fi; \
if [ X"${MANCOMPRESSED}" = X"" -a X"${MANZ}" != X"" ]; then \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> [Automatic manual page handling]"; \
${ECHO_MSG} "${_PKGSRC_IN}> Compressing manual pages for ${PKGNAME}"; \
Use one gonzo regular expression to recognise man and catman pages.
1998-06-22 11:12:48 +02:00
for manpage in $$newmanpages; do \
Add automatic manual page handling. This is backwards-compatible with the MANx and CATx definitions in package Makefiles - if neither MANx or CATx is specified, the PLIST will be searched for the package's manual pages. If automatic manual page handling is in operation, and PKG_VERBOSE is set, echo the manual page names.
1998-06-18 13:45:33 +02:00
manpage=`${ECHO} $$manpage | ${SED} -e 's|\.gz$$||'`; \
Use -h instead of -L in test(1) when testing for a symbolic link.
1999-08-24 21:03:08 +02:00
if [ -h ${PREFIX}/$$manpage ]; then \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
set - `${LS} -l ${PREFIX}/$$manpage`; \
Add automatic manual page handling. This is backwards-compatible with the MANx and CATx definitions in package Makefiles - if neither MANx or CATx is specified, the PLIST will be searched for the package's manual pages. If automatic manual page handling is in operation, and PKG_VERBOSE is set, echo the manual page names.
1998-06-18 13:45:33 +02:00
shift `expr $$# - 1`; \
Improve ELF shared lib handling, so that it now works on Solaris. In the process, it was found that the Solaris ln(1) doesn't work quite the way its manual page claims it does. In particular, "ln -fs a b" will fail if b already exists, rather than deleting the link. Work around this by using an explicit rm(1) prior to calling "ln -s".
1999-03-09 16:31:59 +01:00
${RM} -f ${PREFIX}/$$manpage.gz; \
${LN} -s $${1}.gz ${PREFIX}/$$manpage.gz; \
Add automatic manual page handling. This is backwards-compatible with the MANx and CATx definitions in package Makefiles - if neither MANx or CATx is specified, the PLIST will be searched for the package's manual pages. If automatic manual page handling is in operation, and PKG_VERBOSE is set, echo the manual page names.
1998-06-18 13:45:33 +02:00
${RM} ${PREFIX}/$$manpage; \
else \
${GZIP_CMD} ${PREFIX}/$$manpage; \
fi; \
if [ X"${PKG_VERBOSE}" != X"" ]; then \
${ECHO_MSG} "$$manpage"; \
fi; \
done; \
Remove a bunch of unneeded ()s. Inspired by OpenBSD.
2000-06-03 01:30:52 +02:00
fi
Add a _DO_SHLIB_CHECKS definition to the individual defs.${OPSYS}.mk files, and use it in bsd.pkg.mk. This definition controls the fixup of PLISTs at installation time, and running of ldconfig if necessary.
2001-11-21 14:17:14 +01:00
.if ${_DO_SHLIB_CHECKS} == "yes"
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
. if ${PKG_INSTALLATION_TYPE} == "overwrite"
Introduce a PKGSRC_MESSAGE_RECIPIENTS, which takes the login names of users to whom the MESSAGE file should be mailed at package installation time, and mail the MESSAGE file at the "make install" stage (if PKGSRC_MESSAGE_RECIPIENTS is not empty). Inspired by PR 22328 from Jeremy Reed.
2003-08-07 18:14:05 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
${MAKE} ${MAKEFLAGS} do-shlib-handling SHLIB_PLIST_MODE=0
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
. endif
Add a _DO_SHLIB_CHECKS definition to the individual defs.${OPSYS}.mk files, and use it in bsd.pkg.mk. This definition controls the fixup of PLISTs at installation time, and running of ldconfig if necessary.
2001-11-21 14:17:14 +01:00
.endif
Add automatic ${VARIABLE} handling for MESSAGE files. Convert most MESSAGE files to new syntax (${VARIABLE} gets replaced, not @VARIABLE@, nor @@VARIABLE@@). By default, substitutions are done for LOCALBASE, PKGNAME, PREFIX, X11BASE, X11PREFIX; additional patterns can be added via MESSAGE_SUBST. Clean up some packages while I'm there; add RCS tags to most MESSAGEs. Remove some uninteresting MESSAGEs.
2001-01-29 12:34:21 +01:00
.ifdef MESSAGE
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> Please note the following:"
@${ECHO_MSG} ""
Add automatic ${VARIABLE} handling for MESSAGE files. Convert most MESSAGE files to new syntax (${VARIABLE} gets replaced, not @VARIABLE@, nor @@VARIABLE@@). By default, substitutions are done for LOCALBASE, PKGNAME, PREFIX, X11BASE, X11PREFIX; additional patterns can be added via MESSAGE_SUBST. Clean up some packages while I'm there; add RCS tags to most MESSAGEs. Remove some uninteresting MESSAGEs.
2001-01-29 12:34:21 +01:00
@${CAT} ${MESSAGE}
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
@${ECHO_MSG} ""
Introduce a PKGSRC_MESSAGE_RECIPIENTS, which takes the login names of users to whom the MESSAGE file should be mailed at package installation time, and mail the MESSAGE file at the "make install" stage (if PKGSRC_MESSAGE_RECIPIENTS is not empty). Inspired by PR 22328 from Jeremy Reed.
2003-08-07 18:14:05 +02:00
. if !empty(PKGSRC_MESSAGE_RECIPIENTS)
${_PKG_SILENT}${_PKG_DEBUG} \
(${ECHO} "The ${PKGNAME} package was installed on `${HOSTNAME_CMD}` at `date`"; \
${ECHO} ""; \
${ECHO} "Please note the following:"; \
${ECHO} ""; \
${CAT} ${MESSAGE}; \
${ECHO} "") | \
add hostname to Subject: of PKGSRC_MESSAGE_RECIPIENTS email
2004-10-26 04:23:37 +02:00
${MAIL_CMD} -s"Package ${PKGNAME} installed on `${HOSTNAME_CMD}`" ${PKGSRC_MESSAGE_RECIPIENTS}
Introduce a PKGSRC_MESSAGE_RECIPIENTS, which takes the login names of users to whom the MESSAGE file should be mailed at package installation time, and mail the MESSAGE file at the "make install" stage (if PKGSRC_MESSAGE_RECIPIENTS is not empty). Inspired by PR 22328 from Jeremy Reed.
2003-08-07 18:14:05 +02:00
. endif
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
.endif
.if !defined(NO_PKG_REGISTER)
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} register-pkg
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
.endif # !NO_PKG_REGISTER
${_PKG_SILENT}${_PKG_DEBUG}${TOUCH} ${TOUCH_FLAGS} ${INSTALL_COOKIE}
.if defined(PKG_DEVELOPER) && (${CHECK_SHLIBS} == "YES")
Use '${_PKG_SILENT}${_PKG_DEBUG}' instead of '@' (as seen on the pkgviews-mk branch).
2003-08-23 00:04:03 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} check-shlibs
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
.endif
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
.if defined(PKG_DEVELOPER) && (${CHECK_FILES} == "YES")
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} check-files
.endif
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
# Do handling of shared libs for two cases:
#
# SHLIB_PLIST_MODE=1: when first called via the ${PLIST} target,
# update the PLIST to contain ELF symlink, run
# ldconfig on a.out, etc. (used when called via
# the ${PLIST} target). Will update ${PLIST}.
# SHLIB_PLIST_MODE=0: when called via the real-su-install target,
# actually generate symlinks for ELF, run ldconfig
# for a.out, etc. Will not modify ${PLIST}.
#
# XXX This target could need some cleanup after it was ripped out of
# real-su-install
#
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
_AOUT_AWK = \
BEGIN { linkc = 1 } \
/^@/ { lines[NR] = $$0; next } \
function libtool_release(lib) { \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (gsub("-[^-]+\\.so\\.", ".so.", lib)) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
if (system("${TEST} -h ${PREFIX}/" lib) == 0) { \
rels[NR] = lib; \
} \
} \
} \
/.*\/lib[^\/]+\.so\.[0-9]+\.[0-9]+\.[0-9]+$$/ { \
libtool_release($$0); \
lines[NR] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (sub("-[^-]+\\.so$$", ".so")) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
} \
next \
} \
/.*\/lib[^\/]+\.so\.[0-9]+\.[0-9]+$$/ { \
libtool_release($$0); \
lines[NR] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (sub("-[^-]+\\.so$$", ".so")) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
} \
next \
} \
{ lines[NR] = $$0 } \
END { \
for (i = 0 ; i <= linkc ; i++) \
for (j = 1 ; j < NR ; j++) \
if (lines[j] == links[i]) \
lines[j] = "@comment " lines[j]; \
if (${SHLIB_PLIST_MODE}) \
for (i = 1 ; i <= NR ; i++) { \
print lines[i]; \
if (rels[i] != "") \
print rels[i]; \
} \
}
_DYLIB_AWK= \
/^@/ { lines[NR] = $$0; next } \
function libtool_release(lib) { \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (gsub("\\.so\\.", ".", lib) || gsub("\\.so$$", "", lib)) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
lib = lib ".dylib"; \
if (system("${TEST} -h ${PREFIX}/" lib) == 0) { \
rels[NR] = lib; \
} \
} \
} \
/.*\/lib[^\/]+\.so\.[0-9]+\.[0-9]+\.[0-9]+$$/ { \
libtool_release($$0); \
lines[NR] = $$0; \
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (sub("-[^-]+\\.so$$", ".so")) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
} \
next \
} \
/.*\/lib[^\/]+\.so\.[0-9]+\.[0-9]+$$/ { \
libtool_release($$0); \
lines[NR] = $$0; \
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (sub("-[^-]+\\.so$$", ".so")) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
} \
next \
} \
/.*\/lib[^\/]+\.so\.[0-9]+$$/ { \
libtool_release($$0); \
lines[NR] = $$0; \
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (sub("-[^-]+\\.so$$", ".so")) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
} \
next \
} \
/.*\/lib[^\/]+\.so$$/ { \
lines[NR] = $$0; \
replace use of ${TEST} -e with -f or -d as appropriate. use consistent shell syntax. this allows the pkgsrc framework to work with IRIX 5.3 and Solaris' sh(1) built-in test(1).
2004-07-03 23:59:04 +02:00
if (system("${TEST} -f ${PREFIX}/" $$0) == 0) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
next; \
} \
libtool_release($$0); \
links[linkc++] = $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (sub("-[^-]+\\.so$$", ".so")) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
links[linkc++] = $$0; \
} \
next \
} \
{ lines[NR] = $$0 } \
END { \
for (i = 0 ; i <= linkc ; i++) \
for (j = 1 ; j <= NR ; j++) \
if (lines[j] == links[i]) \
lines[j] = "@comment " lines[j]; \
if (${SHLIB_PLIST_MODE}) \
for (i = 1 ; i <= NR ; i++) { \
print lines[i]; \
if (rels[i] != "") { \
print rels[i]; \
fix a file descriptor leak in the darwin PLIST so->dylib translation script that broke packages that installed too many shared libraries (eg libwww).
2004-03-17 17:36:28 +01:00
cmd = "${LS} -l ${PREFIX}/" rels[i]; \
cmd | getline tgt; \
close(cmd); \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
gsub(".* ", "", tgt); \
if (tgts[tgt] == "") { \
tgts[tgt] = tgt; \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
if (index(tgt, "/") == 1) \
print tgt; \
else { \
prefix=""; \
if (match(rels[i], ".*/") != 0) \
prefix=substr(rels[i],1,RLENGTH); \
print prefix tgt; \
} \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
} \
} \
} \
}
update the PLIST handling for AIX, shared libraries have the .so extension when the runtime linker is used. from garbled@.
2005-02-19 05:25:31 +01:00
# Turn lib*.so.*, lib*.so into lib*.so. Drop duplicates.
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
_AIXLIB_AWK= \
/^@/ { lines[NR] = $$0; next } \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
/.*\/lib[^\/]+\.so(\.[0-9]+)*$$/ { \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("(\\.[0-9]+)*$$", ""); \
update the PLIST handling for AIX, shared libraries have the .so extension when the runtime linker is used. from garbled@.
2005-02-19 05:25:31 +01:00
sub("\\.so$$", ".so"); \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
lines[NR] = $$0; \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
next \
} \
{ lines[NR] = $$0 } \
END { \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
nlibs = 0; \
for (i = 1; i <= NR; i++) { \
for (j = 0; j < nlibs; j++) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
if (libs[j] == lines[i]) \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
break; \
} \
if (j >= nlibs) \
print lines[i]; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
if (match(lines[i], ".*/lib[^/]+\\.a$$")) { \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
libs[nlibs] = lines[i]; \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
nlibs++; \
} \
} \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
}
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-shlib-handling
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
do-shlib-handling:
Fix handling of SHLIB_HANDLING=NO. That is neither stage of do-shlib- handling should be performed. This fixes the build problems for emulators/compat1[23].
2000-12-31 13:21:07 +01:00
.if ${SHLIB_HANDLING} == "YES"
Remove a bunch of unneeded ()s. Inspired by OpenBSD.
2000-06-03 01:30:52 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
egrep -x 'RE' is the same as egrep '^RE$' Not all egrep's know the '-x' flag, so use the latter.
2004-04-11 22:17:37 +02:00
sos=`${EGREP} -h '^.*/lib[^/]+\.so$$' ${PLIST} || ${TRUE}`; \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
if [ "$$sos" != "" ]; then \
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
shlib_type=`${MAKE} ${MAKEFLAGS} show-shlib-type`; \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
if [ "${SHLIB_PLIST_MODE}" = "0" ]; then \
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> [Automatic $$shlib_type shared object handling]"; \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
fi; \
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
case "$$shlib_type" in \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
ELF) ;; \
Fixes for shlib handling on AIX
2005-03-29 10:17:42 +02:00
aixlib) ;; \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
"a.out") \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
${AWK} '${_AOUT_AWK}' <${PLIST} >${PLIST}.tmp ; \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
if [ "${SHLIB_PLIST_MODE}" = "1" ]; then \
${MV} ${PLIST}.tmp ${PLIST}; \
else \
${RM} ${PLIST}.tmp ; \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
fi ; \
egrep -x 'RE' is the same as egrep '^RE$' Not all egrep's know the '-x' flag, so use the latter.
2004-04-11 22:17:37 +02:00
cnt=`${EGREP} -c '^@exec[ ]*${LDCONFIG}$$' ${PLIST} || ${TRUE}`; \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
if [ "${SHLIB_PLIST_MODE}" = "1" ]; then \
if [ $$cnt -eq 0 ]; then \
${ECHO} "@exec ${LDCONFIG}" >> ${PLIST}; \
${ECHO} "@unexec ${LDCONFIG}" >> ${PLIST}; \
fi \
Add logic to handle shared objects correctly for ELF and a.out platforms.
1998-07-14 13:36:11 +02:00
fi; \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
if [ "${SHLIB_PLIST_MODE}" = "0" ]; then \
if [ "${PKG_VERBOSE}" != "" ]; then \
${ECHO_MSG} "$$sos"; \
${ECHO_MSG} "Running ${LDCONFIG}"; \
fi; \
${LDCONFIG} || ${TRUE}; \
fi \
Fill in SHLIB_TYPE in the pkgsrc copy of bsd.own.mk. Use this value in bsd.pkg.mk to work out the type of shared objects when doing automatic shared object handling (ELF, a.out or none).
1998-07-14 17:53:54 +02:00
;; \
Add support for OS X-style dylibs when creating the derived PLIST file.
2001-12-05 17:42:46 +01:00
"dylib") \
Whitespace fixes, no functional changes.
2003-12-04 12:17:40 +01:00
${AWK} '${_DYLIB_AWK}' <${PLIST} >${PLIST}.tmp && \
Add support for OS X-style dylibs when creating the derived PLIST file.
2001-12-05 17:42:46 +01:00
if [ "${SHLIB_PLIST_MODE}" = "1" ]; then \
${MV} ${PLIST}.tmp ${PLIST}; \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
else \
Add support for OS X-style dylibs when creating the derived PLIST file.
2001-12-05 17:42:46 +01:00
${RM} ${PLIST}.tmp ; \
Add another shlib type "aixlib". Handle this in do-shlib-handling: turn *.so into *.a. Pull out the inline awk scripts into variables to make the actual target a little bit clearer.
2003-12-03 19:31:06 +01:00
fi ; \
Add support for OS X-style dylibs when creating the derived PLIST file.
2001-12-05 17:42:46 +01:00
;; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
"*") \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
if [ "${SHLIB_PLIST_MODE}" = "0" ]; then \
${ECHO_MSG} "No shared libraries for ${MACHINE_ARCH}"; \
fi ; \
if [ "${SHLIB_PLIST_MODE}" = "1" ]; then \
for so in $$sos; do \
if [ X"${PKG_VERBOSE}" != X"" ]; then \
${ECHO_MSG} >&2 "Ignoring $$so"; \
fi; \
${SED} -e "s;^$$so$$;@comment No shared objects - &;" \
Don't do input redirection for sed if not necessary.
2001-01-29 15:40:26 +01:00
${PLIST} >${PLIST}.tmp && ${MV} ${PLIST}.tmp ${PLIST}; \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
done; \
fi ; \
Fill in SHLIB_TYPE in the pkgsrc copy of bsd.own.mk. Use this value in bsd.pkg.mk to work out the type of shared objects when doing automatic shared object handling (ELF, a.out or none).
1998-07-14 17:53:54 +02:00
;; \
esac; \
Remove a bunch of unneeded ()s. Inspired by OpenBSD.
2000-06-03 01:30:52 +02:00
fi
Fix handling of SHLIB_HANDLING=NO. That is neither stage of do-shlib- handling should be performed. This fixes the build problems for emulators/compat1[23].
2000-12-31 13:21:07 +01:00
.endif # SHLIB_HANDLING == "YES"
+ Factor out the real-install target from the rest of the .USE ones. This makes it possible to do the next step. There were so many .if make(install) checks that this arguably should be done anyway. + In interactive installs, if the effective uid is not zero when "make install" is attempted, then ${SU_CMD} is invoked (defaults to "su - root -c", but sudo or priv could be used). This allows running as non-root for much more time, which is a significant benefit. This also allows installation of pre-requisite (DEPENDS) packages while building a package as non-root user. + A hook has been added to run ${PRE_ROOT_CMD} before becoming root, so that root can be given write access to ${WRKDIR} directories, for example, on -noroot mounted NFS filesystems.
1999-03-31 11:04:18 +02:00
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
# Check if all binaries and shlibs find their needed libs
# Must be run after "make install", so that files are installed, and
# ${PLIST} exists.
#
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: check-shlibs
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
check-shlibs:
check-shlibs: don't complain about uninstalled pkg on pkgs that have NO_PKG_REGISTER set (like pkg_install)
2001-03-07 01:26:26 +01:00
.if !defined(NO_PKG_REGISTER)
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Replace subshell commands in (...) by {...;} for speed. Based on patch from Chris Coleman <chrisc@vmunix.com>. Timing of "make show-var VARNAME=COMMENT" in pkgsrc/x11: new: 2.744u 0.888s 0:16.69 21.6% 0+0k 809+201io 55pf+0w 2.827u 1.047s 0:07.72 50.0% 0+0k 0+189io 0pf+0w 2.913u 0.992s 0:06.91 56.4% 0+0k 0+168io 0pf+0w 2.978u 0.997s 0:06.88 57.5% 0+0k 0+169io 0pf+0w 3.017u 0.886s 0:07.88 49.3% 0+0k 0+186io 0pf+0w old: 2.921u 0.987s 0:07.04 55.3% 0+0k 2+183io 0pf+0w 2.931u 1.051s 0:08.49 46.8% 0+0k 0+169io 0pf+0w 3.065u 0.816s 0:08.33 46.4% 0+0k 0+174io 0pf+0w 2.962u 0.987s 0:08.08 48.7% 0+0k 0+181io 0pf+0w 2.661u 1.228s 0:08.50 45.6% 0+0k 0+171io 0pf+0w
2001-04-03 14:06:31 +02:00
bins=`${PKG_INFO} -qL ${PKGNAME} | { ${EGREP} -h '/(bin|sbin|libexec)/' || ${TRUE}; }`; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
if [ "${OBJECT_FMT}" = "ELF" ]; then \
Replace subshell commands in (...) by {...;} for speed. Based on patch from Chris Coleman <chrisc@vmunix.com>. Timing of "make show-var VARNAME=COMMENT" in pkgsrc/x11: new: 2.744u 0.888s 0:16.69 21.6% 0+0k 809+201io 55pf+0w 2.827u 1.047s 0:07.72 50.0% 0+0k 0+189io 0pf+0w 2.913u 0.992s 0:06.91 56.4% 0+0k 0+168io 0pf+0w 2.978u 0.997s 0:06.88 57.5% 0+0k 0+169io 0pf+0w 3.017u 0.886s 0:07.88 49.3% 0+0k 0+186io 0pf+0w old: 2.921u 0.987s 0:07.04 55.3% 0+0k 2+183io 0pf+0w 2.931u 1.051s 0:08.49 46.8% 0+0k 0+169io 0pf+0w 3.065u 0.816s 0:08.33 46.4% 0+0k 0+174io 0pf+0w 2.962u 0.987s 0:08.08 48.7% 0+0k 0+181io 0pf+0w 2.661u 1.228s 0:08.50 45.6% 0+0k 0+171io 0pf+0w
2001-04-03 14:06:31 +02:00
shlibs=`${PKG_INFO} -qL ${PKGNAME} | { ${EGREP} -h '/lib/lib.*.so' || ${TRUE}; }`; \
add support for extracting binary/library requires/provides on Darwin, and simplify the logic to reduce code duplication. fixes PR pkg/22492 from Julien T. Letessier.
2003-08-23 11:11:19 +02:00
elif [ "${OBJECT_FMT}" = "Mach-O" ]; then \
shlibs=`${PKG_INFO} -qL ${PKGNAME} | { ${EGREP} -h '/lib/lib.*.dylib' || ${TRUE}; }`; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
else \
shlibs=""; \
fi; \
add support for extracting binary/library requires/provides on Darwin, and simplify the logic to reduce code duplication. fixes PR pkg/22492 from Julien T. Letessier.
2003-08-23 11:11:19 +02:00
if [ "X${LDD}" = X ]; then \
in check-shlibs, throw away warning from `type ldd' and only attempt to run ldd if it is executable. fixes bug where binaries just installed would be run (!) and the install process possibly "hang" on Darwin because Darwin has no ldd!
2003-08-22 15:32:40 +02:00
ldd=`${TYPE} ldd 2>/dev/null | ${AWK} '{ print $$NF }'`;\
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
else \
ldd="${LDD}"; \
fi; \
in check-shlibs, throw away warning from `type ldd' and only attempt to run ldd if it is executable. fixes bug where binaries just installed would be run (!) and the install process possibly "hang" on Darwin because Darwin has no ldd!
2003-08-22 15:32:40 +02:00
if [ -x "$$ldd" ]; then \
for i in $${bins} $${shlibs}; do \
err=`{ $$ldd $$i 2>&1 || ${TRUE}; } | { ${GREP} "not found" || ${TRUE}; }`; \
if [ "${PKG_VERBOSE}" != "" ]; then \
${ECHO} "$$ldd $$i"; \
fi; \
if [ "$$err" != "" ]; then \
${ECHO} "$$i: $$err"; \
error=1; \
fi; \
done; \
fi; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
if [ "$$error" = 1 ]; then \
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
${ECHO} "*** The above programs/libs will not find the listed shared libraries"; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${ECHO} " at runtime. Please fix the package (add -Wl,-R.../lib in the right places)!"; \
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
${SHCOMMENT} Might not error-out for non-pkg-developers; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
exit 1; \
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
fi
check-shlibs: don't complain about uninstalled pkg on pkgs that have NO_PKG_REGISTER set (like pkg_install)
2001-03-07 01:26:26 +01:00
.endif # NO_PKG_REGISTER
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
.if !target(show-shlib-type)
# Show the shared lib type being built: one of ELF, a.out, dylib, or none
.PHONY: show-shlib-type
show-shlib-type:
Allow an empty USE_LANGUAGES setting to not break in show-shlib-type. If we're not compiling anything, we're likely not installing any shared libraries.
2004-02-06 04:10:29 +01:00
. if empty(USE_LANGUAGES)
Don't echo the debug message for the targets whose output is used by other targets. OK'd by jlam. This should close pkg/24377.
2004-10-03 06:24:20 +02:00
@${ECHO} "none"
Get rid of all occurrences of cpp-like tests for ${OPSYS} in bsd.pkg.mk, and use the abstracted means of determining generic shared lib type per operating system.
2004-03-11 23:57:22 +01:00
. elif ${_OPSYS_SHLIB_TYPE} == "ELF/a.out"
Don't echo the debug message for the targets whose output is used by other targets. OK'd by jlam. This should close pkg/24377.
2004-10-03 06:24:20 +02:00
@cd ${WRKDIR} && \
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
sotype=none; \
if [ "X${MKPIC}" != "Xno" -a "X${NOPIC}" = "X" ]; then \
show-shlib-type: use return(0) instead of exit(0). eliminates a warning from non-gcc. ok'd by agc.
2004-01-12 11:08:37 +01:00
${ECHO} "int main() { return(0); }" > a.$$$$.c; \
Pass the PATH when executing ${CC} so that the compiler may be found when CC is something like ccache or distcc that relies on finding the true compiler in the PATH.
2004-02-05 08:37:48 +01:00
${SETENV} PATH=${PATH} \
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
${CC} ${CFLAGS} a.$$$$.c -o a.$$$$.out; \
case `${FILE_CMD} a.$$$$.out` in \
*ELF*dynamically*) \
sotype=ELF ;; \
*shared*library*) \
sotype="a.out" ;; \
*dynamically*) \
sotype="a.out" ;; \
esac; \
fi; \
${ECHO} "$$sotype"; \
${RM} -f a.$$$$.c a.$$$$.out
Get rid of all occurrences of cpp-like tests for ${OPSYS} in bsd.pkg.mk, and use the abstracted means of determining generic shared lib type per operating system.
2004-03-11 23:57:22 +01:00
. else
Don't echo the debug message for the targets whose output is used by other targets. OK'd by jlam. This should close pkg/24377.
2004-10-03 06:24:20 +02:00
@${ECHO} ${_OPSYS_SHLIB_TYPE}
Allow an empty USE_LANGUAGES setting to not break in show-shlib-type. If we're not compiling anything, we're likely not installing any shared libraries.
2004-02-06 04:10:29 +01:00
. endif # USE_LANGUAGES
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
.endif
- Move all CHECK_FILES_SKIP definitions to a central place in bsd.pkg.mk. - Move the previous block of code down in the file so that all used variables are defined (specially PKG_SYSCONFDIR). Fixes a problem noticed by kristerw@'s bulk build in comms/minicom. - Ignore diff's return code, which aborts make in NetBSD 1.6.2. Also noticed by kristerw@'s bulk build. - Use full paths to do the checks, instead of relative to ${PREFIX}. Less ambiguity. Matches should be turned into regular expressions that anchor to a whole line (tried that, but found some problems). - Turn CHECK_FILES to NO by default. As said in the previous point, there are still some problems that have to be fixed and minor improvements to be done. And I have no time to fix this ATM. Yes, this definitely needs more testing. I'm sorry for all the noise. (But hey! you should set and try this feature locally! ;-)
2005-01-15 11:50:17 +01:00
# CHECK_FILES_SKIP is a list of file names that will be skipped when
# analyzing file lists in the check-files target. This is useful to
# avoid getting errors triggered by changes in directories not really
# handled by pkgsrc.
#
# We have to do this here, i.e., quite late in bsd.pkg.mk parsing, so
# that the variables used in them are defined. Otherwise we may get
# problems with the 'for' expressions.
.if defined(INFO_DIR)
CHECK_FILES_SKIP+= ${PREFIX}/${INFO_DIR}/dir
.endif
CHECK_FILES_SKIP+= ${PREFIX}/emul/linux/proc
.for e c in ${CONF_FILES} ${SUPPORT_FILES}
CHECK_FILES_SKIP+= ${c}/.pkgsrc
.endfor
.for e c o g m in ${CONF_FILES_PERMS} ${SUPPORT_FILES_PERMS}
CHECK_FILES_SKIP+= ${c}/.pkgsrc
.endfor
.for d in ${MAKE_DIRS} ${OWN_DIRS}
CHECK_FILES_SKIP+= ${d}
.endfor
.for d o g m in ${MAKE_DIRS_PERMS} ${OWN_DIRS_PERMS}
CHECK_FILES_SKIP+= ${d}
.endfor
Make CHECK_FILES_SKIP_CMD into a single grep, using a ${var:@foo@bar@} loop.
2005-01-24 19:44:38 +01:00
CHECK_FILES_SKIP_CMD= | ${GREP} -v ${CHECK_FILES_SKIP:@f@-e ${f:Q}@}
- Move all CHECK_FILES_SKIP definitions to a central place in bsd.pkg.mk. - Move the previous block of code down in the file so that all used variables are defined (specially PKG_SYSCONFDIR). Fixes a problem noticed by kristerw@'s bulk build in comms/minicom. - Ignore diff's return code, which aborts make in NetBSD 1.6.2. Also noticed by kristerw@'s bulk build. - Use full paths to do the checks, instead of relative to ${PREFIX}. Less ambiguity. Matches should be turned into regular expressions that anchor to a whole line (tried that, but found some problems). - Turn CHECK_FILES to NO by default. As said in the previous point, there are still some problems that have to be fixed and minor improvements to be done. And I have no time to fix this ATM. Yes, this definitely needs more testing. I'm sorry for all the noise. (But hey! you should set and try this feature locally! ;-)
2005-01-15 11:50:17 +01:00
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
# Check if the generated PLIST matches the list of really installed files.
#
.PHONY: check-files
check-files:
.if !defined(NO_PKG_REGISTER)
${_PKG_SILENT}${_PKG_DEBUG} \
errors=0; \
diff -u ${WRKDIR}/.prefix.pre ${WRKDIR}/.prefix.post \
- Move all CHECK_FILES_SKIP definitions to a central place in bsd.pkg.mk. - Move the previous block of code down in the file so that all used variables are defined (specially PKG_SYSCONFDIR). Fixes a problem noticed by kristerw@'s bulk build in comms/minicom. - Ignore diff's return code, which aborts make in NetBSD 1.6.2. Also noticed by kristerw@'s bulk build. - Use full paths to do the checks, instead of relative to ${PREFIX}. Less ambiguity. Matches should be turned into regular expressions that anchor to a whole line (tried that, but found some problems). - Turn CHECK_FILES to NO by default. As said in the previous point, there are still some problems that have to be fixed and minor improvements to be done. And I have no time to fix this ATM. Yes, this definitely needs more testing. I'm sorry for all the noise. (But hey! you should set and try this feature locally! ;-)
2005-01-15 11:50:17 +01:00
>${WRKDIR}/.files.diff || ${TRUE}; \
${GREP} '^+/' ${WRKDIR}/.files.diff | ${SED} "s|^+||" | ${SORT} \
>${WRKDIR}/.files.added; \
${GREP} '^-/' ${WRKDIR}/.files.diff | ${SED} "s|^-||" | ${SORT} \
>${WRKDIR}/.files.deleted; \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
if ${TEST} `${WC} -l ${WRKDIR}/.files.deleted | \
${AWK} '{ print $$1; }'` -gt 0; \
then \
echo "*** The following files have been deleted from" \
"${PREFIX}!"; \
${SED} "s|^| |" <${WRKDIR}/.files.deleted; \
errors=1; \
fi; \
- Move all CHECK_FILES_SKIP definitions to a central place in bsd.pkg.mk. - Move the previous block of code down in the file so that all used variables are defined (specially PKG_SYSCONFDIR). Fixes a problem noticed by kristerw@'s bulk build in comms/minicom. - Ignore diff's return code, which aborts make in NetBSD 1.6.2. Also noticed by kristerw@'s bulk build. - Use full paths to do the checks, instead of relative to ${PREFIX}. Less ambiguity. Matches should be turned into regular expressions that anchor to a whole line (tried that, but found some problems). - Turn CHECK_FILES to NO by default. As said in the previous point, there are still some problems that have to be fixed and minor improvements to be done. And I have no time to fix this ATM. Yes, this definitely needs more testing. I'm sorry for all the noise. (But hey! you should set and try this feature locally! ;-)
2005-01-15 11:50:17 +01:00
${GREP} '^[A-Za-z]' ${PLIST} | ${SED} "s|^|${PREFIX}/|" | \
${SORT} >${WRKDIR}/.files.expected; \
Add the check-files target: when enabled, this verifies that installed packages do not install more files than expected (nor delete existing files), aside other sanity checks in PKG_SYSCONFDIR and VARBASE. This behavior is only enabled if PKG_DEVELOPER is set and CHECK_FILES is YES (the default). Should let us catch problems in other systems, as some packages install different files depending on the OS they are being built (which is different to see). Furthermore, since the sanity checks done in PKG_SYSCONFDIR and VARBASE are quite agressive, only enable them when CHECK_FILES_STRICT is YES (defaults to NO). Developers should enable this feature to detect errors, but this can't be a default yet. Otherwise, lots of packages could be marked broken in bulk builds (they really are, according to "cleanliness" rules, but most of them are non-trivial to fix).
2005-01-14 15:36:31 +01:00
if ! ${CMP} -s ${WRKDIR}/.files.expected ${WRKDIR}/.files.added; then \
echo "*** The PLIST does not match installed files!"; \
echo " The following files were not expected in ${PREFIX}:";\
diff -u ${WRKDIR}/.files.expected ${WRKDIR}/.files.added | \
${GREP} '^+[^+]' | ${SED} "s|^+| |"; \
errors=1; \
fi; \
if [ ${CHECK_FILES_STRICT} = "YES" ] && \
! ${CMP} -s ${WRKDIR}/.sysconfdir.pre ${WRKDIR}/.sysconfdir.post; \
then \
echo "*** The package has modified ${PKG_SYSCONFDIR}" \
"contents directly!"; \
echo " The offending files/directories are:"; \
diff -u ${WRKDIR}/.sysconfdir.pre ${WRKDIR}/.sysconfdir.post | \
${GREP} '^+[^+]' | ${SED} "s|^+| |"; \
errors=1; \
fi; \
if [ ${CHECK_FILES_STRICT} = "YES" ] && \
! ${CMP} -s ${WRKDIR}/.varbase.pre ${WRKDIR}/.varbase.post; then \
echo "*** The package has modified ${VARBASE} contents" \
"directly!"; \
echo " The offending files/directories are:"; \
diff -u ${WRKDIR}/.varbase.pre ${WRKDIR}/.varbase.post |\
${GREP} '^+[^+]' | ${SED} "s|^+| |"; \
errors=1; \
fi; \
${RM} -f ${WRKDIR}/.files.added ${WRKDIR}/.files.deleted \
${WRKDIR}/.files.diff ${WRKDIR}/.files.expected; \
${TEST} $$errors -eq 0
.endif
whitespace fix -- does nobody use emacs here?
2005-04-25 18:47:18 +02:00
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
.PHONY: acquire-extract-lock acquire-patch-lock acquire-tools-lock
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.PHONY: acquire-wrapper-lock acquire-configure-lock acquire-build-lock
Finish off the locking work, by adding locks around "make install" and "make package", from a nudge by Greg Oster. No objections, but lots of mail, received from pkgsrc developers, to committing this during the stability freeze, but I'd really like this to go in pkgsrc-2004Q3.
2004-09-11 09:26:03 +02:00
.PHONY: acquire-install-lock acquire-package-lock
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
acquire-extract-lock:
${_ACQUIRE_LOCK}
acquire-patch-lock:
${_ACQUIRE_LOCK}
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
acquire-tools-lock:
${_ACQUIRE_LOCK}
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
acquire-wrapper-lock:
Merge pkgsrc/mk from the buildlink2 branch back into the main trunk.
2002-08-22 10:21:26 +02:00
${_ACQUIRE_LOCK}
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
acquire-configure-lock:
${_ACQUIRE_LOCK}
acquire-build-lock:
${_ACQUIRE_LOCK}
Finish off the locking work, by adding locks around "make install" and "make package", from a nudge by Greg Oster. No objections, but lots of mail, received from pkgsrc developers, to committing this during the stability freeze, but I'd really like this to go in pkgsrc-2004Q3.
2004-09-11 09:26:03 +02:00
acquire-install-lock:
${_ACQUIRE_LOCK}
acquire-package-lock:
${_ACQUIRE_LOCK}
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
.PHONY: release-extract-lock release-patch-lock release-tools-lock
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.PHONY: release-wrapper-lock release-configure-lock release-build-lock
Finish off the locking work, by adding locks around "make install" and "make package", from a nudge by Greg Oster. No objections, but lots of mail, received from pkgsrc developers, to committing this during the stability freeze, but I'd really like this to go in pkgsrc-2004Q3.
2004-09-11 09:26:03 +02:00
.PHONY: release-install-lock release-package-lock
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
release-extract-lock:
${_RELEASE_LOCK}
release-patch-lock:
${_RELEASE_LOCK}
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
release-tools-lock:
${_RELEASE_LOCK}
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
release-wrapper-lock:
Merge pkgsrc/mk from the buildlink2 branch back into the main trunk.
2002-08-22 10:21:26 +02:00
${_RELEASE_LOCK}
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
release-configure-lock:
${_RELEASE_LOCK}
release-build-lock:
${_RELEASE_LOCK}
Finish off the locking work, by adding locks around "make install" and "make package", from a nudge by Greg Oster. No objections, but lots of mail, received from pkgsrc developers, to committing this during the stability freeze, but I'd really like this to go in pkgsrc-2004Q3.
2004-09-11 09:26:03 +02:00
release-install-lock:
${_RELEASE_LOCK}
release-package-lock:
${_RELEASE_LOCK}
Work out the type of shared libraries once, at "make install" time, and do this dynamically. This means that we no longer have to have a separate pkgsrc copy of bsd.own.mk, which was a maintenance nightmare, and usually out of date, and quite costly too (with a calculation of whether NetBSD-current was being run every time the file was read). Also, on a.out, add ldconfig lines to the generated PLIST only if they're not already there, rather than if they were previously there, and ignore the return value from ldconfig, which cannot be trusted. The ldconfig calls have largely been made obsolete by the use of -rpath, but keep it for just now for backwards compatibility.
1999-04-13 16:18:11 +02:00
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
################################################################
# Skeleton targets start here
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
#
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# You shouldn't have to change these. Either add the pre-* or
# post-* targets/scripts or redefine the do-* targets. These
# targets don't do anything other than checking for cookies and
# call the necessary targets/scripts.
################################################################
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: fetch
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(fetch)
fetch:
"fetch" is also a distinct build phase.
2004-01-27 03:36:59 +01:00
@cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} real-fetch PKG_PHASE=fetch
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: extract
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(extract)
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
extract: checksum ${WRKDIR} acquire-extract-lock ${EXTRACT_COOKIE} release-extract-lock
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: patch
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(patch)
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
patch: extract acquire-patch-lock ${PATCH_COOKIE} release-patch-lock
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
.PHONY: tools
.if !target(tools)
tools: patch acquire-tools-lock ${TOOLS_COOKIE} release-tools-lock
.endif
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.PHONY: wrapper
.if !target(wrapper)
wrapper: tools acquire-wrapper-lock ${WRAPPER_COOKIE} release-wrapper-lock
Merge pkgsrc/mk from the buildlink2 branch back into the main trunk.
2002-08-22 10:21:26 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: configure
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(configure)
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
configure: wrapper acquire-configure-lock ${CONFIGURE_COOKIE} release-configure-lock
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: build
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(build)
Add coarse-grained locking to pkgsrc builds, by means of a beefed up version of the bare-bones code in PR 7590, from David Maxwell. The definition governing the type of locking used is PKGSRC_LOCKTYPE, which can take any of the values "none", "sleep", and "once". The default is "none". If "sleep" locking is used, and process A is building a package, when process B attempts to build the same package, process B will sleep for PKGSRC_SLEEPSECS seconds, and attempt to grab the lock again. Coarse-grained locking uses the OBJHOSTNAME definition to ensure that the PID space is regular for shlock(1) to do its work. The pkgsrc/pkgtools/shlock package has been provided for environments where shlock is not standard.
2002-01-19 22:29:41 +01:00
build: configure acquire-build-lock ${BUILD_COOKIE} release-build-lock
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: test
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
.if !target(test)
test: build ${TEST_COOKIE}
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: install
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(install)
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
install: ${_PKGSRC_BUILD_TARGETS} acquire-install-lock ${INSTALL_COOKIE} release-install-lock
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: package
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(package)
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
package: install acquire-package-lock ${PACKAGE_COOKIE} release-package-lock
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
.if !target(replace)
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
replace: ${_PKGSRC_BUILD_TARGETS} real-replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: undo-replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
.if !target(undo-replace)
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
undo-replace: real-undo-replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${EXTRACT_COOKIE}:
Wrap EXTRACT_COOKIE like all others that could have an interactive step and check for it if doing a batch build
2003-05-11 09:06:23 +02:00
.if ${INTERACTIVE_STAGE:Mextract} == "extract" && defined(BATCH)
@${ECHO} "*** The extract stage of this package requires user interaction"
@${ECHO} "*** Please extract manually with \"cd ${PKGDIR} && ${MAKE} extract\""
@${TOUCH} ${INTERACTIVE_COOKIE}
@${FALSE}
.else
Support DEPENDS_TARGET="install clean"; from Allan Barrett in PR pkg/25059.
2004-04-05 10:06:07 +02:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} real-extract DEPENDS_TARGET=${DEPENDS_TARGET:Q} PKG_PHASE=extract
Wrap EXTRACT_COOKIE like all others that could have an interactive step and check for it if doing a batch build
2003-05-11 09:06:23 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${PATCH_COOKIE}:
Introduce concept of the "phase" that we're in as we progress through fetching, extracting, configuring, building, etc. of a package. We can check what phase we're in by examining the value of ${PKG_PHASE} and comparing against PHASES_AFTER_<phase>, which list phases that are "greater than or equal to" <phase>. One useful example of how to use PKG_PHASE is: .if !empty(PHASES_AFTER_EXTRACT:${PKG_PHASE}) # # Some variable settings or targets here that rely on dependencies to # already be installed, or ${WRKDIR} to be created, etc., as these are # things that should have happened by the time "make extract" is # completed. # .endif
2004-01-21 19:13:27 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} real-patch PKG_PHASE=patch
Wrap EXTRACT_COOKIE like all others that could have an interactive step and check for it if doing a batch build
2003-05-11 09:06:23 +02:00
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
${TOOLS_COOKIE}:
Introduce concept of the "phase" that we're in as we progress through fetching, extracting, configuring, building, etc. of a package. We can check what phase we're in by examining the value of ${PKG_PHASE} and comparing against PHASES_AFTER_<phase>, which list phases that are "greater than or equal to" <phase>. One useful example of how to use PKG_PHASE is: .if !empty(PHASES_AFTER_EXTRACT:${PKG_PHASE}) # # Some variable settings or targets here that rely on dependencies to # already be installed, or ${WRKDIR} to be created, etc., as these are # things that should have happened by the time "make extract" is # completed. # .endif
2004-01-21 19:13:27 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} real-tools PKG_PHASE=tools
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
${WRAPPER_COOKIE}:
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${SETENV} ${BUILD_ENV} ${MAKE} ${MAKEFLAGS} real-wrapper PKG_PHASE=wrapper
Wrap EXTRACT_COOKIE like all others that could have an interactive step and check for it if doing a batch build
2003-05-11 09:06:23 +02:00
Allow the creation of error handlers for shell commands in make targets. PKG_ERROR_CLASSES is a list of error handlers, PKG_ERROR_HANDLER.<class>, to create. The default error handler will rethrow the exit code after emitting an error message that may be specified per-class with PKG_ERROR_MSG.<class>. Create error messages for the configure and build phases that inform the user of places to go looking for why the build didn't complete.
2005-01-10 19:34:48 +01:00
PKG_ERROR_CLASSES+= configure
PKG_ERROR_MSG.configure= \
"" \
"There was an error during the \`\`configure'' phase." \
"Please investigate the following for more information:"
.if defined(GNU_CONFIGURE)
PKG_ERROR_MSG.configure+= \
" * config.log" \
" * ${WRKLOG}" \
""
.else
PKG_ERROR_MSG.configure+= \
" * log of the build" \
" * ${WRKLOG}" \
""
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${CONFIGURE_COOKIE}:
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
.if ${INTERACTIVE_STAGE:Mconfigure} == "configure" && defined(BATCH)
@${ECHO} "*** The configuration stage of this package requires user interaction"
For interactive configure, say "Please configure manually...".
2002-07-03 08:58:31 +02:00
@${ECHO} "*** Please configure manually with \"cd ${PKGDIR} && ${MAKE} configure\""
Make bulk builds aware of packages with INTERACTIVE_STAGE and mark them as "not packaged" instead of "broken".
2002-07-21 13:53:18 +02:00
@${TOUCH} ${INTERACTIVE_COOKIE}
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
@${FALSE}
.else
Allow the creation of error handlers for shell commands in make targets. PKG_ERROR_CLASSES is a list of error handlers, PKG_ERROR_HANDLER.<class>, to create. The default error handler will rethrow the exit code after emitting an error message that may be specified per-class with PKG_ERROR_MSG.<class>. Create error messages for the configure and build phases that inform the user of places to go looking for why the build didn't complete.
2005-01-10 19:34:48 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${SETENV} ${BUILD_ENV} ${MAKE} ${MAKEFLAGS} real-configure PKG_PHASE=configure || ${PKG_ERROR_HANDLER.configure}
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
.endif
Allow the creation of error handlers for shell commands in make targets. PKG_ERROR_CLASSES is a list of error handlers, PKG_ERROR_HANDLER.<class>, to create. The default error handler will rethrow the exit code after emitting an error message that may be specified per-class with PKG_ERROR_MSG.<class>. Create error messages for the configure and build phases that inform the user of places to go looking for why the build didn't complete.
2005-01-10 19:34:48 +01:00
PKG_ERROR_CLASSES+= build
PKG_ERROR_MSG.build= \
"" \
"There was an error during the \`\`build'' phase." \
"Please investigate the following for more information:" \
" * log of the build" \
" * ${WRKLOG}" \
""
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${BUILD_COOKIE}:
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
.if ${INTERACTIVE_STAGE:Mbuild} == "build" && defined(BATCH)
@${ECHO} "*** The build stage of this package requires user interaction"
@${ECHO} "*** Please build manually with \"cd ${PKGDIR} && ${MAKE} build\""
Make bulk builds aware of packages with INTERACTIVE_STAGE and mark them as "not packaged" instead of "broken".
2002-07-21 13:53:18 +02:00
@${TOUCH} ${INTERACTIVE_COOKIE}
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
@${FALSE}
.else
Allow the creation of error handlers for shell commands in make targets. PKG_ERROR_CLASSES is a list of error handlers, PKG_ERROR_HANDLER.<class>, to create. The default error handler will rethrow the exit code after emitting an error message that may be specified per-class with PKG_ERROR_MSG.<class>. Create error messages for the configure and build phases that inform the user of places to go looking for why the build didn't complete.
2005-01-10 19:34:48 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${SETENV} ${BUILD_ENV} ${MAKE} ${MAKEFLAGS} real-build PKG_PHASE=build || ${PKG_ERROR_HANDLER.build}
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
.endif
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
${TEST_COOKIE}:
Rename PATH_ENV to BUILD_ENV, which better expresses what it's for. BUILD_ENV is used to set the environment for the sub-make processes that are invoked for each phase after and including buildlink.
2004-02-21 13:29:23 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${SETENV} ${BUILD_ENV} ${MAKE} ${MAKEFLAGS} real-test PKG_PHASE=test
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${INSTALL_COOKIE}:
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
.if ${INTERACTIVE_STAGE:Minstall} == "install" && defined(BATCH)
@${ECHO} "*** The installation stage of this package requires user interaction"
@${ECHO} "*** Please install manually with \"cd ${PKGDIR} && ${MAKE} install\""
Make bulk builds aware of packages with INTERACTIVE_STAGE and mark them as "not packaged" instead of "broken".
2002-07-21 13:53:18 +02:00
@${TOUCH} ${INTERACTIVE_COOKIE}
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
@${FALSE}
.else
Rename PATH_ENV to BUILD_ENV, which better expresses what it's for. BUILD_ENV is used to set the environment for the sub-make processes that are invoked for each phase after and including buildlink.
2004-02-21 13:29:23 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${SETENV} ${BUILD_ENV} ${MAKE} ${MAKEFLAGS} real-install PKG_PHASE=install
Deprecate IS_INTERACTIVE, and introduce a finer-grained INTERACTIVE_STAGE definition. INTERACTIVE_STAGE can take any of the values: fetch, configure, build and install Multiple values are allowed: e.g. INTERACTIVE_STAGE= configure install Explain INTERACTIVE_STAGE and its use in documentation. Patches provided by Chris Pinnock (cjep@netbsd.org).
2002-07-02 13:26:05 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${PACKAGE_COOKIE}:
Rename PATH_ENV to BUILD_ENV, which better expresses what it's for. BUILD_ENV is used to set the environment for the sub-make processes that are invoked for each phase after and including buildlink.
2004-02-21 13:29:23 +01:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${SETENV} ${BUILD_ENV} ${MAKE} ${MAKEFLAGS} real-package PKG_PHASE=package
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.PHONY: extract-message patch-message tools-message wrapper-message
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
.PHONY: configure-message build-message test-message
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
extract-message:
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> Extracting for ${PKGNAME}"
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
patch-message:
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> Patching for ${PKGNAME}"
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
tools-message:
@${ECHO_MSG} "${_PKGSRC_IN}> Overriding tools for ${PKGNAME}"
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
wrapper-message:
@${ECHO_MSG} "${_PKGSRC_IN}> Creating toolchain wrappers for ${PKGNAME}"
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
configure-message:
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> Configuring for ${PKGNAME}"
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
build-message:
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> Building for ${PKGNAME}"
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
test-message:
@${ECHO_MSG} "${_PKGSRC_IN}> Testing for ${PKGNAME}"
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.PHONY: extract-cookie patch-cookie tools-cookie wrapper-cookie
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
.PHONY: configure-cookie build-cookie test-cookie
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
extract-cookie:
${_PKG_SILENT}${_PKG_DEBUG}${ECHO} ${PKGNAME} >> ${EXTRACT_COOKIE}
patch-cookie:
${_PKG_SILENT}${_PKG_DEBUG} ${TOUCH} ${TOUCH_FLAGS} ${PATCH_COOKIE}
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
tools-cookie:
${_PKG_SILENT}${_PKG_DEBUG} ${TOUCH} ${TOUCH_FLAGS} ${TOOLS_COOKIE}
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
wrapper-cookie:
${_PKG_SILENT}${_PKG_DEBUG} ${TOUCH} ${TOUCH_FLAGS} ${WRAPPER_COOKIE}
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
configure-cookie:
${_PKG_SILENT}${_PKG_DEBUG} ${TOUCH} ${TOUCH_FLAGS} ${CONFIGURE_COOKIE}
build-cookie:
${_PKG_SILENT}${_PKG_DEBUG} ${TOUCH} ${TOUCH_FLAGS} ${BUILD_COOKIE}
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
test-cookie:
${_PKG_SILENT}${_PKG_DEBUG} ${TOUCH} ${TOUCH_FLAGS} ${TEST_COOKIE}
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
.ORDER: pre-fetch do-fetch post-fetch
.ORDER: extract-message install-depends pre-extract do-extract post-extract extract-cookie
.ORDER: patch-message pre-patch do-patch post-patch patch-cookie
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
.ORDER: tools-message pre-tools do-tools post-tools tools-cookie
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.ORDER: wrapper-message pre-wrapper do-wrapper post-wrapper wrapper-cookie
Fix for _CONFIGURE_PREREQ targets not being run when a package defines a do-configure target. To do this I've introduced a new target, pre-configure-overrides, that depends on ${_CONFIGURE_PREREQ}. It's run after pre-configure but before do-configure. This fixes pkg/22077, and allows benchmarks/benchfft to build correctly on a -current cats box. Reviewed by wiz@
2003-07-08 11:20:30 +02:00
.ORDER: configure-message pre-configure pre-configure-override do-configure post-configure configure-cookie
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
.ORDER: build-message pre-build do-build post-build build-cookie
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
.ORDER: test-message pre-test do-test post-test test-cookie
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
# Please note that the order of the following targets is important, and
# should not be modified (.ORDER is not recognised by make(1) in a serial
# make i.e. without -j n)
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.PHONY: real-fetch real-extract real-patch
.PHONY: real-tools real-wrapper
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
.PHONY: real-configure real-build real-test real-install real-package
.PHONY: real-replace real-undo-replace
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
real-fetch: pre-fetch do-fetch post-fetch
real-extract: extract-message install-depends pre-extract do-extract post-extract extract-cookie
real-patch: patch-message pre-patch do-patch post-patch patch-cookie
Move the hide-tools portion of bsd.buildlink2.mk into tools.mk, and put the tools in ${WRKDIR}/.tools/bin instead of in ${BUILDLINK_DIR}, and remove the need for buildlink2 to use USE_GNU_TOOLS. In the modified USE_GNU_TOOLS implementation, the following implementation- specific variables have the following meanings: _TOOLS_REPLACE.<tool> means that we want ${AWK}, ${SED}, etc. symlinked into ${TOOLS_DIR} as awk, sed, etc. _TOOLS_NEED_GNU.<tool> means that we want to use the pkgsrc version of <tool> symlinked info ${TOOLS_DIR} as awk, sed, etc. If this is "YES", then it always trumps _TOOLS_REPLACE.<tool>. And we want nothing to happen if we're building the pkgsrc GNU tool itself. The modified USE_GNU_TOOLS implementation should also hopefully fix the circular dependency problem. Create a new target "tools" that is run after "patch" and before "buildlink" that populates the ${TOOLS_DIR} directory. This ensures that it's always run at the right time, instead of relying on pre-buildlink or pre-configure, which may be cancelled by NO_BUILDLINK or NO_CONFIGURE. XXX There is some possible fallout in texinfo.mk with the MAKE_ENV and XXX CONFIGURE_ENV settings for INSTALL_INFO and MAKEINFO. It looks XXX like the MAKE_ENV and CONFIGURE_ENV settings should move from XXX tools.mk into texinfo.mk, and they no longer need to be XXX conditional on USE_BUILDLINK2. I'll leave it to the texinfo.mk XXX dude (Hi, Stoned!).
2003-08-16 10:50:17 +02:00
real-tools: tools-message pre-tools do-tools post-tools tools-cookie
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
real-wrapper: wrapper-message pre-wrapper do-wrapper post-wrapper wrapper-cookie
Fix for _CONFIGURE_PREREQ targets not being run when a package defines a do-configure target. To do this I've introduced a new target, pre-configure-overrides, that depends on ${_CONFIGURE_PREREQ}. It's run after pre-configure but before do-configure. This fixes pkg/22077, and allows benchmarks/benchfft to build correctly on a -current cats box. Reviewed by wiz@
2003-07-08 11:20:30 +02:00
real-configure: configure-message pre-configure pre-configure-override do-configure post-configure configure-cookie
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
real-build: build-message pre-build do-build post-build build-cookie
Add a test target to bsd.pkg.mk: It runs pre-test, do-test, and post-test. The default pre-test and post-test targets are empty; the default do-test target runs TEST_TARGET (in each TEST_DIRS (defaults to BUILD_DIRS)) if it is set and does nothing otherwise. Thus the test target defaults to doing nothing. The usual way to use this is to set TEST_TARGET to the target to run (e.g. TEST_TARGET=check for GNU packages). If PKGSRC_RUN_TEST is set to yes (in mk.conf, defaults to no), the test target is run before install. (approved by agc)
2002-11-07 05:22:29 +01:00
real-test: test-message pre-test do-test post-test test-cookie
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
real-install: do-su-install
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
real-package: do-su-package
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
real-replace: do-su-replace
real-undo-replace: do-su-undo-replace
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
_SU_TARGET= \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
if [ `${ID} -u` = `${ID} -u ${ROOT_USER}` ]; then \
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
${MAKE} ${MAKEFLAGS} $$realtarget; \
Make this work on Solaris again. Minor formatting changes around the pkg-su-install target. Pass on any PKG_DEBUG_LEVEL as well as FORCE_PKG_REGISTER definitions.
1999-04-06 16:11:10 +02:00
elif [ "X${BATCH}" != X"" ]; then \
`mtree' -> `$$action' in _SU_TARGET
2002-04-29 11:12:20 +02:00
${ECHO_MSG} "Warning: Batch mode, not superuser, can't run $$action for ${PKGNAME}."; \
ROOT_USER is parameterized, so use it in human-readable printed messages as well.
2004-04-26 05:07:46 +02:00
${ECHO_MSG} "Become ${ROOT_USER} and try again to ensure correct permissions."; \
+ Factor out the real-install target from the rest of the .USE ones. This makes it possible to do the next step. There were so many .if make(install) checks that this arguably should be done anyway. + In interactive installs, if the effective uid is not zero when "make install" is attempted, then ${SU_CMD} is invoked (defaults to "su - root -c", but sudo or priv could be used). This allows running as non-root for much more time, which is a significant benefit. This also allows installation of pre-requisite (DEPENDS) packages while building a package as non-root user. + A hook has been added to run ${PRE_ROOT_CMD} before becoming root, so that root can be given write access to ${WRKDIR} directories, for example, on -noroot mounted NFS filesystems.
1999-03-31 11:04:18 +02:00
else \
Make this work on Solaris again. Minor formatting changes around the pkg-su-install target. Pass on any PKG_DEBUG_LEVEL as well as FORCE_PKG_REGISTER definitions.
1999-04-06 16:11:10 +02:00
args=""; \
if [ "X${FORCE_PKG_REGISTER}" != X"" ]; then \
args="FORCE_PKG_REGISTER=1"; \
fi; \
if [ "X${PKG_DEBUG_LEVEL}" != X"" ]; then \
args="$$args PKG_DEBUG_LEVEL=${PKG_DEBUG_LEVEL}"; \
fi; \
if [ "X${PRE_ROOT_CMD}" != "X${TRUE}" ]; then \
${ECHO} "*** WARNING *** Running: ${PRE_ROOT_CMD}"; \
${PRE_ROOT_CMD}; \
fi; \
ROOT_USER is parameterized, so use it in human-readable printed messages as well.
2004-04-26 05:07:46 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> Becoming ${ROOT_USER}@`${HOSTNAME_CMD}` to $$action ${PKGBASE}."; \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "`${ECHO} ${SU_CMD} | ${AWK} '{ print $$1 }'` ";\
Add a variable SU_CMD_PATH_APPEND containing a colon separated list of directories. When ${SU_CMD} is used, we add this list to the end of the PATH environment variable. This fixes PR#17532 from Hauke Fath.
2002-07-23 21:10:27 +02:00
${SU_CMD} "cd ${.CURDIR}; ${SETENV} PATH=$${PATH}:${SU_CMD_PATH_APPEND} ${MAKE} $$args ${MAKEFLAGS} $$realtarget $$realflags"; \
+ Factor out the real-install target from the rest of the .USE ones. This makes it possible to do the next step. There were so many .if make(install) checks that this arguably should be done anyway. + In interactive installs, if the effective uid is not zero when "make install" is attempted, then ${SU_CMD} is invoked (defaults to "su - root -c", but sudo or priv could be used). This allows running as non-root for much more time, which is a significant benefit. This also allows installation of pre-requisite (DEPENDS) packages while building a package as non-root user. + A hook has been added to run ${PRE_ROOT_CMD} before becoming root, so that root can be given write access to ${WRKDIR} directories, for example, on -noroot mounted NFS filesystems.
1999-03-31 11:04:18 +02:00
fi
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-su-install
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
do-su-install:
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
extractname=`${CAT} ${EXTRACT_COOKIE}`; \
case "$$extractname" in \
"") ${ECHO_MSG} "*** Warning: ${WRKDIR} may contain an older version of ${PKGBASE}" ;; \
"${PKGNAME}") ;; \
In the do-su-install target, when the ${WRKDIR} contains an outdated version of the package, and an attempt is made to install the package, first "make clean" and then "make build" to pick up the new version of the package before continuing with "make install", rather than printing an error message and dying.
2002-04-18 22:54:51 +02:00
*) ${ECHO_MSG} "*** Warning: Package version $$extractname in ${WRKDIR}"; \
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
${ECHO_MSG} "*** Current version ${PKGNAME} in pkgsrc directory"; \
In the do-su-install target, when the ${WRKDIR} contains an outdated version of the package, and an attempt is made to install the package, first "make clean" and then "make build" to pick up the new version of the package before continuing with "make install", rather than printing an error message and dying.
2002-04-18 22:54:51 +02:00
${ECHO_MSG} "*** Cleaning and rebuilding the newer version of the package..."; \
${MAKE} clean && ${MAKE} build ;; \
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
esac
@${ECHO_MSG} "${_PKGSRC_IN}> Installing for ${PKGNAME}"
${_PKG_SILENT}${_PKG_DEBUG} \
realtarget="real-su-install"; \
Allow customizing the message for _SU_TARGET by substituting for "$action", where action can be "install", "package", "create user for", etc., instead of blindly always saying "install". Define "action" before calling ${_SU_TARGET} whereever it is used.
2001-10-01 23:15:11 +02:00
action="install"; \
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
${_SU_TARGET}
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-su-package
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
do-su-package:
@${ECHO_MSG} "${_PKGSRC_IN}> Packaging ${PKGNAME}"
${_PKG_SILENT}${_PKG_DEBUG} \
realtarget="real-su-package"; \
Allow customizing the message for _SU_TARGET by substituting for "$action", where action can be "install", "package", "create user for", etc., instead of blindly always saying "install". Define "action" before calling ${_SU_TARGET} whereever it is used.
2001-10-01 23:15:11 +02:00
action="package"; \
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
${_SU_TARGET}
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-su-replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
do-su-replace:
@${ECHO_MSG} "${_PKGSRC_IN}> Replacing ${PKGNAME}"
${_PKG_SILENT}${_PKG_DEBUG} \
realtarget="real-su-replace"; \
action="replace"; \
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
${_SU_TARGET}
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: do-su-undo-replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
do-su-undo-replace:
@${ECHO_MSG} "${_PKGSRC_IN}> Undoing Replacement of ${PKGNAME}"
${_PKG_SILENT}${_PKG_DEBUG} \
realtarget="real-su-undo-replace"; \
action="undo-replace"; \
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
${_SU_TARGET}
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
Following on from the removal of the automatic "is there a script to run" logic, more speedups for building packages. Get rid of a .USE macro, and do not spawn sub-makes for the pre-, do- and post-target stages, replacing them with standard make(1) targets. Timing information as follows (multiple runs performed, best results taken): 800 MHz Celeron, 128 MB, local pkgsrc, local obj scripts/, pre,do,post-*: 0.731u 0.261s 0:02.04 48.5% 0+0k 29+168io 9pf+0w no scripts/, pre,do,post-*: 0.678u 0.242s 0:01.30 70.0% 0+0k 0+169io 0pf+0w no scripts/, no pre,do,post-*: 0.267u 0.089s 0:00.90 37.7% 0+0k 0+155io 0pf+0w 40 MHz Sparc, 36 MB, nfs pkgsrc, local obj scripts/, pre,do,post-*: 22.590u 6.839s 0:33.31 88.3% 0+0k 121+254io 0pf+0w no scripts/, pre,do,post-*: 22.481u 6.442s 0:33.30 86.8% 0+0k 120+251io 0pf+0w no scripts/, no pre,do,post-*: 8.534u 4.189s 0:16.48 77.1% 0+0k 105+242io 0pf+0w
2001-11-08 10:29:56 +01:00
# Empty pre-* and post-* targets
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Initial commit of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3 into mk/wrapper. The buildlink3 code is modified to simply hook its transformations into the wrapper script framework. The wrapper script framework has some new features: * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Much improved debugging output. It's possible to output the wrapper work log in-line with normal output by setting WRAPPER_LOG to "stderr". Important differences in behaviour from the old buildlink3 code include: * Only move the -l options to the end of the command line, leaving the -L options in-place. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la The old --fix-la syntax no longer works. 20040818 ======== * Initial release of a new wrapper script framework that encapsulates the non-buildlink-related code and moves it out of mk/buildlink3. These features include: * making MIPSpro accept GCC options * making MIPSpro "ucode" accept GCC options * making SunPro accept GCC options * making "ld" accept -Wl,* options and silently removing the "-Wl," * (NEW) making UnixWare GCC accept -rpath options and silently converting them into an appropriate LD_RUN_PATH One major benefit of this is that the buildlink3 code is now much tighter and easier to understand since it concerns itself solely with buildlink-related details. I haven't yet optimized the wrapper cache, so the new wrapper scripts may take slightly longer to execute than the old buildlink3 wrapper scripts, but I'll be improving this over time. 20040821 ======== * Move the inclusion of $cmd_sink outside of the main loop in wrapper.sh so that the $cmd_sink script can be used to globally scan and process the arguments. Move the LD_RUN_PATH code to a cmd-sink-unixware-gcc script. Garbage-collect the now unused export_vars-related code. * Add cmd-sink-aix-xlc for AIX xlc that munges -Wl,-R* into an appropriate -blibpath option. * Add cmd-sink-interix-gcc and cmd-sink-interix-ld that errors out when it sees -fpic/-fPIC and -shared/-Bshareable, respectively (requested by <tv>). * Move the code that converts full paths to shared libraries into the "-Ldir -llib" equivalents from the buildlink3 code into wrapper/logic. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Move the code that checks for absolute rpaths from the buildlink3 code into wrapper/arg-source. Remove the same from bsd.buildlink3.mk and gen-transform.sh. * Only move the -l options to the end of the command line, leaving the -L options in-place. * Add more debugging code. 20040824 ======== * Fix quoting problems after arguments are transformed. Remove the hack that was inserted that magically made almost everything work because we do it the right way now. * Move the inclusion of $logic outside of the main loop in wrapper.sh so that the $logic script doesn't have to worry about underflowing the argument buffer. * Encapsulate the loop in wrapper.sh that fills the argument buffer entirely within the arg-source script. * Move from the logic script into the arg-source script the transformations that merge or split arguments. * Fix bug where skipargs was effectively being ignored if it was more than 1. * Handle the whitespace in transformations in the logic script that turn one library option into multiple library options, e.g. "-lreadline" -> "-ledit -ltermcap". * Allow you to specify an environment variable WRAPPER_SKIP_TRANSFORM for whether you wish to skip the transformation step in the logic script. This is intended for testing purposes. * Added check_prog() and init_lib() functions to the shell code library to make it more reusable outside of the wrapper framework. * Allow the msg_log() function to output to "stdout" or "stderr". If you want to have all of the logging appear on the screen, then you can now set WRAPPER_LOG=stderr. * Make some of the script components not overridable on a per-wrapper basis. * Add a gen-transform.sh script that generates transformation sedfiles. The "transform" script is used to transform arguments, while the "untransform" script is used to unwrap files. Move the no-rpath logic from buildlink3/gen-transform.sh into wrapper/gen-transform.sh since it's not buildlink3-specific. * Check for a non-empty blibpath before adding the option in cmd-sink-aix-xlc. * Extend the autodetection of the libtool mode to detect "compile" and "uninstall". * Add a cmd-sink-libtool script that doesn't pass linker options to libtool unless we're in "link" mode. * Set _USE_RPATH to "yes" for UnixWare so that the wrappers will see the rpath options and convert them to a LD_RUN_PATH definition. * Add more debugging code. 20040826 ======== * Rewrite buildlink3/gen-transform.sh to produce more precise sed commands. Drop some unused commands from the mini-language, and add a few more that are more restrictive in their scope. * Fix problem where repeated options weren't properly handled by some of sed commands. It's not enough that they're "global replace", since some patterns match separator characters before and after each option. We must repeat those patterns twice to catch all instances correctly. * Fix problem noted in both PR pkg/24760 and PR pkg/25500, where -L/usr/lib/* was being mangled improperly. * Remove the top-level "buildlink" target; instead, make buildlinking occur as part of the "wrapper" target. * Add more debugging code. 20040828 ======== * Added a head_queue function to shell-lib that returns the head of the named queue without popping it off the front of the queue. * Strip consecutive, repeated library options from the command line when we read it in the logic script. * Be more careful about not underflowing the argument buffer. 20040906 ======== * shell-lib was moved into pkgsrc/mk/scripts; correct references to that file in the wrapper code. * Use opt-sub instead of sub-mangle when protecting -I/usr/include/* and -L/usr/lib/* from buildlink transformations. This avoids adding lines that look like "-I-I..." in the transformation sedfiles. * mangle and sub-mangle are only meant to transform directories in -I, -L, and rpath options, so remove the lines in buildlink3/gen-transform.sh that transformed bare directories. * Fix bug in strip-slashdot where the "." wasn't backquoted and thus matched all characters instead of only the "." character. * Change the libtool wrapper to use a modified buildcmd script that doesn't rearrange any of the arguments. This should fix spurious problems where libtool doesn't understand how to parse the command line when the -l options are moved to the end of the argument list. * Fix bug in the logic script where the $cachearg and $cachedarg weren't being properly set at all times, which caused the cache to contain the wrong transformed argument. 20040907 ======== * Support automatically passing "ABI" flags to the compiler and linker depending on the value of ${ABI}. Currently supports the SunPro compiler with ${ABI} == 64 and the MIPSPro compiler with ${ABI} as any of 32, n32, o32, and 64. * Move back the code that splits absolute paths to shared libraries from arg-source back into logic. This allows us to correctly skip splitting those paths based on the previous option. Also add a sanity check that the library name in the split argument doesn't contain a "/" since shell globs are not as precise as REs. * Don't transform the path given after --dynamic-linker (used by GNU ld for ELF linkage). * Add the ability for the libtool wrapper to be called just to unwrap an existing libtool archive by running: libtool --mode=unwrap -o libfoo.la 20040914 ======== * Add a loop in libtool-fix-la to ensure that all of the options listed in the dependency_libs lines of *.lai files are processed. This fixes a buildlink3 leakage bug. * Merge the gen-transform.sh scripts between buildlink3 and wrapper and place them all in wrapper. This makes sense since the commands simply allow for many types of transformations, which buildlink3 takes advantage of, but there is nothing inherently buildlink-ish about those commands. * Don't directly manipulate SUBST_SED.unwrap. Instead, create the value of SUBST_SED.unwrap by combining several other variables (currently just _UNWRAP_SED) to ensure that the correct ordering is preserved. * Correct some confusing debugging messages.
2004-09-21 17:01:38 +02:00
.for name in fetch extract patch tools wrapper configure build test install-script install package
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if !target(pre-${name})
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
pre-${name}:
@${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if !target(post-${name})
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
post-${name}:
@${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endfor
# Reinstall
#
# Special target to re-run install
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: reinstall
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(reinstall)
Back out the 'make reinstall' change from rev. 1.557: On 'make reinstall', the DEPENDS_TARGET is also set to reinstall, and for pkgs with dependencies this results in all dependencies being removed and reinstalled.
2000-09-01 15:48:51 +02:00
reinstall:
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${RM} -f ${INSTALL_COOKIE} ${PACKAGE_COOKIE} ${PLIST}
Support DEPENDS_TARGET="install clean"; from Allan Barrett in PR pkg/25059.
2004-04-05 10:06:07 +02:00
${_PKG_SILENT}${_PKG_DEBUG}DEPENDS_TARGET=${DEPENDS_TARGET:Q} ${MAKE} ${MAKEFLAGS} install
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Add a deinstall target, so that packages can be removed easily. This implementation was taken from FreeBSD.
1998-02-17 16:07:02 +01:00
# Deinstall
#
# Special target to remove installation
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: deinstall do-su-deinstall
Add a deinstall target, so that packages can be removed easily. This implementation was taken from FreeBSD.
1998-02-17 16:07:02 +01:00
.if !target(deinstall)
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
deinstall: do-su-deinstall
Use "su" automatically for "make deinstall" target, too.
1999-04-03 23:03:50 +02:00
Set UPDATE_RUNNING=YES when deinstalling due to a update or replace target. Use this to trigger appropriate args to pkg_delete so PKG_PRESERVE packages can be upgraded.
2004-11-02 01:03:09 +01:00
.if !target(do-su-deinstall)
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
do-su-deinstall:
use PKGBASE instead of PKGNAME when printing informational messages such as "Deinstalling", as this lead to confusing/misleading output when the (previously) installed pkg has a different version number than the pkgsrc version.
2003-09-28 10:52:13 +02:00
@${ECHO_MSG} "${_PKGSRC_IN}> Deinstalling for ${PKGBASE}"
Use "su" automatically for "make deinstall" target, too.
1999-04-03 23:03:50 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Add a just-in-time su mode for "make package" by factoring out the code whcih does the su(1), and using that. Also make the deinstall target use the same common code.
2001-09-12 12:38:12 +02:00
realtarget="real-su-deinstall"; \
realflags="DEINSTALLDEPENDS=${DEINSTALLDEPENDS}"; \
Allow customizing the message for _SU_TARGET by substituting for "$action", where action can be "install", "package", "create user for", etc., instead of blindly always saying "install". Define "action" before calling ${_SU_TARGET} whereever it is used.
2001-10-01 23:15:11 +02:00
action="deinstall"; \
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
${_SU_TARGET}
Set UPDATE_RUNNING=YES when deinstalling due to a update or replace target. Use this to trigger appropriate args to pkg_delete so PKG_PRESERVE packages can be upgraded.
2004-11-02 01:03:09 +01:00
.endif
Let "make deinstall" remove any version installed, not only the one currently in pkgsrc. Enabled unconditionally by request of several people.
1999-08-21 03:17:59 +02:00
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if (${DEINSTALLDEPENDS} != "NO")
. if (${DEINSTALLDEPENDS} != "ALL")
Add '-r' to root-install-flags if DEINSTALLDEPENDS is set to != "ALL". (This doesn't affect the "make update" target, it just makes "make deinstall DEINSTALLDEPENDS=YES" behave sane).
1999-09-09 00:01:19 +02:00
# used for removing stuff in bulk builds
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
real-su-deinstall-flags+= -r -R
Add '-r' to root-install-flags if DEINSTALLDEPENDS is set to != "ALL". (This doesn't affect the "make update" target, it just makes "make deinstall DEINSTALLDEPENDS=YES" behave sane).
1999-09-09 00:01:19 +02:00
# used for "update" target
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
real-su-deinstall-flags+= -r
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
. endif
. ifdef PKG_VERBOSE
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
real-su-deinstall-flags+= -v
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Set UPDATE_RUNNING=YES when deinstalling due to a update or replace target. Use this to trigger appropriate args to pkg_delete so PKG_PRESERVE packages can be upgraded.
2004-11-02 01:03:09 +01:00
. ifdef PKG_PRESERVE
. if (${UPDATE_RUNNING} == "YES")
# used to update w/o removing any files
real-su-deinstall-flags+= -N -f
. endif
. endif
Let "make deinstall" remove any version installed, not only the one currently in pkgsrc. Enabled unconditionally by request of several people.
1999-08-21 03:17:59 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: real-su-deinstall
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
real-su-deinstall:
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
"make deinstall" now tries to just delete this particular package first before trying to delete all packages with the same PKGBASE. This is better semantics in a pkgviews world.
2003-09-05 17:00:10 +02:00
found="`${PKG_INFO} -e \"${PKGNAME}\" || ${TRUE}`"; \
case "$$found" in \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
"") found="`${PKG_BEST_EXISTS} \"${PKGWILDCARD}\" || ${TRUE}`" ;; \
"make deinstall" now tries to just delete this particular package first before trying to delete all packages with the same PKGBASE. This is better semantics in a pkgviews world.
2003-09-05 17:00:10 +02:00
esac; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
if [ "$$found" != "" ]; then \
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
${ECHO} Running ${PKG_DELETE} ${real-su-deinstall-flags} $$found ; \
${PKG_DELETE} ${real-su-deinstall-flags} $$found || ${TRUE} ; \
Let "make deinstall" remove any version installed, not only the one currently in pkgsrc. Enabled unconditionally by request of several people.
1999-08-21 03:17:59 +02:00
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if (${DEINSTALLDEPENDS} != "NO") && (${DEINSTALLDEPENDS} != "ALL")
Also remove pkgs installed via BUILD_DEPENDS, they are not noted in the installed pkg and this won't be caught by "pkg_delete -r". There are still situations where pkgs can be left behind if they were pulled in as BUILD_DEPENDS of any depends installed via the "depends" target. Maybe Makefile-level recursion should be used instead of "pkg_delete -r" to implement this...
2000-02-26 22:58:55 +01:00
@${SHCOMMENT} Also remove BUILD_DEPENDS:
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. for pkg in ${BUILD_DEPENDS:C/:.*$//}
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
found="`${PKG_BEST_EXISTS} \"${pkg}\" || ${TRUE}`"; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
if [ "$$found" != "" ]; then \
${ECHO} Running ${PKG_DELETE} $$found; \
${PKG_DELETE} ${real-su-deinstall-flags} $$found || ${TRUE}; \
Also remove pkgs installed via BUILD_DEPENDS, they are not noted in the installed pkg and this won't be caught by "pkg_delete -r". There are still situations where pkgs can be left behind if they were pulled in as BUILD_DEPENDS of any depends installed via the "depends" target. Maybe Makefile-level recursion should be used instead of "pkg_delete -r" to implement this...
2000-02-26 22:58:55 +01:00
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endfor
. endif # DEINSTALLDEPENDS
Replace deinstall-depends target with call to "pkg_delete -R" (now that we have it for sure).
1999-04-20 22:28:11 +02:00
@${RM} -f ${INSTALL_COOKIE} ${PACKAGE_COOKIE}
.endif # target(deinstall)
Allow (recursive) cleaning out and deinstallation of required packages: make clean CLEANDEPENDS=YES make deinstall DEINSTALLDEPENDS=YES Both CLEANDEPENDS and DEINSTALLDEPENDS default to NO, so this isn't different from the current behaviour.
1999-03-03 23:53:49 +01:00
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
################################################################
# Some more targets supplied for users' convenience
################################################################
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
# The 'update' target can be used to update a package and all
# currently installed packages that depend upon this package.
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: update
Do not attempt to update a package (deleting the package and all its dependencies) if IGNORE is set. This puts the 'update' target in line with 'install', 'package', and other targets that use the same checks.
2002-04-06 01:26:49 +02:00
.if !target(update)
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.if exists(${DDIR})
RESUMEUPDATE?= YES
Fix clean-update target to remove ${DDIR} if it didn't exist before. This prevents 'make update' from "resuming" an update that never even began. Thus, a sequence of make clean-update make update is now safe, no matter if 'make update' was started previously or not.
2000-07-18 10:33:49 +02:00
CLEAR_DIRLIST?= NO
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
update:
${_PKG_SILENT}${_PKG_DEBUG}${ECHO_MSG} \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
"${_PKGSRC_IN}> Resuming update for ${PKGNAME}"
Make UPDATE_TARGET=replace somewhat useful: don't remove the original package in this case.
2003-05-04 03:20:13 +02:00
. if ${REINSTALL} != "NO" && ${UPDATE_TARGET} != "replace"
Make REINSTALL=YES useful for 'make update' by explicitly deinstalling each package before attempting to install it. Fixes a problem if 'make clean-update' was called at the wrong time. Make the 'clean-update' target warn about a preserved ${DDIR} and usage of 'make update REINSTALL=YES' if necessary.
2000-07-21 08:30:04 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Set UPDATE_RUNNING=YES when deinstalling due to a update or replace target. Use this to trigger appropriate args to pkg_delete so PKG_PRESERVE packages can be upgraded.
2004-11-02 01:03:09 +01:00
${MAKE} ${MAKEFLAGS} deinstall UPDATE_RUNNING=YES DEINSTALLDEPENDS=ALL
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.else
RESUMEUPDATE?= NO
Fix clean-update target to remove ${DDIR} if it didn't exist before. This prevents 'make update' from "resuming" an update that never even began. Thus, a sequence of make clean-update make update is now safe, no matter if 'make update' was started previously or not.
2000-07-18 10:33:49 +02:00
CLEAR_DIRLIST?= YES
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
update:
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} ${DDIR}
Make UPDATE_TARGET=replace somewhat useful: don't remove the original package in this case.
2003-05-04 03:20:13 +02:00
. if ${UPDATE_TARGET} != "replace"
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG}if ${PKG_INFO} -qe ${PKGBASE}; then \
Correct whitespace after UPDATE_RUNNING changes
2004-11-02 09:31:45 +01:00
${MAKE} ${MAKEFLAGS} deinstall UPDATE_RUNNING=YES DEINSTALLDEPENDS=ALL \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
|| (${RM} ${DDIR} && ${FALSE}); \
Remove depend directory file in "update" target if the deinstallation fails because otherwise the next invokation of the "update" target will not attempt to remove the package again. Fix provided by Aaron J. Grier in PR pkg/15827.
2002-03-22 20:08:44 +01:00
fi
Make UPDATE_TARGET=replace somewhat useful: don't remove the original package in this case.
2003-05-04 03:20:13 +02:00
. endif
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.endif
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
${MAKE} ${MAKEFLAGS} ${UPDATE_TARGET} KEEP_WRKDIR=YES \
Support DEPENDS_TARGET="install clean"; from Allan Barrett in PR pkg/25059.
2004-04-05 10:06:07 +02:00
DEPENDS_TARGET=${DEPENDS_TARGET:Q}
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Fix broken "update" target.
2000-06-03 22:26:05 +02:00
[ ! -s ${DDIR} ] || for dep in `${CAT} ${DDIR}` ; do \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
(if cd ../.. && cd "$${dep}" ; then \
Standardise on the format of continuation lines. No functional change.
2002-05-13 13:43:05 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> Installing in $${dep}" && \
Make UPDATE_TARGET=replace somewhat useful: don't remove the original package in this case.
2003-05-04 03:20:13 +02:00
if [ "(" "${RESUMEUPDATE}" = "NO" -o \
"${REINSTALL}" != "NO" ")" -a \
"${UPDATE_TARGET}" != "replace" ] ; then \
Correct whitespace after UPDATE_RUNNING changes
2004-11-02 09:31:45 +01:00
${MAKE} ${MAKEFLAGS} deinstall UPDATE_RUNNING=YES; \
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
fi && \
Use UPDATE_TARGET (new) instead of DEPENDS_TARGET for 'make update'. This allows the user to set DEPENDS_TARGET=update, which will now automatically update all outdated prerequisite packages.
2000-10-22 10:22:24 +02:00
${MAKE} ${MAKEFLAGS} ${UPDATE_TARGET} \
Support DEPENDS_TARGET="install clean"; from Allan Barrett in PR pkg/25059.
2004-04-05 10:06:07 +02:00
DEPENDS_TARGET=${DEPENDS_TARGET:Q} ; \
Change the 'update' target to skip non-existent (removed) directories. This fixes PR pkg/9181 by Richard Rauch.
2000-02-01 17:30:07 +01:00
else \
Standardise on the format of continuation lines. No functional change.
2002-05-13 13:43:05 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> Skipping removed directory $${dep}"; \
Change the 'update' target to skip non-existent (removed) directories. This fixes PR pkg/9181 by Richard Rauch.
2000-02-01 17:30:07 +01:00
fi) ; \
done
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.if ${NOCLEAN} == "NO"
Change the 'update' target to skip non-existent (removed) directories. This fixes PR pkg/9181 by Richard Rauch.
2000-02-01 17:30:07 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${MAKE} ${MAKEFLAGS} clean-update CLEAR_DIRLIST=YES
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: clean-update
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
clean-update:
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} ${DDIR}
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
if [ -s ${DDIR} ] ; then \
for dep in `${CAT} ${DDIR}` ; do \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
(if cd ../.. && cd "$${dep}" ; then \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${MAKE} ${MAKEFLAGS} clean ; \
Change the 'update' target to skip non-existent (removed) directories. This fixes PR pkg/9181 by Richard Rauch.
2000-02-01 17:30:07 +01:00
else \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Skipping removed directory $${dep}";\
Change the 'update' target to skip non-existent (removed) directories. This fixes PR pkg/9181 by Richard Rauch.
2000-02-01 17:30:07 +01:00
fi) ; \
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
done ; \
fi
Fix clean-update target to remove ${DDIR} if it didn't exist before. This prevents 'make update' from "resuming" an update that never even began. Thus, a sequence of make clean-update make update is now safe, no matter if 'make update' was started previously or not.
2000-07-18 10:33:49 +02:00
.if ${CLEAR_DIRLIST} != "NO"
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} clean
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.else
Change the 'update' target to skip non-existent (removed) directories. This fixes PR pkg/9181 by Richard Rauch.
2000-02-01 17:30:07 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${MAKE} ${MAKEFLAGS} clean update-dirlist \
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
DIRLIST="`${CAT} ${DDIR}`" PKGLIST="`${CAT} ${DLIST}`"
Make REINSTALL=YES useful for 'make update' by explicitly deinstalling each package before attempting to install it. Fixes a problem if 'make clean-update' was called at the wrong time. Make the 'clean-update' target warn about a preserved ${DDIR} and usage of 'make update REINSTALL=YES' if necessary.
2000-07-21 08:30:04 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${ECHO_MSG} \
"${_PKGSRC_IN}> Warning: preserved leftover directory list. Your next";\
${ECHO_MSG} "${_PKGSRC_IN}> \`\`${MAKE} update'' may fail. It is advised to use";\
${ECHO_MSG} "${_PKGSRC_IN}> \`\`${MAKE} update REINSTALL=YES'' instead!"
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.endif
Do not attempt to update a package (deleting the package and all its dependencies) if IGNORE is set. This puts the 'update' target in line with 'install', 'package', and other targets that use the same checks.
2002-04-06 01:26:49 +02:00
.endif # !target(update)
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: update-dirlist
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
update-dirlist:
${_PKG_SILENT}${_PKG_DEBUG}${MKDIR} -p ${WRKDIR}
.ifdef PKGLIST
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. for __tmp__ in ${PKGLIST}
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${ECHO} >>${DLIST} "${__tmp__}"
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endfor
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.endif
.ifdef DIRLIST
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. for __tmp__ in ${DIRLIST}
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${ECHO} >>${DDIR} "${__tmp__}"
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endfor
Change the "update" target so it becomes more useful (and less dangerous!). Here is a summary of the changed behaviour: * A "make update" will no longer clean the subtree before running. * If called a second time, it will resume package installation, in case a previous "make update" was interrupted for some reason. (just make sure you don't call "make clean" in between ;-) * Unless "NOCLEAN" is set, "make update" will clean up the subtree after successfully updating _all_ packages (to avoid a dirty directory structure for future makes). * A "clean-update" target has been introduced to clean up the subtree that gets updated. This should be used prior to an initial "make update" if you suspect that one of the packages that get updated may be dirty. XXX: this still does not automatically check for or update any obsolete ´upstream' packages (e.g., if gtk depends on glib, a "make update" in the gtk directory won't automatically update glib, you have to do it the other way round).
1999-10-31 20:43:02 +01:00
.endif
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
${DDIR}: ${DLIST}
${_PKG_SILENT}${_PKG_DEBUG} \
Don't do input redirection for sed if not necessary.
2001-01-29 15:40:26 +01:00
ddir=`${SED} 's:-[^-]*$$::' ${DLIST}`; \
Fix the ${DDIR} target to work with multi-component paths. This makes the update target work with pkgsrc installations that have more (or less) than two components (like /src/NetBSD-current/pkgsrc). Thanks for Matthias Scheler for noting this!
2000-07-18 10:21:41 +02:00
${ECHO} >${DDIR}; \
for pkg in $${ddir} ; do \
quote args to pkg_info, to prevent some shell trying to interpret them.
2002-09-16 12:04:39 +02:00
if ${PKG_INFO} -b "$${pkg}" >/dev/null 2>&1 ; then \
${PKG_INFO} -b "$${pkg}" | ${SED} -ne \
Revise previous multi-component-pathname fix: do not depend on the directory name of a package being the same as the package name when creating ${DDIR} entries.
2000-07-19 02:08:32 +02:00
's,\([^/]*/[^/]*\)/Makefile:.*,\1,p' | \
${HEAD} -1 >>${DDIR}; \
Fix the ${DDIR} target to work with multi-component paths. This makes the update target work with pkgsrc installations that have more (or less) than two components (like /src/NetBSD-current/pkgsrc). Thanks for Matthias Scheler for noting this!
2000-07-18 10:21:41 +02:00
fi ; \
done
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
When creating the ${DLIST} file in the update target, don't create the ${WRKDIR} every time with "mkdir -p" - instead, use the ${WRKDIR} target, so that any symlinks to it are created properly. Also, when creating the ${DLIST} file in the update target, don't pipe the output of pkg_info -R through "tail -n +4" - simply use pkg_info -qR to achieve the same results.
2002-04-18 17:54:34 +02:00
${DLIST}: ${WRKDIR}
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Add another fix to previous; reverse the lines in the output. From pkg_delete, recursion is depth-first, which is wrong for "make update". Instead, make sure that "closer" dependencies are at the top of the file.
2004-04-16 16:27:38 +02:00
{ ${PKG_DELETE} -n "${PKGWILDCARD}" 2>&1 | \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
${GREP} '^ ' | \
Add another fix to previous; reverse the lines in the output. From pkg_delete, recursion is depth-first, which is wrong for "make update". Instead, make sure that "closer" dependencies are at the top of the file.
2004-04-16 16:27:38 +02:00
${AWK} '{ l[NR]=$$0 } END { for (i=NR;i>0;--i) print l[i] }' \
|| ${TRUE}; } > ${DLIST}
Add 'update' target to update a package and all depending packages.
1999-08-30 00:13:01 +02:00
Add new "info" target which calls "pkg_info" for the package in the current directory.
2000-07-30 10:52:41 +02:00
# The 'info' target can be used to display information about a package.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: info
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
info:
quote args to pkg_info, to prevent some shell trying to interpret them.
2002-09-16 12:04:39 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${PKG_INFO} "${PKGWILDCARD}"
Add new "info" target which calls "pkg_info" for the package in the current directory.
2000-07-30 10:52:41 +02:00
Add new "check" target which invokes "pkg_admin check" for the current package.
2000-11-12 16:20:06 +01:00
# The 'check' target can be used to check an installed package.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: check
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
check:
quote args to pkg_info, to prevent some shell trying to interpret them.
2002-09-16 12:04:39 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${PKG_ADMIN} check "${PKGWILDCARD}"
Add new "check" target which invokes "pkg_admin check" for the current package.
2000-11-12 16:20:06 +01:00
Add a "list" target which invokes "pkg_info -L" for the current package.
2002-09-08 08:59:26 +02:00
# The 'list' target can be used to list the files installed by a package.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: list
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
list:
quote args to pkg_info, to prevent some shell trying to interpret them.
2002-09-16 12:04:39 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${PKG_INFO} -L "${PKGWILDCARD}"
Add a "list" target which invokes "pkg_info -L" for the current package.
2002-09-08 08:59:26 +02:00
* Remove old PLIST-mi/md.shared/md.static that was only used for perl in a previous (pkgsrc) life. * add "lint" target to run pkglint * Add commented out direction on how to use bzip2 compressed binary packages * Some cleanup (clean target, ...) Reviewed by Thomas Klauser <wiz@netbsd.org>.
2000-12-30 12:19:04 +01:00
# Run pkglint:
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: lint
* Remove old PLIST-mi/md.shared/md.static that was only used for perl in a previous (pkgsrc) life. * add "lint" target to run pkglint * Add commented out direction on how to use bzip2 compressed binary packages * Some cleanup (clean target, ...) Reviewed by Thomas Klauser <wiz@netbsd.org>.
2000-12-30 12:19:04 +01:00
lint:
lint: no news is good news, don't tell us about "OK" items from pkglint.
2003-02-15 01:26:58 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${LOCALBASE}/bin/pkglint | ${GREP} -v ^OK
Add new "tarup" target which allows to create a binary package from an installed package with the "pkg_tarup" utility.
2001-07-10 13:51:57 +02:00
# Create a binary package from an install package using "pkg_tarup"
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: tarup
Add new "tarup" target which allows to create a binary package from an installed package with the "pkg_tarup" utility.
2001-07-10 13:51:57 +02:00
tarup:
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if ${PKG_INSTALLATION_TYPE} == "overwrite"
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
${RM} -f ${PACKAGES}/All/${PKGNAME}${PKG_SUFX}; \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
${SETENV} PKG_DBDIR=${_PKG_DBDIR} PKG_SUFX=${PKG_SUFX} \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
PKGREPOSITORY=${PACKAGES}/All \
${LOCALBASE}/bin/pkg_tarup ${PKGNAME}; \
for CATEGORY in ${CATEGORIES}; do \
${MKDIR} ${PACKAGES}/$$CATEGORY; \
cd ${PACKAGES}/$$CATEGORY; \
${RM} -f ${PKGNAME}${PKG_SUFX}; \
${LN} -s ../All/${PKGNAME}${PKG_SUFX}; \
Add new "tarup" target which allows to create a binary package from an installed package with the "pkg_tarup" utility.
2001-07-10 13:51:57 +02:00
done
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.endif
* Remove old PLIST-mi/md.shared/md.static that was only used for perl in a previous (pkgsrc) life. * add "lint" target to run pkglint * Add commented out direction on how to use bzip2 compressed binary packages * Some cleanup (clean target, ...) Reviewed by Thomas Klauser <wiz@netbsd.org>.
2000-12-30 12:19:04 +01:00
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
# shared code for replace and undo-replace
_REPLACE= \
Correct whitespace after UPDATE_RUNNING changes
2004-11-02 09:31:45 +01:00
if [ -f ${_PKG_DBDIR}/$$oldpkgname/+REQUIRED_BY ]; then \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
${MV} ${_PKG_DBDIR}/$$oldpkgname/+REQUIRED_BY ${WRKDIR}/.req; \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
fi; \
Correct whitespace after UPDATE_RUNNING changes
2004-11-02 09:31:45 +01:00
${MAKE} deinstall UPDATE_RUNNING=YES; \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
$$replace_action; \
if [ -f ${WRKDIR}/.req ]; then \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
${MV} ${WRKDIR}/.req ${_PKG_DBDIR}/$$newpkgname/+REQUIRED_BY; \
for pkg in `${CAT} ${_PKG_DBDIR}/$$newpkgname/+REQUIRED_BY`; do \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
${SETENV} NEWPKGNAME=$$newpkgname \
${AWK} '/^@pkgdep '$$oldpkgname'/ { print "@pkgdep " ENVIRON["NEWPKGNAME"]; next } { print }' \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
< ${_PKG_DBDIR}/$$pkg/+CONTENTS > ${_PKG_DBDIR}/$$pkg/+CONTENTS.$$$$ && \
${MV} ${_PKG_DBDIR}/$$pkg/+CONTENTS.$$$$ ${_PKG_DBDIR}/$$pkg/+CONTENTS; \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
done; \
fi
# replace a package in place - not for the faint-hearted
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: real-su-replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
real-su-replace:
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
${ECHO_MSG} "*** WARNING - experimental target - data loss may be experienced ***"; \
Make the "replace" and "undo-replace" targets more useful: + Check for the existence of pkg_tarup, and fail if it can't be found. + Place the binary package of the replaced-package in the ${WRKDIR}, rather than overwriting any instance in ${PKGREPOSITORY}. + Don't make a binary package when undoing the replacement.
2002-04-02 11:41:39 +02:00
if [ -x ${LOCALBASE}/bin/pkg_tarup ]; then \
${SETENV} PKGREPOSITORY=${WRKDIR} ${LOCALBASE}/bin/pkg_tarup ${PKGBASE}; \
else \
${ECHO} "No ${LOCALBASE}/bin/pkg_tarup binary - can't pkg_tarup ${PKGBASE}"; \
exit 1; \
fi
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Fix cases where we were calling "${PKG_BEST_EXISTS} <pkgbase>". <pkgbase> isn't actually a pkgpattern; what we really want is <pkgwildcard>. This should fix breakage with the "update" and "replace" targets introduced in revisions 1.1275 and 1.1278.
2003-09-13 20:26:15 +02:00
oldpkgname=`${PKG_BEST_EXISTS} "${PKGWILDCARD}"`; \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
newpkgname=${PKGNAME}; \
${ECHO} "$$oldpkgname" > ${WRKDIR}/.replace; \
replace_action="${MAKE} install"; \
${_REPLACE}
# undo the replacement of a package - not for the faint-hearted either
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: real-su-undo-replace
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
real-su-undo-replace:
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
if [ ! -f ${WRKDIR}/.replace ]; then \
${ECHO_MSG} "No replacement to undo"; \
exit 1; \
Make the "replace" and "undo-replace" targets more useful: + Check for the existence of pkg_tarup, and fail if it can't be found. + Place the binary package of the replaced-package in the ${WRKDIR}, rather than overwriting any instance in ${PKGREPOSITORY}. + Don't make a binary package when undoing the replacement.
2002-04-02 11:41:39 +02:00
fi
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
${ECHO_MSG} "*** WARNING - experimental target - data loss may be experienced ***"; \
oldpkgname=${PKGNAME}; \
newpkgname=`${CAT} ${WRKDIR}/.replace`; \
Make the "replace" and "undo-replace" targets more useful: + Check for the existence of pkg_tarup, and fail if it can't be found. + Place the binary package of the replaced-package in the ${WRKDIR}, rather than overwriting any instance in ${PKGREPOSITORY}. + Don't make a binary package when undoing the replacement.
2002-04-02 11:41:39 +02:00
replace_action="${SETENV} ${PKG_ADD} ${WRKDIR}/$$newpkgname${PKG_SUFX}"; \
Add two new EXPERIMENTAL targets: replace - this target updates a package in place, fixing up references to and from other packages to the updated package. undo-replace - undoes all the work of the previous replace operation. For this target to work, the ${WRKDIR} must be preserved. The replace target first makes a binary package of the existing installed package, then a copy of the +REQUIRED_BY file is taken, if it exists, and then the existing package is deleted. The new package is installed, and the preserved +REQUIRED_BY file is copied back into place, using its contents to modify the +CONTENTS files of all the packages which require it. The undo-replace shares code with the replace target, and does the same operation, but in reverse. Warnings are displayed when either of these targets is used, since they are experimental targets right now.
2002-03-04 12:56:54 +01:00
${_REPLACE}; \
${RM} ${WRKDIR}/.replace
Move mirror-distfiles target around a bit so it can even be invoked on platforms that are prohibited by ONLY_FOR_ARCH/OPSYS etc. so we can provide distfile-backups even for them.
1999-04-09 04:06:53 +02:00
# This is for the use of sites which store distfiles which others may
# fetch - only fetch the distfile if it is allowed to be
# re-distributed freely
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: mirror-distfiles
Move mirror-distfiles target around a bit so it can even be invoked on platforms that are prohibited by ONLY_FOR_ARCH/OPSYS etc. so we can provide distfile-backups even for them.
1999-04-09 04:06:53 +02:00
mirror-distfiles:
* Prevent " etc. in RESTRICTED (NO_{SRC,BIN}_ON_{CDROM,FTP}) and BROKEN from producing errors * Rework IGNORE handling: - IGNORE messages must now include the pkg's name - if multiple reasons exist for a pkg to be ignored, they are now all printed - IGNORECMD is gone * Make 'mirror-distfiles' not print a warning if NO_SRC_ON_FTP is set (I really don't want to see that list every day!) * Catch one leftover 'depends' -> 'install-depends'
2000-08-19 17:32:15 +02:00
.if !defined(NO_SRC_ON_FTP)
Garbage collect NO_CHECK_DEPENDS: it is meaningless since revision 1.942.
2004-07-31 14:24:02 +02:00
@${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} fetch NO_SKIP=yes
Move mirror-distfiles target around a bit so it can even be invoked on platforms that are prohibited by ONLY_FOR_ARCH/OPSYS etc. so we can provide distfile-backups even for them.
1999-04-09 04:06:53 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Cleaning up
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: pre-clean
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(pre-clean)
pre-clean:
@${DO_NADA}
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: clean
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(clean)
clean: pre-clean
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
. if (${CLEANDEPENDS} != "NO") && (!empty(BUILD_DEPENDS) || !empty(DEPENDS))
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} clean-depends
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
@${ECHO_MSG} "${_PKGSRC_IN}> Cleaning for ${PKGNAME}"
Clean up `make clean' a little. In particular, make it safe with WRKOBJDIR when doing a simultaneous build on another platform.
2000-06-03 21:20:55 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
if [ -d ${WRKDIR} ]; then \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
if [ -w ${WRKDIR} ]; then \
${RM} -rf ${WRKDIR}; \
else \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> ${WRKDIR} not writable, skipping"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
fi; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. ifdef WRKOBJDIR
Clean up `make clean' a little. In particular, make it safe with WRKOBJDIR when doing a simultaneous build on another platform.
2000-06-03 21:20:55 +02:00
-${_PKG_SILENT}${_PKG_DEBUG} \
* Fix a bug in clean-depends: the `sort -u' was ineffective because the `tr' was missing. * Do make(1) pattern substitutions for fetch-list-recursive. * Eliminate subshells in clean-depends and fetch-list-recursive. * Make the rmdir for the WRKOBJDIR removal silent if it fails. * Slight optimization to fetch-list-one-pkg: don't do the basename operation twice.
2000-06-03 23:51:57 +02:00
${RMDIR} ${BUILD_DIR} 2>/dev/null; \
Clean up `make clean' a little. In particular, make it safe with WRKOBJDIR when doing a simultaneous build on another platform.
2000-06-03 21:20:55 +02:00
${RM} -f ${WRKDIR_BASENAME}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
* Move the clean-depends target near the clear target * some whitespace change to enhance readability
2000-05-31 18:59:44 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: clean-depends
* Move the clean-depends target near the clear target * some whitespace change to enhance readability
2000-05-31 18:59:44 +02:00
.if !target(clean-depends)
clean-depends:
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
. if !empty(BUILD_DEPENDS) || !empty(DEPENDS)
* Fix a bug in clean-depends: the `sort -u' was ineffective because the `tr' was missing. * Do make(1) pattern substitutions for fetch-list-recursive. * Eliminate subshells in clean-depends and fetch-list-recursive. * Make the rmdir for the WRKOBJDIR removal silent if it fails. * Slight optimization to fetch-list-one-pkg: don't do the basename operation twice.
2000-06-03 23:51:57 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
for i in `${MAKE} ${MAKEFLAGS} show-all-depends-dirs-excl`; do \
rework how the DEPENDS tree is traversed in the 'clean-depends' target. now, each branch is only traversed once instead of possibly multiple times. depending on the size of the DEPENDS list, the speedup is _huge_. For example make clean CLEANDEPENDS=yes in x11/gnome now takes around 12 minutes instead of 4 hours on a DECstation 5000/240. Thanks to Hubert for help and suggestions on this. Look for more changes to come as this same problem exists in other areas.
2000-06-28 18:05:44 +02:00
cd ${.CURDIR}/../../$$i && \
${MAKE} ${MAKEFLAGS} CLEANDEPENDS=NO clean; \
* Move the clean-depends target near the clear target * some whitespace change to enhance readability
2000-05-31 18:59:44 +02:00
done
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
eliminate a few more makes from being forked off in the clean-depends area
2000-06-28 19:35:53 +02:00
.endif
* Move the clean-depends target near the clear target * some whitespace change to enhance readability
2000-05-31 18:59:44 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: pre-distclean
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(pre-distclean)
pre-distclean:
@${DO_NADA}
.endif
* Move the clean-depends target near the clear target * some whitespace change to enhance readability
2000-05-31 18:59:44 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: cleandir
Add "cleandir" target to comform with the rest of our source tree.
1998-10-04 22:56:58 +02:00
.if !target(cleandir)
cleandir: clean
.endif
* Move the clean-depends target near the clear target * some whitespace change to enhance readability
2000-05-31 18:59:44 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: distclean
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(distclean)
distclean: pre-distclean clean
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${_PKG_SILENT}${ECHO_MSG} "${_PKGSRC_IN}> Dist cleaning for ${PKGNAME}"
Remove a bunch of unneeded ()s. Inspired by OpenBSD.
2000-06-03 01:30:52 +02:00
${_PKG_SILENT}${_PKG_DEBUG}if [ -d ${_DISTDIR} ]; then \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
cd ${_DISTDIR} && \
${TEST} -z "${DISTFILES}" || ${RM} -f ${DISTFILES}; \
In the distclean target, if PKG_RESUME_TRANSFERS is set, clean the temp files too.
2005-03-22 23:49:15 +01:00
if [ "${PKG_RESUME_TRANSFERS:M[Yy][Ee][Ss]}" ]; then \
${TEST} -z "${DISTFILES}.temp" || ${RM} -f ${DISTFILES}.temp; \
fi; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${TEST} -z "${PATCHFILES}" || ${RM} -f ${PATCHFILES}; \
Remove a bunch of unneeded ()s. Inspired by OpenBSD.
2000-06-03 01:30:52 +02:00
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(DIST_SUBDIR) && exists(DIST_SUBDIR)
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
-${_PKG_SILENT}${_PKG_DEBUG}${RMDIR} ${_DISTDIR}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Upon "make distclean", also rm the README.html file.
1999-09-08 16:58:34 +02:00
-${_PKG_SILENT}${_PKG_DEBUG}${RM} -f README.html
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
# Prints out a script to fetch all needed files (no checksumming).
.PHONY: fetch-list
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.if !target(fetch-list)
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fetch-list:
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
@${ECHO} '#!/bin/sh'
@${ECHO} '#'
@${ECHO} '# This is an auto-generated script, the result of running'
Make BUILD_DIR always point to a physical path by invoking /bin/pwd instead of relying on the shell's builtin pwd. This makes BUILD_DIR consistently point to the correct directory regardless of the definition of SHELL. This fixes elusive some buildlink2 errors due to the fact that BUILDLINK_DIR is derived from BUILD_DIR and some paths were incorrectly being translated.
2003-01-15 21:55:38 +01:00
@${ECHO} '# `${MAKE} fetch-list'"'"' in directory "'"`${PWD_CMD}`"'"'
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
@${ECHO} '# on host "'"`${UNAME} -n`"'" on "'"`date`"'".'
@${ECHO} '#'
@${MAKE} ${MAKEFLAGS} fetch-list-recursive
Recursively print /bin/sh code for all dependent package as well as the current package (IF called from a package dir and not via bsd.port.subdir.mk; in the latter case it's assumed that the dependent packages will be scanned later anyways). This fixes PR 4992 by Mario Kemper <magick@bundy.lip.owl.de>.
1998-02-20 22:31:20 +01:00
.endif # !target(fetch-list)
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.PHONY: fetch-list-recursive
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.if !target(fetch-list-recursive)
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
Recursively print /bin/sh code for all dependent package as well as the current package (IF called from a package dir and not via bsd.port.subdir.mk; in the latter case it's assumed that the dependent packages will be scanned later anyways). This fixes PR 4992 by Mario Kemper <magick@bundy.lip.owl.de>.
1998-02-20 22:31:20 +01:00
fetch-list-recursive:
* Fix a bug in clean-depends: the `sort -u' was ineffective because the `tr' was missing. * Do make(1) pattern substitutions for fetch-list-recursive. * Eliminate subshells in clean-depends and fetch-list-recursive. * Make the rmdir for the WRKOBJDIR removal silent if it fails. * Slight optimization to fetch-list-one-pkg: don't do the basename operation twice.
2000-06-03 23:51:57 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
for dir in `${MAKE} ${MAKEFLAGS} show-all-depends-dirs`; do \
(cd ../../$$dir && \
${MAKE} ${MAKEFLAGS} fetch-list-one-pkg \
| ${AWK} ' \
/^[^#]/ { FoundSomething = 1 } \
/^unsorted/ { gsub(/[[:space:]]+/, " \\\n\t") } \
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
/^echo/ { gsub(/;[[:space:]]+/, "\n") } \
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
{ block[line_c++] = $$0 } \
END { if (FoundSomething) \
for (line = 0; line < line_c; line++) \
print block[line] } \
') \
Recursively print /bin/sh code for all dependent package as well as the current package (IF called from a package dir and not via bsd.port.subdir.mk; in the latter case it's assumed that the dependent packages will be scanned later anyways). This fixes PR 4992 by Mario Kemper <magick@bundy.lip.owl.de>.
1998-02-20 22:31:20 +01:00
done
.endif # !target(fetch-list-recursive)
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
.PHONY: fetch-list-one-pkg
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.if !target(fetch-list-one-pkg)
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
Recursively print /bin/sh code for all dependent package as well as the current package (IF called from a package dir and not via bsd.port.subdir.mk; in the latter case it's assumed that the dependent packages will be scanned later anyways). This fixes PR 4992 by Mario Kemper <magick@bundy.lip.owl.de>.
1998-02-20 22:31:20 +01:00
fetch-list-one-pkg:
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
. if !empty(_ALLFILES)
@${ECHO}
@${ECHO} '#'
Make BUILD_DIR always point to a physical path by invoking /bin/pwd instead of relying on the shell's builtin pwd. This makes BUILD_DIR consistently point to the correct directory regardless of the definition of SHELL. This fixes elusive some buildlink2 errors due to the fact that BUILDLINK_DIR is derived from BUILD_DIR and some paths were incorrectly being translated.
2003-01-15 21:55:38 +01:00
@location=`${PWD_CMD} | ${AWK} -F / '{ print $$(NF-1) "/" $$NF }'`; \
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
${ECHO} '# Need additional files for ${PKGNAME} ('$$location')...'
@${ECHO} '#'
@${MKDIR} ${_DISTDIR}
. for fetchfile in ${_ALLFILES}
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
. if defined(_FETCH_MESSAGE)
@(cd ${_DISTDIR}; \
if [ ! -f ${fetchfile:T} ]; then \
${ECHO}; \
filesize=`${AWK} ' \
/^Size/ && $$2 == "(${fetchfile})" { print $$4 } \
' ${DISTINFO_FILE}` || true; \
${ECHO} '# Prompt user to get ${fetchfile} ('$${filesize-???}' bytes) manually:'; \
${ECHO} '#'; \
${ECHO} ${_FETCH_MESSAGE:Q}; \
fi)
Some tweaks to "fetch" and "fetch-list" re: DYNAMIC_MASTER_SITES: - - For "fetch", move the assignment of the "getsite.sh" script to a more natural place, to get the full benefit of, and consistent handling for, setting MASTER_SORT and MASTER_SORT_REGEX - - For "fetch-list", embed the "getsite.sh" script into the generated script, so that the calculation can be properly done at run-time. ("fetch-list" also respects MASTER_SORT and MASTER_SORT_REGEX.)
2002-03-10 18:15:29 +01:00
. elif defined(DYNAMIC_MASTER_SITES)
@(cd ${_DISTDIR}; \
if [ ! -f ${fetchfile:T} ]; then \
${ECHO}; \
filesize=`${AWK} ' \
/^Size/ && $$2 == "(${fetchfile})" { print $$4 } \
' ${DISTINFO_FILE}` || true; \
${ECHO} '# Fetch ${fetchfile} ('$${filesize-???}' bytes):'; \
${ECHO} '#'; \
${ECHO} '${SH} -s ${fetchfile:T} <<"EOF" |('; \
${CAT} ${FILESDIR}/getsite.sh; \
${ECHO} EOF; \
${ECHO} read unsorted_sites; \
${ECHO} 'unsorted_sites="$${unsorted_sites} ${_MASTER_SITE_BACKUP}"'; \
${ECHO} sites='"'${ORDERED_SITES:Q}'"'; \
${ECHO} "${MKDIR} ${_DISTDIR}"; \
${ECHO} 'cd ${_DISTDIR} && [ -f ${fetchfile} -o -f ${fetchfile:T} ] ||'; \
${ECHO} 'for site in $$sites; do'; \
${ECHO} ' ${FETCH_CMD} ${FETCH_BEFORE_ARGS} "$${site}${fetchfile:T}" ${FETCH_AFTER_ARGS} && break ||'; \
${ECHO} ' ${ECHO} ${fetchfile} not fetched'; \
${ECHO} done; \
${ECHO} ')'; \
fi)
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
. else
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
@(cd ${_DISTDIR}; \
Generalize the handling for packages where "fetch" and "fetch-list" only emit a message and don't actually fetch anything. This allows us to make the output of "fetch-list" for these packages consistent with other packages. While we're in here, integrate DYNAMIC_MASTER_SITES with the ${ORDERED_SITES} macro. The only functional change here is that ${MASTER_SITE_OVERRIDE} is now respected. Still to do -- something appropriate for "fetch-list" for these packages, like sourcing "getsites.sh" into the generated script. (Well, "package", but there are two others that do something similar in their "Makefile".) Also eliminate the misbegotten _FETCH_ALLFILES macro -- now that only "fetch" uses it, move it's functionality directly under "do-fetch".
2002-03-04 20:41:03 +01:00
if [ ! -f ${fetchfile:T} ]; then \
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
${ECHO}; \
filesize=`${AWK} ' \
/^Size/ && $$2 == "(${fetchfile})" { print $$4 } \
' ${DISTINFO_FILE}` || true; \
${ECHO} '# Fetch ${fetchfile} ('$${filesize-???}' bytes):'; \
${ECHO} '#'; \
Correctly fetch ALLFILES containing '=' characters
2002-09-29 03:38:32 +02:00
${ECHO} 'unsorted_sites="${SITES_${fetchfile:T:S/=/--/}} ${_MASTER_SITE_BACKUP}"'; \
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
${ECHO} sites='"'${ORDERED_SITES:Q}'"'; \
${ECHO} "${MKDIR} ${_DISTDIR}"; \
${ECHO} 'cd ${_DISTDIR} && [ -f ${fetchfile} -o -f ${fetchfile:T} ] ||'; \
${ECHO} 'for site in $$sites; do'; \
${ECHO} ' ${FETCH_CMD} ${FETCH_BEFORE_ARGS} "$${site}${fetchfile:T}" ${FETCH_AFTER_ARGS} && break ||'; \
${ECHO} ' ${ECHO} ${fetchfile} not fetched'; \
${ECHO} done; \
fi)
Some tweaks to "fetch" and "fetch-list" re: DYNAMIC_MASTER_SITES: - - For "fetch", move the assignment of the "getsite.sh" script to a more natural place, to get the full benefit of, and consistent handling for, setting MASTER_SORT and MASTER_SORT_REGEX - - For "fetch-list", embed the "getsite.sh" script into the generated script, so that the calculation can be properly done at run-time. ("fetch-list" also respects MASTER_SORT and MASTER_SORT_REGEX.)
2002-03-10 18:15:29 +01:00
. endif # defined(_FETCH_MESSAGE) || defined(DYNAMIC_MASTER_SITES)
Comprehensive update to the fetch-list target (and friends), incorporating many suggestions made by Bernd Ernesti on "tech-pkg". Closes PR pkg/15539. - - Partially revert 1.895, which made 'fetch-list' ape 'fetch', so instead check if a file exists on the building host before emitting anything for it into the generated script. Update to use the new 'fetch' procedure on the fetching host, once it is determined that the file is needed, and generally clean up the output, so that it's a human readable shell script with comments. - - Introduce utility target 'show-all-depends-dirs', to do the necessary recursing over dependencies quickly and correctly. Add a couple of variations to show how it's done, and bracket the whole thing with ".if make()", so it's invisible to the dependency graph for all the other targets. The key improvement over the old way, is that it utilizes an "awk" associative array to guarantee that no package subdirectory is visited twice. For the package with the hairiest dependency graph, "x11/gnome", this leads to something like a 1,000-fold improvement! - - "sort -u" over the result of 'fetch-list-one-pkg' goes bye-bye. The new "awk" post-processor begins emitting output as soon as there is a file to fetch, so the user doesn't have to wait for the whole thing to finish before he sees anything. - - Get rid of the old hook for non-recursing for a top-level 'fetch-list', which was never used. Previously, a top-level 'fetch-list' couldn't work anyway, because the progress reports ruined the shell script. Remove 'fetch-list' from the targets that recurse from the top, and instead use a different 'fetch-list' target in "bsd.pkg.subdir.mk" to do the right thing. Currently, that right thing includes recursing in the single-category case, or if ${SPECIFIC_PKGS} is set, and not recursing in the top-level, all packages case (where it would be perfectly pointless).
2002-02-28 15:42:39 +01:00
. endfor
. endif # !empty(_ALLFILES)
Recursively print /bin/sh code for all dependent package as well as the current package (IF called from a package dir and not via bsd.port.subdir.mk; in the latter case it's assumed that the dependent packages will be scanned later anyways). This fixes PR 4992 by Mario Kemper <magick@bundy.lip.owl.de>.
1998-02-20 22:31:20 +01:00
.endif # !target(fetch-list-one-pkg)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Checksumming utilities
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: makesum
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(makesum)
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
makesum: fetch uptodate-digest
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
newfile=${DISTINFO_FILE}.$$$$; \
Don't change distinfo file if no checksums/sizes are modified.
2001-04-21 14:40:58 +02:00
if [ -f ${DISTINFO_FILE} ]; then \
Reduce chances that distinfo ends up with a bogus RCS Id or none at all.
2002-07-02 15:31:40 +02:00
${GREP} '^.NetBSD' ${DISTINFO_FILE} > $$newfile || \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
(${ECHO_N} "$$" > $$newfile && \
${ECHO_N} "NetBSD" >> $$newfile && \
Reduce chances that distinfo ends up with a bogus RCS Id or none at all.
2002-07-02 15:31:40 +02:00
${ECHO} "$$" >> $$newfile) \
Don't change distinfo file if no checksums/sizes are modified.
2001-04-21 14:40:58 +02:00
else \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "$$" > $$newfile; \
${ECHO_N} "NetBSD" >> $$newfile; \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
${ECHO} "$$" >> $$newfile; \
Don't change distinfo file if no checksums/sizes are modified.
2001-04-21 14:40:58 +02:00
fi; \
${ECHO} "" >> $$newfile; \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
cd ${DISTDIR}; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
for sumfile in "" ${_CKSUMFILES}; do \
Fix a pasto, pointed out by John Refling in PR 10094, which prevented files from being ignored when making checksums for distfiles.
2000-05-11 10:14:35 +02:00
if [ "X$$sumfile" = "X" ]; then continue; fi; \
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
for a in "" ${DIGEST_ALGORITHMS}; do \
if [ "X$$a" = "X" ]; then continue; fi; \
${DIGEST} $$a $$sumfile >> $$newfile; \
done; \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
${WC} -c $$sumfile | ${AWK} '{ print "Size (" $$2 ") = " $$1 " bytes" }' >> $$newfile; \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
done; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
for ignore in "" ${_IGNOREFILES}; do \
Fix a pasto, pointed out by John Refling in PR 10094, which prevented files from being ignored when making checksums for distfiles.
2000-05-11 10:14:35 +02:00
if [ "X$$ignore" = "X" ]; then continue; fi; \
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
for a in "" ${DIGEST_ALGORITHMS}; do \
if [ "X$$a" = "X" ]; then continue; fi; \
${ECHO} "$$a ($$ignore) = IGNORE" >> $$newfile; \
done; \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
done; \
Deprecate DIGEST_FILE and PATCH_SUM_FILE (replaced by DIGEST_FILE, and not used anymore in pkgsrc).
2001-06-09 14:15:59 +02:00
if [ -f ${DISTINFO_FILE} ]; then \
${AWK} '$$2 ~ /\(patch-[a-z0-9]+\)/ { print $$0 }' < ${DISTINFO_FILE} >> $$newfile; \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
fi; \
Add definitions for CMP pointing to cmp utility and use it in bsd.pkg.mk.
2001-07-13 08:16:02 +02:00
if ${CMP} -s $$newfile ${DISTINFO_FILE}; then \
Don't change distinfo file if no checksums/sizes are modified.
2001-04-21 14:40:58 +02:00
${RM} -f $$newfile; \
Improve message if distinfo parts don't get modified. Closes pkg/12718.
2001-04-23 22:53:50 +02:00
${ECHO_MSG} "=> distinfo: distfiles part unchanged."; \
Don't change distinfo file if no checksums/sizes are modified.
2001-04-21 14:40:58 +02:00
else \
${MV} $$newfile ${DISTINFO_FILE}; \
fi
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
.if !target(makepatchsum)
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
makepatchsum mps: uptodate-digest
Update "makepatchsum" target to produce a placeholder patch-sum file if and only if there are no patches and patch-sum already exists. This is necessary to meet the original purpose of that target, which is to "catch" users who happen to have old patches lying around, for whatever reason. Also, don't update patch-sum if it wouldn't change.
1999-10-14 07:31:22 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
newfile=${DISTINFO_FILE}.$$$$; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
if [ -f ${DISTINFO_FILE} ]; then \
Deprecate DIGEST_FILE and PATCH_SUM_FILE (replaced by DIGEST_FILE, and not used anymore in pkgsrc).
2001-06-09 14:15:59 +02:00
${AWK} '$$2 !~ /\(patch-[a-z0-9]+\)/ { print $$0 }' < ${DISTINFO_FILE} >> $$newfile; \
Fix 'make mdi' when no previous distinfo-file exists.
2001-06-09 14:56:21 +02:00
else \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "$$" > $$newfile; \
${ECHO_N} "NetBSD" >> $$newfile; \
Fix 'make mdi' when no previous distinfo-file exists.
2001-06-09 14:56:21 +02:00
${ECHO} "$$" >> $$newfile; \
Reduce chances that distinfo ends up with a bogus RCS Id or none at all.
2002-07-02 15:31:40 +02:00
${ECHO} "" >> $$newfile; \
Update "makepatchsum" target to produce a placeholder patch-sum file if and only if there are no patches and patch-sum already exists. This is necessary to meet the original purpose of that target, which is to "catch" users who happen to have old patches lying around, for whatever reason. Also, don't update patch-sum if it wouldn't change.
1999-10-14 07:31:22 +02:00
fi; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
if [ -d ${PATCHDIR} ]; then \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
(cd ${PATCHDIR}; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
for sumfile in "" patch-*; do \
Update "makepatchsum" target to produce a placeholder patch-sum file if and only if there are no patches and patch-sum already exists. This is necessary to meet the original purpose of that target, which is to "catch" users who happen to have old patches lying around, for whatever reason. Also, don't update patch-sum if it wouldn't change.
1999-10-14 07:31:22 +02:00
if [ "X$$sumfile" = "X" ]; then continue; fi; \
if [ "X$$sumfile" = "Xpatch-*" ]; then break; fi; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
case $$sumfile in \
Update "makepatchsum" target to produce a placeholder patch-sum file if and only if there are no patches and patch-sum already exists. This is necessary to meet the original purpose of that target, which is to "catch" users who happen to have old patches lying around, for whatever reason. Also, don't update patch-sum if it wouldn't change.
1999-10-14 07:31:22 +02:00
patch-local-*) ;; \
Don't generate checksums for patches that will get ignored in the apply step anyway (*.orig, *.rej).
2001-04-21 14:56:56 +02:00
*.orig|*.rej|*~) continue ;; \
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
*) ${ECHO} "${PATCH_DIGEST_ALGORITHM} ($$sumfile) = `${SED} -e '/\$$NetBSD.*/d' $$sumfile | ${DIGEST} ${PATCH_DIGEST_ALGORITHM}`" >> $$newfile;; \
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
esac; \
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
done); \
Update "makepatchsum" target to produce a placeholder patch-sum file if and only if there are no patches and patch-sum already exists. This is necessary to meet the original purpose of that target, which is to "catch" users who happen to have old patches lying around, for whatever reason. Also, don't update patch-sum if it wouldn't change.
1999-10-14 07:31:22 +02:00
fi; \
Add definitions for CMP pointing to cmp utility and use it in bsd.pkg.mk.
2001-07-13 08:16:02 +02:00
if ${CMP} -s $$newfile ${DISTINFO_FILE}; then \
Don't change distinfo file if no checksums/sizes are modified.
2001-04-21 14:40:58 +02:00
${RM} -f $$newfile; \
Improve message if distinfo parts don't get modified. Closes pkg/12718.
2001-04-23 22:53:50 +02:00
${ECHO_MSG} "=> distinfo: patches part unchanged."; \
Don't change distinfo file if no checksums/sizes are modified.
2001-04-21 14:40:58 +02:00
else \
${MV} $$newfile ${DISTINFO_FILE}; \
fi
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
.endif
Deprecate DIGEST_FILE and PATCH_SUM_FILE (replaced by DIGEST_FILE, and not used anymore in pkgsrc).
2001-06-09 14:15:59 +02:00
# This target is done by invoking a sub-make so that DISTINFO_FILE gets
Distinfo changes. + move the distfile digest/checksum value from files/md5 to distinfo + move the patch digest/checksum values from files/patch-sum to distinfo + include distfile filesizes in distinfo If there's no distinfo file (the name comes from FreeBSD, no point in being gratuitously incompatible), then use existing files/{md5,patch-sum} files.
2001-04-17 11:07:01 +02:00
# re-evaluated after the "makepatchsum" target is made. This can be
# made into:
#makedistinfo mdi: makepatchsum makesum
# once a combined distinfo file exists for all packages
.if !target(makedistinfo)
makedistinfo mdi distinfo: makepatchsum
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} makesum
Add logic to provide md5 checksums on the patch files, placing them in a file called files/patch-sum. This has the benefit of only attempting to patch packages with official patches. This should mean that you can untar a pkgsrc.tar.gz over a previous tree, and still have it do the right thing, ignoring old patch files. Patches which have an entry in the files/patch-sum file will be applied, and then any local patches. If there are local patches to a package, these should go in patch files called patches/patch-local-* (these files are ignored when calculating the files/patch-sum file, and also when checking the patches/patch-* files). The calculation of the patch checksum ignores any NetBSD RCS Id in the patch-* files. This removes a nasty ordering dependency for developers making patches to packages. Patch file checksums will only be checked if the files/patch-sum file exists. Add a target "makepatchsum", analogous to "makesum", which makes a new files/patch-sum file.
1999-07-09 15:14:21 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: checksum
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(checksum)
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
checksum: fetch uptodate-digest
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Deprecate DIGEST_FILE and PATCH_SUM_FILE (replaced by DIGEST_FILE, and not used anymore in pkgsrc).
2001-06-09 14:15:59 +02:00
if [ ! -f ${DISTINFO_FILE} ]; then \
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
${ECHO_MSG} "=> No checksum file."; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
else \
Fix a problem introduced in the previous commit, noticed by Matthiases Scheler and Drochner. Handle missing checksums in distinfo properly. Tested by myself and tron.
2005-02-25 14:05:52 +01:00
(cd ${DISTDIR}; OK="true"; missing=""; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
for file in "" ${_CKSUMFILES}; do \
if [ "X$$file" = X"" ]; then continue; fi; \
Fix a problem introduced in the previous commit, noticed by Matthiases Scheler and Drochner. Handle missing checksums in distinfo properly. Tested by myself and tron.
2005-02-25 14:05:52 +01:00
filesummed=false; \
for a in ${DIGEST_ALGORITHMS}; do \
Don't fail if there are multiple digest values for the same file with the same algorithm in the distinfo file - pointed out by Robert Elz in PR 29973. Additional testing by wiz - thanks!
2005-04-16 11:26:22 +02:00
CKSUM2=`${AWK} 'NF == 4 && $$1 == "'$$a'" && $$2 == "('$$file')" && $$3 == "=" {print $$4; exit}' ${DISTINFO_FILE}`; \
Fix a problem introduced in the previous commit, noticed by Matthiases Scheler and Drochner. Handle missing checksums in distinfo properly. Tested by myself and tron.
2005-02-25 14:05:52 +01:00
case "$${CKSUM2}" in \
"") ${ECHO_MSG} "=> No $$a checksum recorded for $$file."; \
;; \
*) filesummed=true; \
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
CKSUM=`${DIGEST} $$a < $$file`; \
if [ "$$CKSUM2" = "IGNORE" ]; then \
${ECHO_MSG} "=> Checksum for $$file is set to IGNORE in checksum file even though"; \
${ECHO_MSG} " the file is not in the "'$$'"{IGNOREFILES} list."; \
OK="false"; \
elif [ "$$CKSUM" = "$$CKSUM2" ]; then \
${ECHO_MSG} "=> Checksum $$a OK for $$file."; \
else \
${ECHO_MSG} "=> Checksum $$a mismatch for $$file."; \
OK="false"; \
Fix a problem introduced in the previous commit, noticed by Matthiases Scheler and Drochner. Handle missing checksums in distinfo properly. Tested by myself and tron.
2005-02-25 14:05:52 +01:00
fi ;; \
esac; \
As seen on tech-pkg - allow multiple digest algorithms to be used to ensure the integrity of distfiles and dist patches. For now, the default algorithms are SHA1 and RMD160, set as a whitespace-separated list in the DIGEST_ALGORITHMS definition. The DIGEST_ALGORITHM definition is deprecated. Patchfiles will still use simply SHA1, since we are trying to detect a binary "has this file changed", rather than proect against tampering. In short, if someone can modify the patch file, they can modify the distinfo file holding its digest information. This value is set in the new PATCH_DIGEST_ALGORITHM definition. Triggered by the breaking of SHA1, as reported in http://www.schneier.com/blog/archives/2005/02/sha1_broken.html
2005-02-22 22:01:10 +01:00
done; \
Fix a problem introduced in the previous commit, noticed by Matthiases Scheler and Drochner. Handle missing checksums in distinfo properly. Tested by myself and tron.
2005-02-25 14:05:52 +01:00
case "$$filesummed" in \
false) missing="$$missing $$file"; \
OK=false ;; \
esac; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
done; \
for file in "" ${_IGNOREFILES}; do \
if [ "X$$file" = X"" ]; then continue; fi; \
Don't fail if there are multiple digest values for the same file with the same algorithm in the distinfo file - pointed out by Robert Elz in PR 29973. Additional testing by wiz - thanks!
2005-04-16 11:26:22 +02:00
CKSUM2=`${AWK} 'NF == 4 && $$3 == "=" && $$2 == "('$$file')"{print $$4; exit}' ${DISTINFO_FILE}`; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
if [ "$$CKSUM2" = "" ]; then \
In informational messages to the user, use '=>' in preference to '>>', so that cut-n-paste into send-pr will work correctly. Fixes PR 9173 from ITOH Yasufumi.
2000-03-10 17:07:37 +01:00
${ECHO_MSG} "=> No checksum recorded for $$file, file is in "'$$'"{IGNOREFILES} list."; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
OK="false"; \
elif [ "$$CKSUM2" != "IGNORE" ]; then \
Add support for message digests other than md5 into pkgsrc. This is done by using the pkgsrc/pkgtools/digest package, and adding dependencies in bsd.pkg.mk to make sure that that package is up to date. There is no discernable performance decrease in using sha1 checksums over md5 ones, and most people advocate the use of sha1 or rmd160 over md5. As part of this change, move the initial setting of LOCALBASE, X11BASE and CROSSBASE into bsd.prefs.mk from bsd.pkg.mk. Addresses PR 11809.
2001-03-06 15:50:45 +01:00
${ECHO_MSG} "=> Checksum for $$file is not set to IGNORE in checksum file even though"; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${ECHO_MSG} " the file is in the "'$$'"{IGNOREFILES} list."; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
OK="false"; \
fi; \
done; \
if [ "$$OK" != "true" ]; then \
Fix a problem introduced in the previous commit, noticed by Matthiases Scheler and Drochner. Handle missing checksums in distinfo properly. Tested by myself and tron.
2005-02-25 14:05:52 +01:00
case "$$missing" in \
"") ;; \
*) ${ECHO_MSG} "Missing checksums for $$missing";; \
esac; \
Deprecate DIGEST_FILE and PATCH_SUM_FILE (replaced by DIGEST_FILE, and not used anymore in pkgsrc).
2001-06-09 14:15:59 +02:00
${ECHO_MSG} "Make sure the Makefile and checksum file (${DISTINFO_FILE})"; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${ECHO_MSG} "are up to date. If you want to override this check, type"; \
make -> ${MAKE}, as per FreeBSD's PR 6341.
1998-04-19 14:48:07 +02:00
${ECHO_MSG} "\"${MAKE} NO_CHECKSUM=yes [other args]\"."; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
exit 1; \
fi) ; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
fi
.endif
Add bin-install package: will install a binary package if available, else run a "make package". XXX needs to be extended to also look on some FTP sites.
2000-07-06 18:45:21 +02:00
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
# List of sites carrying binary pkgs. Variables "rel" and "arch" are
# replaced with OS release ("1.5", ...) and architecture ("mipsel", ...)
rename BINPKG_SITE to BINPKG_SITES Adresses PR pkg/19004 by tv@pobox.com
2002-11-10 19:20:44 +01:00
BINPKG_SITES?= \
s/netbsd.org/NetBSD.org/, s/freebsd.org/FreeBSD.org/
2004-02-25 10:24:52 +01:00
ftp://ftp.NetBSD.org/pub/NetBSD/packages/$${rel}/$${arch}
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
# List of flags to pass to pkg_add(1) for bin-install:
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
BIN_INSTALL_FLAGS?= # -v
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if ${PKG_INSTALLATION_TYPE} == "pkgviews"
We want to add consistently add to the default view if we use pkgviews and pkg_add'ing.
2003-09-08 03:06:01 +02:00
PKG_ARGS_ADD= -W ${LOCALBASE} -w ${DEFAULT_VIEW}
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.endif
_BIN_INSTALL_FLAGS= ${BIN_INSTALL_FLAGS}
_BIN_INSTALL_FLAGS+= ${PKG_ARGS_ADD}
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
Add bin-install package: will install a binary package if available, else run a "make package". XXX needs to be extended to also look on some FTP sites.
2000-07-06 18:45:21 +02:00
# Install binary pkg, without strict uptodate-check first
Fix PR 17883 by adding code to perform just-in-time su(1) functionality for the bin-install target.
2003-10-18 14:29:26 +02:00
.PHONY: real-su-bin-install
real-su-bin-install:
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
@found="`${PKG_BEST_EXISTS} \"${PKGWILDCARD}\" || ${TRUE}`"; \
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
if [ "$$found" != "" ]; then \
trim whitespace from a message.
2003-12-30 06:07:41 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> $$found is already installed - perhaps an older version?"; \
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
${ECHO_MSG} "*** If so, you may wish to \`\`pkg_delete $$found'' and install"; \
${ECHO_MSG} "*** this package again by \`\`${MAKE} bin-install'' to upgrade it properly."; \
${SHCOMMENT} ${ECHO_MSG} "*** or use \`\`${MAKE} bin-update'' to upgrade it and all of its dependencies."; \
exit 1; \
fi
Add bin-install package: will install a binary package if available, else run a "make package". XXX needs to be extended to also look on some FTP sites.
2000-07-06 18:45:21 +02:00
@if [ -f ${PKGFILE} ] ; then \
${ECHO_MSG} "Installing from binary pkg ${PKGFILE}" ; \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
${PKG_ADD} ${_BIN_INSTALL_FLAGS} ${PKGFILE} ; \
Add bin-install package: will install a binary package if available, else run a "make package". XXX needs to be extended to also look on some FTP sites.
2000-07-06 18:45:21 +02:00
else \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
rel=`${UNAME} -r | ${SED} 's@\.\([0-9]*\)[\._].*@\.\1@'`; \
bin-install: use $MACHINE_ARCH over calling uname
2002-02-14 00:02:14 +01:00
arch=${MACHINE_ARCH}; \
rename BINPKG_SITE to BINPKG_SITES Adresses PR pkg/19004 by tv@pobox.com
2002-11-10 19:20:44 +01:00
for site in ${BINPKG_SITES} ; do \
Migrate some OS-specific variable definitions into OS-specific files. Include it in bsd.prefs.mk to allow their use much earlier in Makefiles, and try to use the variable references instead of directly invoking the programs in bsd.prefs.mk and bsd.pkg.mk where possible.
2001-06-12 14:49:55 +02:00
${ECHO} Trying `eval ${ECHO} $$site`/All ; \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
${SHCOMMENT} ${ECHO} ${SETENV} PKG_PATH="`eval ${ECHO} $$site`/All" ${PKG_ADD} ${_BIN_INSTALL_FLAGS} ${PKGNAME}${PKG_SUFX} ; \
Migrate some OS-specific variable definitions into OS-specific files. Include it in bsd.prefs.mk to allow their use much earlier in Makefiles, and try to use the variable references instead of directly invoking the programs in bsd.prefs.mk and bsd.pkg.mk where possible.
2001-06-12 14:49:55 +02:00
if ${SETENV} PKG_PATH="`eval ${ECHO} $$site`/All" ${PKG_ADD} ${BIN_INSTALL_FLAGS} ${PKGNAME}${PKG_SUFX} ; then \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
${ECHO} "${PKGNAME} successfully installed."; \
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
break ; \
fi ; \
done ; \
quote args to pkg_info, to prevent some shell trying to interpret them.
2002-09-16 12:04:39 +02:00
if ! ${PKG_INFO} -qe "${PKGNAME}" ; then \
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
${SHCOMMENT} Cycle through some FTP server here ;\
${ECHO_MSG} "Installing from source" ; \
bin-install: set DEPENDS_TARGET=bin-install, so depends are still installed via binaries (if possible), even if there is no binary pkg available for the current pkg. Implementation still allows overriding DEPENDS_TARGET=build (etc.) to force installation of dependencies by compiling if desired. Patch submitted by Marko Schuetz <MarkoSchuetz@web.de> in private mail.
2002-04-25 23:58:55 +02:00
${MAKE} ${MAKEFLAGS} package \
Support DEPENDS_TARGET="install clean"; from Allan Barrett in PR pkg/25059.
2004-04-05 10:06:07 +02:00
DEPENDS_TARGET=${DEPENDS_TARGET:Q} && \
make the bin-install target look at some FTP servers (stored in BINPKG_SITES). As discussed on tech-pkg.
2001-05-20 03:58:19 +02:00
${MAKE} ${MAKEFLAGS} clean ; \
fi ; \
Add bin-install package: will install a binary package if available, else run a "make package". XXX needs to be extended to also look on some FTP sites.
2000-07-06 18:45:21 +02:00
fi
Fix PR 17883 by adding code to perform just-in-time su(1) functionality for the bin-install target.
2003-10-18 14:29:26 +02:00
.PHONY: bin-install
bin-install:
@${ECHO_MSG} "${_PKGSRC_IN}> Binary install for ${PKGNAME}"
${_PKG_SILENT}${_PKG_DEBUG} \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
realtarget="real-su-bin-install"; \
Fix PR 17883 by adding code to perform just-in-time su(1) functionality for the bin-install target.
2003-10-18 14:29:26 +02:00
action="binary install"; \
${_SU_TARGET}
Add bin-install package: will install a binary package if available, else run a "make package". XXX needs to be extended to also look on some FTP sites.
2000-07-06 18:45:21 +02:00
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
################################################################
# The special package-building targets
# You probably won't need to touch these
################################################################
Add support for SVR4_PKGNAME: SVR4 has a lenght limitation of the package name: Only 9 characters are allowed. Thats not a real problem since gensolpkg, which is used to create a SVR4 packages, truncates the PKGNAME to 9 characters, but there is a second problem. Normaly you have a vendor identifier in that package name. gensolpkg uses at the moment TNF, so we only have 6 characters left, and that is insufficient for a few packages like the amanda ones. Where the real lenght should be limited to 5 characters so one can choose to use a vendor string up to 4 characters. SVR4_PKGNAME should be only added to a few packages where the truncation of the PKGNAME does not produce an unique package name, e.g the amanda packages. You will need pkgsrc/pkgtools/gensolpkg 1.9 which will be commited in the next few days to use SVR4_PKGNAME. TODO: add SVR4_PKGNAME support to pkglint.
2001-01-21 23:41:03 +01:00
# Set to "html" by the README.html target to generate HTML code,
# or to "svr4" to print SVR4 (Solaris, ...) short package names, from
# SVR4_PKGNAME variable.
# This variable is passed down via build-depends-list and run-depends-list
Cosmetic mods: PACKAGE_NAME_AS_LINK, which can have a YES or NO value, changed to PACKAGE_NAME_TYPE, which can have an html or name value. s/echo/${ECHO}/
1998-08-27 16:56:02 +02:00
PACKAGE_NAME_TYPE?= name
Enhance look & feel of the generated README.html's.
1998-02-20 23:06:18 +01:00
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Nobody should want to override this unless PKGNAME is simply bogus.
add definition for HTML_PKGNAME used by the readme generation
2002-11-12 14:42:50 +01:00
HTML_PKGNAME=<a href="../../${PKGPATH:S/&/\&amp;/g:S/>/\&gt;/g:S/</\&lt;/g}/README.html">${PKGNAME:S/&/\&amp;/g:S/>/\&gt;/g:S/</\&lt;/g}</A>
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: package-name
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(package-name)
package-name:
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if (${PACKAGE_NAME_TYPE} == "html")
Inline HTMLIFY using make(1) subsitution rules. Eliminates two more sed(1) invocations per dependency. XXX This is really quite stupid, in that the characters we're quoting are also shell metacharacters that are passed unquoted to the shell in other places. CVS ----------------------------------------------------------------------
2000-06-04 00:59:04 +02:00
@${ECHO} '<a href="../../${PKGPATH:S/&/\&amp;/g:S/>/\&gt;/g:S/</\&lt;/g}/README.html">${PKGNAME:S/&/\&amp;/g:S/>/\&gt;/g:S/</\&lt;/g}</A>'
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. elif (${PACKAGE_NAME_TYPE} == "svr4")
Add support for SVR4_PKGNAME: SVR4 has a lenght limitation of the package name: Only 9 characters are allowed. Thats not a real problem since gensolpkg, which is used to create a SVR4 packages, truncates the PKGNAME to 9 characters, but there is a second problem. Normaly you have a vendor identifier in that package name. gensolpkg uses at the moment TNF, so we only have 6 characters left, and that is insufficient for a few packages like the amanda ones. Where the real lenght should be limited to 5 characters so one can choose to use a vendor string up to 4 characters. SVR4_PKGNAME should be only added to a few packages where the truncation of the PKGNAME does not produce an unique package name, e.g the amanda packages. You will need pkgsrc/pkgtools/gensolpkg 1.9 which will be commited in the next few days to use SVR4_PKGNAME. TODO: add SVR4_PKGNAME support to pkglint.
2001-01-21 23:41:03 +01:00
@${ECHO} ${SVR4_PKGNAME}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else
Get rid of another fork.
2000-06-03 15:36:13 +02:00
@${ECHO} ${PKGNAME}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif # PACKAGE_NAME_TYPE
Enhance look & feel of the generated README.html's.
1998-02-20 23:06:18 +01:00
.endif # !target(package-name)
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: make-readme-html-help
Change COMMENT handling: COMMENTs are now a variable in the Makefile instead of a pkg/COMMENT file. The COMMENT var should be in the maintainer block after the homepage. Modify bsd.pkg.mk, pkglint, url2pkg, and port2pkg (last one untested) for the new behaviour. Document new state in Packages.txt. This should save lots of inodes, and lots of time when untarring/updating. Idea by Alistair Crooks. For the time being, accept pkg/COMMENT instead of a COMMENT var to avoid a flag day.
2001-02-16 14:06:17 +01:00
.if !target(make-readme-html-help)
make-readme-html-help:
Use the make :Q modifier for quoting strings for the shell, instead of manually doing it (in some places).
2001-02-21 16:01:29 +01:00
@${ECHO} '${PKGNAME:S/&/\&amp;/g:S/>/\&gt;/g:S/</\&lt;/g}</a>: <TD>'${COMMENT:S/&/\&amp;/g:S/>/\&gt;/g:S/</\&lt;/g:Q}
Change COMMENT handling: COMMENTs are now a variable in the Makefile instead of a pkg/COMMENT file. The COMMENT var should be in the maintainer block after the homepage. Modify bsd.pkg.mk, pkglint, url2pkg, and port2pkg (last one untested) for the new behaviour. Document new state in Packages.txt. This should save lots of inodes, and lots of time when untarring/updating. Idea by Alistair Crooks. For the time being, accept pkg/COMMENT instead of a COMMENT var to avoid a flag day.
2001-02-16 14:06:17 +01:00
.endif # !target(make-readme-html-help)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Show (recursively) all the packages this package depends on.
Change print-depends to look into /var/db/pkg, instead of pkgsrc (implements $PACKAGE_DEPENDS_QUICK) -- Thomas Klausner hinted me at some interresting problem WRT the print-depends package. If something depends on suse_linus-6.*, that package will register whatever suse_linux there is right now on the system (in his case V6.1). What print-depends then does is also dig out any pkgs required for suse_linux, based on what's in pkgsrc, which is of course _wrong_, as this most likely doesn't match the installed version.[*] The right thing instead is to print whatever the (currently installed) suse_linux package depends on, which can be read from "pkg_info -qf" output, looking at the @pkgdep lines. Of course this only works if the pkg is actually installed, so it cannot be used for README.html generation etc., but that's bad luck. One positive aspect of using the @pkgdep list instead of going down recursively is that this saves some time and processes. On my P133 it's going down from ~11s to 0.5s for the pkgsrc/x11/kde package. During a "make package" of kde itself (not counting the depending pkgs) this will cut down like 20s (once for the fake-pkg target, and another one for the pkg_create args - maybe a 3rd time for pkg_create in fake-pkg, but I don't want to check that right now). Anyways, a general solution would be to use the "pkg_info -qf | grep @pkgdep" in print-depends instead of recursion (only!) in the places that we know that the depends are installed. Which is the PKG_ARGS assignment (for pkg_create) and the fake-pkg target. [*] Plus it may be questioned that we should not go down recursively for pkg depends anyways, but this is used in too many subtle ways throughout the system to change.
2000-03-09 15:05:26 +01:00
# If PACKAGE_DEPENDS_WITH_PATTERNS is set, print as pattern (if possible)
Bump PKGTOOLS_REQD to 20000202. As a consequence of this - unconditionally enable the size code - use wildcards consistently (PACKAGE_DEPENDS_WITH_PATTERNS=true)
2000-02-02 17:28:55 +01:00
PACKAGE_DEPENDS_WITH_PATTERNS?=true
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: run-depends-list
* More target renaming: depends-list => build-depends-list package-depends => run-depends-list print-depends-list => print-build-depends-list print-package-depends => print-run-depends-list * While there, fix print-run-depends-list to not seperate items by \n. Changes discussed with wiz.
2000-08-01 04:51:34 +02:00
.if !target(run-depends-list)
run-depends-list:
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. for dep in ${DEPENDS}
When displaying a list of "dependencies" in the run-depends-list target, don't display what we're doing if PKG_DEBUG_LEVEL is non-zero - the output of run-depends-list is fed to pkg_create(1) and encoded in the binary package as a list of @pkgdep lines, and expanding the output is not the right thing to do.
2001-05-24 15:28:36 +02:00
@pkg="${dep:C/:.*//}"; \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
dir="${dep:C/[^:]*://}"; \
Allow using absolute paths as second component of DEPENDS again.
2000-06-04 02:50:48 +02:00
cd ${.CURDIR}; \
In package-depends, don't print warnings about non-existing dir if the information from that dir isn't used. Reviewed by: wiz
2000-06-24 01:41:43 +02:00
if ${PACKAGE_DEPENDS_WITH_PATTERNS}; then \
${ECHO} "$$pkg"; \
else \
if cd $$dir 2>/dev/null; then \
Clean up the previous commits: * in package-depends split handling of DEPENDS and RUN_DEPENDS again. The fact that the code was equal for both cases (and thus broken!) is no excuse to fold the two together. (What's the difference? DEPENDS has a pkg name/patter as first value, RUN_DEPENDS has a executable name of file/path) * cut down two more ()s * fix binpkg-list to find binary pkgs again. Remember, $PACKAGES is where binary packages are placed, under a well defined directory structure (including dir "All" and category dirs). * Sort available binary pkgs in README.html by machine again * Inline one more use of the "package-path" target. It's unused now. Anyone care if we remove it? * Fix variable name in print-pkg-depend-sizes: the first component of DEPENDS is a package-name/patter, not a program.
2000-06-03 18:52:32 +02:00
${MAKE} ${MAKEFLAGS} package-name PACKAGE_NAME_TYPE=${PACKAGE_NAME_TYPE}; \
* Move the clean-depends target near the clear target * some whitespace change to enhance readability
2000-05-31 18:59:44 +02:00
else \
In package-depends, don't print warnings about non-existing dir if the information from that dir isn't used. Reviewed by: wiz
2000-06-24 01:41:43 +02:00
${ECHO_MSG} "Warning: \"$$dir\" non-existent -- @pkgdep registration incomplete" >&2; \
fi; \
Clean up the previous commits: * in package-depends split handling of DEPENDS and RUN_DEPENDS again. The fact that the code was equal for both cases (and thus broken!) is no excuse to fold the two together. (What's the difference? DEPENDS has a pkg name/patter as first value, RUN_DEPENDS has a executable name of file/path) * cut down two more ()s * fix binpkg-list to find binary pkgs again. Remember, $PACKAGES is where binary packages are placed, under a well defined directory structure (including dir "All" and category dirs). * Sort available binary pkgs in README.html by machine again * Inline one more use of the "package-path" target. It's unused now. Anyone care if we remove it? * Fix variable name in print-pkg-depend-sizes: the first component of DEPENDS is a package-name/patter, not a program.
2000-06-03 18:52:32 +02:00
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endfor
* More target renaming: depends-list => build-depends-list package-depends => run-depends-list print-depends-list => print-build-depends-list print-package-depends => print-run-depends-list * While there, fix print-run-depends-list to not seperate items by \n. Changes discussed with wiz.
2000-08-01 04:51:34 +02:00
.endif # target(run-depends-list)
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# Build a package but don't check the package cookie
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: repackage
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(repackage)
repackage: pre-repackage package
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: pre-repackage
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
pre-repackage:
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${RM} -f ${PACKAGE_COOKIE}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
# Build a package but don't check the cookie for installation, also don't
# install package cookie
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: package-noinstall
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(package-noinstall)
package-noinstall:
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG}cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} PACKAGE_NOINSTALL=yes real-package
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
################################################################
# Dependency checking
################################################################
Sometimes, the dependencies installed are a function of which variables are set, and it'd be nice to know which ones before the default dependencies are installed. Create a new target "pre-install-depends" that is executed before dependencies are installed. Move the display of BUILD_DEFS (build-defs-message) into pre-install-depends so that they are displayed before dependencies are installed.
2004-08-05 01:22:38 +02:00
.PHONY: install-depends pre-install-depends
pre-install-depends:
forgot one depends -> install-depends rename Hinted at by wiz
2000-08-07 01:39:23 +02:00
.if !target(install-depends)
Introduce another definition, FATAL_OBJECT_FMT_SKEW, defaulting to "yes". This halts the execution if an installed package's object format is not the same as the object format currently in use. It can be set to "no" on the rare occasion where the interface to the pre-requisite packages is one of a binary, not a library.
2000-01-14 12:58:21 +01:00
# Tells whether to halt execution if the object formats differ
FATAL_OBJECT_FMT_SKEW?= yes
Introduce another definition WARN_NO_OBJECT_FMT, defaulting to yes, which governs whether a package which was installed before OBJECT_FMT was added to the BUILD_DEFS will display an warning message. To shut up the message, set WARN_NO_OBJECT_FMT to "no".
2000-01-17 15:49:55 +01:00
WARN_NO_OBJECT_FMT?= yes
Introduce another definition, FATAL_OBJECT_FMT_SKEW, defaulting to "yes". This halts the execution if an installed package's object format is not the same as the object format currently in use. It can be set to "no" on the rare occasion where the interface to the pre-requisite packages is one of a binary, not a library.
2000-01-14 12:58:21 +01:00
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
install-depends: pre-install-depends
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
. if !empty(DEPENDS) || !empty(BUILD_DEPENDS)
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. if defined(NO_DEPENDS)
Change BUILD_DEPENDS semantics: first component is now a package name+version/pattern, no more executable/patchname/whatnot. While there, introduce BUILD_USES_MSGFMT as shorthand to pull in devel/gettext unless /usr/bin/msgfmt exists (i.e. on post-1.5 -current). Patch by Alistair Crooks <agc@netbsd.org>
2001-03-27 05:19:43 +02:00
@${DO_NADA}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. else # !DEPENDS
. for dep in ${DEPENDS} ${BUILD_DEPENDS}
Change BUILD_DEPENDS semantics: first component is now a package name+version/pattern, no more executable/patchname/whatnot. While there, introduce BUILD_USES_MSGFMT as shorthand to pull in devel/gettext unless /usr/bin/msgfmt exists (i.e. on post-1.5 -current). Patch by Alistair Crooks <agc@netbsd.org>
2001-03-27 05:19:43 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Minor consistency issues, eliminate grep-before-awk, and fix a bug that was introduced with BUILD_DEPENDS that have special targets.
2000-06-03 16:17:49 +02:00
pkg="${dep:C/:.*//}"; \
dir="${dep:C/[^:]*://:C/:.*$//}"; \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
found=`${PKG_BEST_EXISTS} "$$pkg" || ${TRUE}`; \
Add support for two environment variables to help update all your installed packages and only rebuild everything once: STOP_DOWNLEVEL_AFTER_FIRST if set makes "make show-downlevel" stop a bit earlier (since with the other one below you will be only interested in the first downlevel pkg found). REBUILD_DOWNLEVEL_DEPENDS if set forces the pattern match for dependencies to fail whenever the installed pkg is not the version in the makefile (i.e. "make show-downlevel" would print a version mismatch for the dependecy). This causes all downlevel dependencies (and everything depending on them) to be rebuild. Enhancements, like making this all work with make command line flags and settings in /etc/mk.conf are welcome. A way to stop "make show-downlevel" through all upper levels of recursion imediately would be very usefull too.
2001-09-10 22:03:17 +02:00
if [ "X$$REBUILD_DOWNLEVEL_DEPENDS" != "X" ]; then \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
pkgname=`cd $$dir ; ${MAKE} ${MAKEFLAGS} show-var VARNAME=PKGNAME`; \
if [ "X$$found" != "X" -a "X$$found" != "X$${pkgname}" ]; then \
${ECHO_MSG} "ignoring old installed package \"$$found\""; \
Add support for two environment variables to help update all your installed packages and only rebuild everything once: STOP_DOWNLEVEL_AFTER_FIRST if set makes "make show-downlevel" stop a bit earlier (since with the other one below you will be only interested in the first downlevel pkg found). REBUILD_DOWNLEVEL_DEPENDS if set forces the pattern match for dependencies to fail whenever the installed pkg is not the version in the makefile (i.e. "make show-downlevel" would print a version mismatch for the dependecy). This causes all downlevel dependencies (and everything depending on them) to be rebuild. Enhancements, like making this all work with make command line flags and settings in /etc/mk.conf are welcome. A way to stop "make show-downlevel" through all upper levels of recursion imediately would be very usefull too.
2001-09-10 22:03:17 +02:00
found=""; \
fi; \
fi; \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
if [ "$$found" != "" ]; then \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
instobjfmt=`${PKG_INFO} -B "$$found" | ${AWK} -F'=[ \t]*' '/^OBJECT_FMT/ {print $$2; exit}'`; \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
if [ "$$instobjfmt" = "" ]; then \
Introduce another definition WARN_NO_OBJECT_FMT, defaulting to yes, which governs whether a package which was installed before OBJECT_FMT was added to the BUILD_DEFS will display an warning message. To shut up the message, set WARN_NO_OBJECT_FMT to "no".
2000-01-17 15:49:55 +01:00
if [ "X${WARN_NO_OBJECT_FMT}" != "Xno" ]; then \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
${ECHO} "WARNING: Unknown object format for installed package $$found - continuing"; \
Introduce another definition WARN_NO_OBJECT_FMT, defaulting to yes, which governs whether a package which was installed before OBJECT_FMT was added to the BUILD_DEFS will display an warning message. To shut up the message, set WARN_NO_OBJECT_FMT to "no".
2000-01-17 15:49:55 +01:00
fi; \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
elif [ "$$instobjfmt" != "${OBJECT_FMT}" ]; then \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
${ECHO} "Installed package $$found is an $$instobjfmt package."; \
Add a check, inspired by Thor Lancelot Simon: When an installed package is found as part of a DEPENDS pre-requisite, check its object format. If there is none in the installed package, that package was probably built before 29th September 1999, which was when the OBJECT_FMT definition was added to the BUILD_DEFS, so print a warning and continue. If the object format of the installed package and ${OBJECT_FMT} are the same, then continue. If the object format of the installed package and ${OBJECT_FMT} differ, then print an error message explaining the error, and exit.
2000-01-14 12:39:31 +01:00
${ECHO} "You are building an ${OBJECT_FMT} package, which will not inter-operate."; \
In cases where we need the best match for a pkgpattern, use "${PKG_BEST_EXIST} pkgpattern" instead of "${PKG_INFO} -e pkgpattern". The latter can return multiple package names if there are multiple versions of a piece of software installed. PKG_BEST_EXIST is defined to be "${PKG_ADMIN} -b -d ${_PKG_DBDIR} -s "" lsbest", so it searches for the best installed package that matches the given pkgpattern or else returns the empty string. Bump PKGTOOLS_REQD to 20030912 since pkg_admin(1) needs to know about "-b" and "-d <dir>".
2003-09-13 07:55:14 +02:00
${ECHO} "Please update the $$found package to ${OBJECT_FMT}"; \
Let "FATAL_OBJECT_FMT_SKEW=no" really work.
2000-08-07 12:23:38 +02:00
if [ "X${FATAL_OBJECT_FMT_SKEW}" != "Xno" ]; then \
Introduce another definition, FATAL_OBJECT_FMT_SKEW, defaulting to "yes". This halts the execution if an installed package's object format is not the same as the object format currently in use. It can be set to "no" on the rare occasion where the interface to the pre-requisite packages is one of a binary, not a library.
2000-01-14 12:58:21 +01:00
exit 1; \
fi; \
Add a check, inspired by Thor Lancelot Simon: When an installed package is found as part of a DEPENDS pre-requisite, check its object format. If there is none in the installed package, that package was probably built before 29th September 1999, which was when the OBJECT_FMT definition was added to the BUILD_DEFS, so print a warning and continue. If the object format of the installed package and ${OBJECT_FMT} are the same, then continue. If the object format of the installed package and ${OBJECT_FMT} differ, then print an error message explaining the error, and exit.
2000-01-14 12:39:31 +01:00
fi; \
We just depend on the package that best matches the pkgpattern, so remove the unreachable code surrounded by checks for matching multiple packages.
2003-09-13 13:32:04 +02:00
${ECHO_MSG} "${_PKGSRC_IN}> Required installed package $$pkg: $${found} found"; \
+ Remove the generated ${PLIST} file when doing a re-install. + Enable wildcard matching in DEPENDS and CONFLICTS definitions.
1998-08-25 18:26:09 +02:00
else \
Change the output of the 'install-depends' target so it fits in 80 columns: before: ===> mozilla-0.m18nb1 depends on installed package: gtk+>=1.2.8 - gtk+-1.2.8 found ===> mozilla-0.m18nb1 depends on installed package: ORBit>=0.5.1 - ORBit-0.5.3nb1 found ===> mozilla-0.m18nb1 depends on installed package: jpeg-6b - jpeg-6b found ===> mozilla-0.m18nb1 depends on installed package: png>=1.0.6 - png-1.0.8 found ===> mozilla-0.m18nb1 depends on file: /usr/pkg/bin/perl - found ===> mozilla-0.m18nb1 depends on executable: autoconf - /usr/pkg/bin/autoconf found ===> mozilla-0.m18nb1 depends on executable: zip - /usr/pkg/bin/zip found ===> mozilla-0.m18nb1 depends on executable: gmake - /usr/pkg/bin/gmake found after: ===> Required installed package gtk+>=1.2.8: gtk+-1.2.8 found ===> Required installed package ORBit>=0.5.1: ORBit-0.5.3nb1 found ===> Required installed package jpeg-6b: jpeg-6b found ===> Required installed package png>=1.0.6: png-1.0.8 found ===> Required file /usr/pkg/bin/perl: found ===> Required executable autoconf: /usr/pkg/bin/autoconf found ===> Required executable zip: /usr/pkg/bin/zip found ===> Required executable gmake: /usr/pkg/bin/gmake found
2000-11-16 06:36:01 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Required package $$pkg: NOT found"; \
Support DEPENDS_TARGET="install clean"; from Allan Barrett in PR pkg/25059.
2004-04-05 10:06:07 +02:00
target=${DEPENDS_TARGET:Q}; \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Verifying $$target for $$dir"; \
+ Remove the generated ${PLIST} file when doing a re-install. + Enable wildcard matching in DEPENDS and CONFLICTS definitions.
1998-08-25 18:26:09 +02:00
if [ ! -d $$dir ]; then \
In informational messages to the user, use '=>' in preference to '>>', so that cut-n-paste into send-pr will work correctly. Fixes PR 9173 from ITOH Yasufumi.
2000-03-10 17:07:37 +01:00
${ECHO_MSG} "=> No directory for $$dir. Skipping.."; \
Remove the old deprecated DEPENDS definition from bsd.pkg.mk. Implement a new DEPENDS definition, which looks for an installed package, building it if not present, and use it in preference to LIB_DEPENDS. This should make the package collection more useful on NetBSD ELF ports.
1998-04-22 16:22:39 +02:00
else \
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
cd $$dir ; \
Prevent MAKEFLAGS from being passed through shell environment. It was redundant because we explicitly pass MAKEFLAGS to a sub-process through command line arguments. This fixes the problem reported in PR pkg/29874 that PKGNAME_REQD is not overridden in recursive MAKE calls.
2005-04-05 05:18:00 +02:00
${SETENV} _PKGSRC_DEPS=", ${PKGNAME}${_PKGSRC_DEPS}" ${MAKE} ${MAKEFLAGS} $$target PKGNAME_REQD=\'$$pkg\' MAKEFLAGS="" || exit 1; \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Returning to build of ${PKGNAME}"; \
- The comment above the _COOKIE settings says "don't change these!!", so use = to set them, not ?=. - _DEPENDS_USE doesn't actually do any macro expansion, so save some processing time by executing it exactly once (remove the .USE operator). - Merge the IGNORE targets. Add an IGNORE target for depends and check-depends (see bottom). - Don't check for installation of a package in ${DEPENDS} when doing a "make depends". Instead, let the return code of make fall through (using && instead of ; in the last clause of each depend recursion block). - Provide a variable IGNORE_FAIL: when set, causes the ignore message to be printed (regardless of IGNORE_SILENT or ECHO_MSG setting) and the build to fail with ${FALSE}. Used with the next addition: - Add a new target, "check-depends", which is called as part of "real-fetch". This validates whether the dependencies of a pkg to be built are installed, and if not, whether those dependencies will prospectively build (will not set IGNORE directly or indirectly through the various disable options). Recurses to "depends" with a DEPENDS_TARGET of "check-depends". ...And now, it is no longer necessary to mark dependent pkgs as unusable when their dependencies are unusable thanks to an incorrect LICENSE, _FOR_PLATFORM setting, and so forth. The check-depends step will validate everything before fetching a single file.
1999-05-25 01:04:52 +02:00
fi; \
fi
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endfor # DEPENDS
. endif # !NO_DEPENDS
. endif # DEPENDS
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: build-depends-list
* More target renaming: depends-list => build-depends-list package-depends => run-depends-list print-depends-list => print-build-depends-list print-package-depends => print-run-depends-list * While there, fix print-run-depends-list to not seperate items by \n. Changes discussed with wiz.
2000-08-01 04:51:34 +02:00
.if !target(build-depends-list)
build-depends-list:
Use show-all-depends-dirs-excl for build-depends-list target. Makes e.g. "make readme" awfully faster. TODO: Similar change for run-depends-list.
2002-03-02 17:09:14 +01:00
@for dir in `${MAKE} ${MAKEFLAGS} show-all-depends-dirs-excl`; \
do \
(cd ../../$$dir && \
${MAKE} ${MAKEFLAGS} package-name) \
done
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Add a cdrom-readme target, a clone of the readme target, for ease of use. The URLs in the generated README.html files can be specified by overriding the CDROM_PKG_URL_HOST and CDROM_PKG_URL_DIR definitions. Document the targets, and clean up some English, in Packages.txt
1998-09-23 15:09:32 +02:00
# If PACKAGES is set to the default (../../pkgsrc/packages), the current
# ${MACHINE_ARCH} and "release" (uname -r) will be used. Otherwise a directory
# structure of ...pkgsrc/packages/`uname -r`/${MACHINE_ARCH} is assumed.
# The PKG_URL is set from FTP_PKG_URL_* or CDROM_PKG_URL_*, depending on
# the target used to generate the README.html file.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: binpkg-list
Add "click here to download"-links.
1998-05-25 02:04:30 +02:00
.if !target(binpkg-list)
binpkg-list:
replace use of ${TEST} -e with -f or -d as appropriate. use consistent shell syntax. this allows the pkgsrc framework to work with IRIX 5.3 and Solaris' sh(1) built-in test(1).
2004-07-03 23:59:04 +02:00
@if ${TEST} -d ${PACKAGES}; then \
Check that the ${PACKAGES} directory exists, before trying to cd there. This modification allows "make readme" to succeed even if the ${PACKAGES} hierarchy does not exist. Fixes PR 12480 from Arto Selonen (arto@selonen.org), and addresses PR 12362 from Masao Uebayashi <uebayasi@soum.co.jp>
2001-05-03 15:01:04 +02:00
cd ${PACKAGES}; \
case ${.CURDIR} in \
*/pkgsrc/packages) \
for pkg in ${PKGREPOSITORYSUBDIR}/${PKGWILDCARD}${PKG_SUFX} ; \
do \
if [ -f "$$pkg" ] ; then \
pkgname=`${ECHO} $$pkg | ${SED} 's@.*/@@'`; \
${ECHO} "<TR><TD>${MACHINE_ARCH}:<TD><a href=\"${PKG_URL}/$$pkg\">$$pkgname</a><TD>(${OPSYS} ${OS_VERSION})"; \
fi ; \
done ; \
;; \
*) \
cd ${PACKAGES}/../..; \
for i in [1-9].*/*; do \
if cd ${PACKAGES}/../../$$i/${PKGREPOSITORYSUBDIR} 2>/dev/null; then \
for j in ${PKGWILDCARD}${PKG_SUFX}; \
do \
if [ -f "$$j" ]; then \
${ECHO} $$i/$$j;\
fi; \
done; \
fi; \
done | ${AWK} -F/ ' \
{ \
release = $$1; \
arch = $$2; \
pkg = $$3; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
gsub("\\.tgz","", pkg); \
Check that the ${PACKAGES} directory exists, before trying to cd there. This modification allows "make readme" to succeed even if the ${PACKAGES} hierarchy does not exist. Fixes PR 12480 from Arto Selonen (arto@selonen.org), and addresses PR 12362 from Masao Uebayashi <uebayasi@soum.co.jp>
2001-05-03 15:01:04 +02:00
if (arch != "m68k" && arch != "mipsel") { \
if (arch in urls) \
urls[arch "/" pkg "/" release] = "<a href=\"${PKG_URL}/" release "/" arch "/${PKGREPOSITORYSUBDIR}/" pkg "${PKG_SUFX}\">" pkg "</a>, " urls[arch]; \
else \
urls[arch "/" pkg "/" release] = "<a href=\"${PKG_URL}/" release "/" arch "/${PKGREPOSITORYSUBDIR}/" pkg "${PKG_SUFX}\">" pkg "</a> "; \
} \
} \
END { \
for (av in urls) { \
split(av, ava, "/"); \
arch=ava[1]; \
pkg=ava[2]; \
release=ava[3]; \
print "<TR><TD>" arch ":<TD>" urls[av] "<TD>(${OPSYS} " release ")"; \
} \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
}' | ${SORT} \
Check that the ${PACKAGES} directory exists, before trying to cd there. This modification allows "make readme" to succeed even if the ${PACKAGES} hierarchy does not exist. Fixes PR 12480 from Arto Selonen (arto@selonen.org), and addresses PR 12362 from Masao Uebayashi <uebayasi@soum.co.jp>
2001-05-03 15:01:04 +02:00
;; \
esac; \
fi
Add "click here to download"-links.
1998-05-25 02:04:30 +02:00
.endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
################################################################
# Everything after here are internal targets and really
# shouldn't be touched by anybody but the release engineers.
################################################################
# This target generates an index entry suitable for aggregation into
# a large index. Format is:
#
This is NetBSD. s/port/package/g
1999-08-10 12:48:23 +02:00
# distribution-name|package-path|installation-prefix|comment| \
- fix comment about what shows up in the index file. - add ONLY_FOR_ARCHS and NOT_FOR_OPSYS to the list of variables that goes in the print-summary-data output. This is the first step in speeding up the 'make index' process.
2003-07-19 04:50:22 +02:00
# description-file|maintainer|categories|build deps|run deps|for arch| \
# not for opsys
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
#
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: describe
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(describe)
describe:
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
@${ECHO_N} "${PKGNAME}|${.CURDIR}|"; \
${ECHO_N} "${PREFIX}|"; \
${ECHO_N} ${COMMENT:Q}; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
if [ -f ${DESCR_SRC} ]; then \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "|${DESCR_SRC}"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
else \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "|/dev/null"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
fi; \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "|${MAINTAINER}|${CATEGORIES}|"; \
- The comment above the _COOKIE settings says "don't change these!!", so use = to set them, not ?=. - _DEPENDS_USE doesn't actually do any macro expansion, so save some processing time by executing it exactly once (remove the .USE operator). - Merge the IGNORE targets. Add an IGNORE target for depends and check-depends (see bottom). - Don't check for installation of a package in ${DEPENDS} when doing a "make depends". Instead, let the return code of make fall through (using && instead of ; in the last clause of each depend recursion block). - Provide a variable IGNORE_FAIL: when set, causes the ignore message to be printed (regardless of IGNORE_SILENT or ECHO_MSG setting) and the build to fail with ${FALSE}. Used with the next addition: - Add a new target, "check-depends", which is called as part of "real-fetch". This validates whether the dependencies of a pkg to be built are installed, and if not, whether those dependencies will prospectively build (will not set IGNORE directly or indirectly through the various disable options). Recurses to "depends" with a DEPENDS_TARGET of "check-depends". ...And now, it is no longer necessary to mark dependent pkgs as unusable when their dependencies are unusable thanks to an incorrect LICENSE, _FOR_PLATFORM setting, and so forth. The check-depends step will validate everything before fetching a single file.
1999-05-25 01:04:52 +02:00
case "A${BUILD_DEPENDS}B${DEPENDS}C" in \
ABC) ;; \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
*) cd ${.CURDIR} && ${ECHO_N} `${MAKE} ${MAKEFLAGS} build-depends-list | ${SORT} -u`;; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
esac; \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "|"; \
if [ "${DEPENDS}" != "" ]; then \
cd ${.CURDIR} && ${ECHO_N} `${MAKE} ${MAKEFLAGS} run-depends-list | ${SORT} -u`; \
* Remove RUN_DEPENDS (left in packages.7 intentionally) * Renamed misc-depends to run-depends, and drop (old) run-depends altogether Code reviewed by wiz.
2000-08-01 04:16:58 +02:00
fi; \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "|"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
if [ "${ONLY_FOR_ARCHS}" = "" ]; then \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "any"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
else \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "${ONLY_FOR_ARCHS}"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
fi; \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "|"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
if [ "${NOT_FOR_OPSYS}" = "" ]; then \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "any"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
else \
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
${ECHO_N} "not ${NOT_FOR_OPSYS}"; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
fi; \
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
${ECHO} ""
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: readmes
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(readmes)
readmes: readme
.endif
Add a cdrom-readme target, a clone of the readme target, for ease of use. The URLs in the generated README.html files can be specified by overriding the CDROM_PKG_URL_HOST and CDROM_PKG_URL_DIR definitions. Document the targets, and clean up some English, in Packages.txt
1998-09-23 15:09:32 +02:00
# This target is used to generate README.html files
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: readme
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(readme)
s/netbsd.org/NetBSD.org/, s/freebsd.org/FreeBSD.org/
2004-02-25 10:24:52 +01:00
FTP_PKG_URL_HOST?= ftp://ftp.NetBSD.org
Add a cdrom-readme target, a clone of the readme target, for ease of use. The URLs in the generated README.html files can be specified by overriding the CDROM_PKG_URL_HOST and CDROM_PKG_URL_DIR definitions. Document the targets, and clean up some English, in Packages.txt
1998-09-23 15:09:32 +02:00
FTP_PKG_URL_DIR?= /pub/NetBSD/packages
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
readme:
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
@cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} README.html PKG_URL=${FTP_PKG_URL_HOST}${FTP_PKG_URL_DIR}
Add a cdrom-readme target, a clone of the readme target, for ease of use. The URLs in the generated README.html files can be specified by overriding the CDROM_PKG_URL_HOST and CDROM_PKG_URL_DIR definitions. Document the targets, and clean up some English, in Packages.txt
1998-09-23 15:09:32 +02:00
.endif
# This target is used to generate README.html files, very like "readme"
# However, a different target was used for ease of use.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: cdrom-readme
Add a cdrom-readme target, a clone of the readme target, for ease of use. The URLs in the generated README.html files can be specified by overriding the CDROM_PKG_URL_HOST and CDROM_PKG_URL_DIR definitions. Document the targets, and clean up some English, in Packages.txt
1998-09-23 15:09:32 +02:00
.if !target(cdrom-readme)
CDROM_PKG_URL_HOST?= file://localhost
CDROM_PKG_URL_DIR?= /usr/pkgsrc/packages
cdrom-readme:
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
@cd ${.CURDIR} && ${MAKE} ${MAKEFLAGS} README.html PKG_URL=${CDROM_PKG_URL_HOST}${CDROM_PKG_URL_DIR}
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
README_NAME= ${TEMPLATES}/README.pkg
Remove the magic control characters in a similar way to Bill Fenner's patch for bsd.port.subdir.mk.
1997-11-21 18:44:11 +01:00
Add any licensing information to the README.html files which are generated for each package.
1998-07-22 11:18:46 +02:00
# set up the correct license information as a sed expression
.ifdef LICENSE
SED_LICENSE_EXPR= -e 's|%%LICENSE%%|<p>Please note that this package has a ${LICENSE} license.</p>|'
.else
SED_LICENSE_EXPR= -e 's|%%LICENSE%%||'
.endif
The Grand Homepagification: - New, optional Makefile variable HOMEPAGE, specifies a URL for the home page of the software if it has one. - The value of HOMEPAGE is used to add a link from the README.html files. - pkglint updated to know about it. The "correct" location for HOMEPAGE in the Makefile is after MAINTAINER, in that same section.
1998-08-20 17:16:34 +02:00
# set up the "more info URL" information as a sed expression
.ifdef HOMEPAGE
Change "This package has a homepage" to "This package has a homepage at <link>." Suggested by David Brownlee.
1998-08-27 13:34:23 +02:00
SED_HOMEPAGE_EXPR= -e 's|%%HOMEPAGE%%|<p>This package has a home page at <a HREF="${HOMEPAGE}">${HOMEPAGE}</a>.</p>|'
The Grand Homepagification: - New, optional Makefile variable HOMEPAGE, specifies a URL for the home page of the software if it has one. - The value of HOMEPAGE is used to add a link from the README.html files. - pkglint updated to know about it. The "correct" location for HOMEPAGE in the Makefile is after MAINTAINER, in that same section.
1998-08-20 17:16:34 +02:00
.else
SED_HOMEPAGE_EXPR= -e 's|%%HOMEPAGE%%||'
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-vulnerabilities-html
Fix the show-vulnerabilities and show-vulnerabilities-html targets to work properly in the absence of a vulnerability list.
2000-09-28 12:50:43 +02:00
show-vulnerabilities-html:
Fix PR 15395, whereby a package entry in the vulnerabilities file, which shared a common prefix with a package being installed, would erroneously be flagged as having a vulnerability. (e.g. samba and sam).
2002-01-28 21:55:26 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
if [ -f ${PKGVULNDIR}/pkg-vulnerabilities ]; then \
In the show-vulnerabilities and show-vulnerabilities-html targets, tighten down further the regexp used to match a vulnerable package.
2002-01-29 16:07:55 +01:00
${AWK} '/^${PKGBASE}[-<>=]+[0-9]/ { gsub("\<", "\\&lt;", $$1); \
Fix the show-vulnerabilities and show-vulnerabilities-html targets to work properly in the absence of a vulnerability list.
2000-09-28 12:50:43 +02:00
gsub("\>", "\\&gt;", $$1); \
Open/close HTML mark-up in the correct order. Noted by Kevin P. Neal.
2003-02-09 15:59:12 +01:00
printf("<LI><STRONG>%s has a %s exploit (see <a href=\"%s\">%s</a> for more details)</STRONG></LI>\n", $$1, $$2, $$3, $$3) }' \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
${PKGVULNDIR}/pkg-vulnerabilities; \
Fix the show-vulnerabilities and show-vulnerabilities-html targets to work properly in the absence of a vulnerability list.
2000-09-28 12:50:43 +02:00
fi
Add the glue for showing any security vulnerabilities to the README.html files.
2000-09-19 21:29:11 +02:00
redo the README.html target for increased speed. For packages with no dependencies the speedup is about 2x for ones like gnome with lots of dependencies, the speedup is aroud 400x. The speedup is obtained by generating the dependency list in a more efficient way that only takes 1 make call per dependency instead of potentially many many make calls.
2003-01-04 23:20:47 +01:00
# If PACKAGES is set to the default (../../packages), the current
# ${MACHINE_ARCH} and "release" (uname -r) will be used. Otherwise a directory
# structure of ...pkgsrc/packages/`uname -r`/${MACHINE_ARCH} is assumed.
# The PKG_URL is set from FTP_PKG_URL_* or CDROM_PKG_URL_*, depending on
# the target used to generate the README.html file.
cleanup of README.html generation code, as provided in PR 7989 by Jaromir Dolecek <dolecek@ics.muni.cz>. While there, mark the README.html as precious, so it doesn't get removed if the build process is interrupted.
1999-07-14 21:34:06 +02:00
.PHONY: README.html
README.html: .PRECIOUS
s/${ENV}/${SETENV}/, noted by Kevin P. Neal in connection with PR 19586.
2003-02-09 15:38:51 +01:00
@${SETENV} AWK=${AWK} BMAKE=${MAKE} ../../mk/scripts/mkdatabase -f $@.tmp1
replace use of ${TEST} -e with -f or -d as appropriate. use consistent shell syntax. this allows the pkgsrc framework to work with IRIX 5.3 and Solaris' sh(1) built-in test(1).
2004-07-03 23:59:04 +02:00
@if ${TEST} -d ${PACKAGES}; then \
redo the README.html target for increased speed. For packages with no dependencies the speedup is about 2x for ones like gnome with lots of dependencies, the speedup is aroud 400x. The speedup is obtained by generating the dependency list in a more efficient way that only takes 1 make call per dependency instead of potentially many many make calls.
2003-01-04 23:20:47 +01:00
cd ${PACKAGES}; \
Make BUILD_DIR always point to a physical path by invoking /bin/pwd instead of relying on the shell's builtin pwd. This makes BUILD_DIR consistently point to the correct directory regardless of the definition of SHELL. This fixes elusive some buildlink2 errors due to the fact that BUILDLINK_DIR is derived from BUILD_DIR and some paths were incorrectly being translated.
2003-01-15 21:55:38 +01:00
case `${PWD_CMD}` in \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
${PKGSRCDIR}/packages) \
redo the README.html target for increased speed. For packages with no dependencies the speedup is about 2x for ones like gnome with lots of dependencies, the speedup is aroud 400x. The speedup is obtained by generating the dependency list in a more efficient way that only takes 1 make call per dependency instead of potentially many many make calls.
2003-01-04 23:20:47 +01:00
MULTIARCH=no; \
;; \
*) \
MULTIARCH=yes; \
;; \
esac; \
cd ${.CURDIR} ; \
fi; \
${AWK} -f ../../mk/scripts/genreadme.awk \
builddependsfile=/dev/null \
dependsfile=/dev/null \
DISTDIR=${DISTDIR} \
MACHINE_ARCH=${MACHINE_ARCH} \
MULTIARCH=$$MULTIARCH \
OPSYS=${OPSYS} \
OS_VERSION=${OS_VERSION} \
PACKAGES=${PACKAGES} \
PKG_SUFX=${PKG_SUFX} \
PKG_URL=${PKG_URL} \
PKGREPOSITORYSUBDIR=${PKGREPOSITORYSUBDIR} \
PKGSRCDIR=${.CURDIR:C|/[^/]*/[^/]*$||} \
when invoking genreadme.awk, set TMPDIR correctly
2003-01-05 01:40:46 +01:00
TMPDIR=${TMPDIR:U/tmp} \
redo the README.html target for increased speed. For packages with no dependencies the speedup is about 2x for ones like gnome with lots of dependencies, the speedup is aroud 400x. The speedup is obtained by generating the dependency list in a more efficient way that only takes 1 make call per dependency instead of potentially many many make calls.
2003-01-04 23:20:47 +01:00
SINGLEPKG=${PKGPATH} \
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
$@.tmp1
redo the README.html target for increased speed. For packages with no dependencies the speedup is about 2x for ones like gnome with lots of dependencies, the speedup is aroud 400x. The speedup is obtained by generating the dependency list in a more efficient way that only takes 1 make call per dependency instead of potentially many many make calls.
2003-01-04 23:20:47 +01:00
@${RM} $@.tmp1
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-pkgtools-version
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
.if !target(show-pkgtools-version)
show-pkgtools-version:
Calculate current pkgtools version once (inline rather than via additional make call), and pass it in the environment to any recursive make calls. 'make fetch' in gnumeric on my system is now > 15% faster checking dependencies.
2000-01-07 13:24:14 +01:00
@${ECHO} ${PKGTOOLS_VERSION}
if ${IDENT} does not exist, assume that the package tools are ok. This is for bootstrapping on systems that don't have RCS installed.
1999-08-31 23:45:45 +02:00
.endif
add print-pkgtools_ver to get version of currently-installed pkg-tools
1998-11-09 15:41:30 +01:00
Add a convenience target called "show-var", which will echo a make(1) variable from the command line: i.e. "make show-var VARNAME=var" will echo var's value
1999-09-28 12:03:16 +02:00
# convenience target, to display make variables from command line
# i.e. "make show-var VARNAME=var", will print var's value
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-var
Add a convenience target called "show-var", which will echo a make(1) variable from the command line: i.e. "make show-var VARNAME=var" will echo var's value
1999-09-28 12:03:16 +02:00
show-var:
show-var: implement quoting via make's :Q modifier, instead of home-brewn.
2001-04-17 18:56:21 +02:00
@${ECHO} ${${VARNAME}:Q}
Add a convenience target called "show-var", which will echo a make(1) variable from the command line: i.e. "make show-var VARNAME=var" will echo var's value
1999-09-28 12:03:16 +02:00
Add a new "show-vars" target which allows to display several makefile variables at once.
2001-10-26 23:42:44 +02:00
# enhanced version of target above, to display multiple variables
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: show-vars
Add a new "show-vars" target which allows to display several makefile variables at once.
2001-10-26 23:42:44 +02:00
show-vars:
.for VARNAME in ${VARNAMES}
@${ECHO} ${${VARNAME}:Q}
.endfor
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: print-build-depends-list
* More target renaming: depends-list => build-depends-list package-depends => run-depends-list print-depends-list => print-build-depends-list print-package-depends => print-run-depends-list * While there, fix print-run-depends-list to not seperate items by \n. Changes discussed with wiz.
2000-08-01 04:51:34 +02:00
.if !target(print-build-depends-list)
print-build-depends-list:
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
. if !empty(BUILD_DEPENDS) || !empty(DEPENDS)
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
@${ECHO_N} 'This package requires package(s) "'
@${ECHO_N} `${MAKE} ${MAKEFLAGS} build-depends-list | ${SORT} -u`
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
@${ECHO} '" to build.'
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: print-run-depends-list
* More target renaming: depends-list => build-depends-list package-depends => run-depends-list print-depends-list => print-build-depends-list print-package-depends => print-run-depends-list * While there, fix print-run-depends-list to not seperate items by \n. Changes discussed with wiz.
2000-08-01 04:51:34 +02:00
.if !target(print-run-depends-list)
print-run-depends-list:
Merge down some more visual cleanup from tv-derecurse to reduce divergence. Part of this, introduce bsd.pkg.use.mk, where the Special Logic invoked by various pkgsrc-wide USE_* variables will be migrated/consolidated.
2005-02-11 16:55:13 +01:00
. if !empty(DEPENDS)
Convert "${ECHO} -n" to "${ECHO_N}", and introduce that into the defs.* files. On most platforms this resolves back to "${ECHO} -n".
2004-04-07 16:26:50 +02:00
@${ECHO_N} 'This package requires package(s) "'
@${ECHO_N} `${MAKE} ${MAKEFLAGS} run-depends-list | ${SORT} -u`
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
@${ECHO} '" to run.'
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
move bulk/{mkreadme,genreadme.awk} to scripts/ as this is a more suitable location. They are not really part of the bulk build stuff.
2003-01-04 22:13:34 +01:00
# This target is used by the mk/scripts/mkreadme script to generate
add a print-summary-data target which will be used by the new bulk README.html generation script. This target gives all information needed by the README.html file in 1 make call.
2002-11-11 22:46:51 +01:00
# README.html files
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: print-summary-data
add a print-summary-data target which will be used by the new bulk README.html generation script. This target gives all information needed by the README.html file in 1 make call.
2002-11-11 22:46:51 +01:00
.if !target(print-summary-data)
print-summary-data:
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
@${ECHO} "depends ${PKGPATH} ${DEPENDS}"
@${ECHO} "build_depends ${PKGPATH} ${BUILD_DEPENDS}"
@${ECHO} "conflicts ${PKGPATH} ${CONFLICTS}"
add a print-summary-data target which will be used by the new bulk README.html generation script. This target gives all information needed by the README.html file in 1 make call.
2002-11-11 22:46:51 +01:00
@${ECHO} "index ${PKGPATH} ${PKGNAME}"
@${ECHO} htmlname ${PKGPATH} ${HTML_PKGNAME:Q}
@${ECHO} homepage ${PKGPATH} ${HOMEPAGE:Q}
@${ECHO} wildcard ${PKGPATH} ${PKGWILDCARD:Q}
@${ECHO} comment ${PKGPATH} ${COMMENT:Q}
@${ECHO} license ${PKGPATH} ${LICENSE:Q}
rework the INDEX file generation. The new approach speeds up things by several orders of magnitude and 'make index' now takes 30 minutes or so instead of several days on my test machine. The approach now is to take one pass through every package and extract some key information including the explicitly listed dependencies. After the data is extracted, the dependencies are flattened in one step which avoids the extremely inefficient recursive make that was previously used.
2003-07-23 11:41:23 +02:00
@if [ "${ONLY_FOR_ARCHS}" = "" ]; then \
${ECHO} "onlyfor ${PKGPATH} any"; \
else \
${ECHO} "onlyfor ${PKGPATH} ${ONLY_FOR_ARCHS}"; \
fi;
add an @ on one of the print-summary-data target lines. This removes some garbage from the database file.
2003-09-15 12:27:12 +02:00
@if [ "${NOT_FOR_OPSYS}" = "" ]; then \
rework the INDEX file generation. The new approach speeds up things by several orders of magnitude and 'make index' now takes 30 minutes or so instead of several days on my test machine. The approach now is to take one pass through every package and extract some key information including the explicitly listed dependencies. After the data is extracted, the dependencies are flattened in one step which avoids the extremely inefficient recursive make that was previously used.
2003-07-23 11:41:23 +02:00
${ECHO} "notfor ${PKGPATH} any"; \
else \
${ECHO} "notfor ${PKGPATH} not ${NOT_FOR_OPSYS}"; \
fi;
@${ECHO} "maintainer ${PKGPATH} ${MAINTAINER}"
@${ECHO} "categories ${PKGPATH} ${CATEGORIES}"
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
@if [ -f ${DESCR_SRC} ]; then \
${ECHO} "descr ${PKGPATH} ${DESCR_SRC:S;${PKGSRCDIR}/;;g}"; \
else \
${ECHO} "descr ${PKGPATH} /dev/null"; \
rework the INDEX file generation. The new approach speeds up things by several orders of magnitude and 'make index' now takes 30 minutes or so instead of several days on my test machine. The approach now is to take one pass through every package and extract some key information including the explicitly listed dependencies. After the data is extracted, the dependencies are flattened in one step which avoids the extremely inefficient recursive make that was previously used.
2003-07-23 11:41:23 +02:00
fi
@${ECHO} "prefix ${PKGPATH} ${PREFIX}"
add a print-summary-data target which will be used by the new bulk README.html generation script. This target gives all information needed by the README.html file in 1 make call.
2002-11-11 22:46:51 +01:00
.endif
Add a "show-license" target, and a "show-licence" alias, to display the package's licence, if it's non-standard.
2001-01-05 18:37:12 +01:00
.if !target(show-license)
show-license show-licence:
@if [ "${LICENSE}" != "" ]; then \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
if ${TEST} -f ${PKGSRCDIR}/licenses/${LICENSE}; then \
Some licenses are pretty long, so use $PAGER instead of $CAT, if $PAGER is set.
2004-01-28 21:29:46 +01:00
if [ "${PAGER}" != "" ]; then \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
${PAGER} ${PKGSRCDIR}/licenses/${LICENSE};\
Some licenses are pretty long, so use $PAGER instead of $CAT, if $PAGER is set.
2004-01-28 21:29:46 +01:00
else \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
${CAT} ${PKGSRCDIR}/licenses/${LICENSE};\
Some licenses are pretty long, so use $PAGER instead of $CAT, if $PAGER is set.
2004-01-28 21:29:46 +01:00
fi \
Check for the existence of the ${LICENSE} file before trying to access it. Pointed out by Thomas Klausner.
2001-03-20 13:39:00 +01:00
else \
${ECHO} "Generic ${LICENSE} information not available"; \
${ECHO} "See the package description (pkg_info -d ${PKGNAME}) for more information."; \
fi \
Add a "show-license" target, and a "show-licence" alias, to display the package's licence, if it's non-standard.
2001-01-05 18:37:12 +01:00
fi
.endif
Add code to include the size of a package into both the installed version and binary packages. The size can be queried then via pkg_info: xfeyrer @ noon% pkg_info -s xv Information for xv-3.10apl1: Size of this package in bytes: 4670692 xfeyrer @ noon% pkg_info -S xv Information for xv-3.10apl1: Size in bytes including required pkgs: 14610165 While doing work on the size code, support for the @src directive was removed (formerly enabled with the pkg_create -s switch, but unused in our pkg system). The new pkg_info -s and -S switches were tested on installed, local (file) and remote (ftp) packages. In bsd.pkg.mk, take special care for pkg_* versions that do not have the pkg_create -s and -S switches and do not record size information there.
1999-11-29 20:48:15 +01:00
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
# Stat all the files of one pkg and sum the sizes up.
#
Add code to include the size of a package into both the installed version and binary packages. The size can be queried then via pkg_info: xfeyrer @ noon% pkg_info -s xv Information for xv-3.10apl1: Size of this package in bytes: 4670692 xfeyrer @ noon% pkg_info -S xv Information for xv-3.10apl1: Size in bytes including required pkgs: 14610165 While doing work on the size code, support for the @src directive was removed (formerly enabled with the pkg_create -s switch, but unused in our pkg system). The new pkg_info -s and -S switches were tested on installed, local (file) and remote (ftp) packages. In bsd.pkg.mk, take special care for pkg_* versions that do not have the pkg_create -s and -S switches and do not record size information there.
1999-11-29 20:48:15 +01:00
# XXX This is intended to be run before pkg_create is called, so the
Add the glue for showing any security vulnerabilities to the README.html files.
2000-09-19 21:29:11 +02:00
# existence of ${PLIST} can be assumed.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: print-pkg-size-this
Tune the print-pkg-size targets to not re-calculate the size for all the depends, but use pkg_info -sS instead. Some code-restructuring and renaming of the targets while there: print-pkg-size-this prints the size of the current pkg, and print-pkg-size-depends prints the size of all the depends ONLY (i.e. NO LONGER including the size of the current pkg). Speeds up "make install" on a pkg requiring teTeX from ~3.2s to ~1.7s.
2000-06-25 05:58:17 +02:00
print-pkg-size-this:
fix typo in last
2000-06-25 08:31:12 +02:00
@${SHCOMMENT} "This pkg's files" ; \
Fix print-pkg-size: when parsing the PLIST, use awk to prepend the proper prefix (defaults to $PREFIX, but can be changed by @cwd from the PLIST). Sent in in PR 9888 by Takahiro Kambe <taca@sky.yamashina.kyoto.jp>
2000-04-15 18:13:07 +02:00
${AWK} 'BEGIN { base = "${PREFIX}/" } \
/^@cwd/ { base = $$2 "/" } \
print-pkg-size-this target: - properly quote any filenames in ${PLIST} before feeding them to xargs. This allows latest textproc/docbook-xsl package to be registered with its correct size because it contains a file with a single quote in its name. XXX pkg_add does not like this filename anyway. - fix the awk script that parse ${PLIST} so pathname containing space (hence awk's NF > 1) are not removed from size calculation. Many thanks to atatat@ to remind me the proper way to quote a shell word.
2003-12-25 17:18:48 +01:00
/^@/ { next } \
{ print base $$0 }' \
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
<${PLIST} \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
| ${SORT} -u \
print-pkg-size-this target: - properly quote any filenames in ${PLIST} before feeding them to xargs. This allows latest textproc/docbook-xsl package to be registered with its correct size because it contains a file with a single quote in its name. XXX pkg_add does not like this filename anyway. - fix the awk script that parse ${PLIST} so pathname containing space (hence awk's NF > 1) are not removed from size calculation. Many thanks to atatat@ to remind me the proper way to quote a shell word.
2003-12-25 17:18:48 +01:00
| ${SED} -e "s/'/'\\\\''/g" -e "s/.*/'&'/" \
Limit xargs to 256 args when passing args to ls for the package size computation. The motivation for this is that kdelibs3 was overflowing ls through xargs.
2004-02-09 06:42:57 +01:00
| ${XARGS} -n 256 ${LS} -ld \
Don't make pkgsrc depend on dc(1) to be present in the system. Since we already use awk so pervasively in pkgsrc, simply use awk in place of dc for simple computations.
2005-01-28 22:05:59 +01:00
| ${AWK} '{ s += $$5; } END { print s; }' \
Add code to include the size of a package into both the installed version and binary packages. The size can be queried then via pkg_info: xfeyrer @ noon% pkg_info -s xv Information for xv-3.10apl1: Size of this package in bytes: 4670692 xfeyrer @ noon% pkg_info -S xv Information for xv-3.10apl1: Size in bytes including required pkgs: 14610165 While doing work on the size code, support for the @src directive was removed (formerly enabled with the pkg_create -s switch, but unused in our pkg system). The new pkg_info -s and -S switches were tested on installed, local (file) and remote (ftp) packages. In bsd.pkg.mk, take special care for pkg_* versions that do not have the pkg_create -s and -S switches and do not record size information there.
1999-11-29 20:48:15 +01:00
Tune the print-pkg-size targets to not re-calculate the size for all the depends, but use pkg_info -sS instead. Some code-restructuring and renaming of the targets while there: print-pkg-size-this prints the size of the current pkg, and print-pkg-size-depends prints the size of all the depends ONLY (i.e. NO LONGER including the size of the current pkg). Speeds up "make install" on a pkg requiring teTeX from ~3.2s to ~1.7s.
2000-06-25 05:58:17 +02:00
# Sizes of required pkgs (only)
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
#
Tune the print-pkg-size targets to not re-calculate the size for all the depends, but use pkg_info -sS instead. Some code-restructuring and renaming of the targets while there: print-pkg-size-this prints the size of the current pkg, and print-pkg-size-depends prints the size of all the depends ONLY (i.e. NO LONGER including the size of the current pkg). Speeds up "make install" on a pkg requiring teTeX from ~3.2s to ~1.7s.
2000-06-25 05:58:17 +02:00
# XXX This is intended to be run before pkg_create is called, so the
kill some trailing whitespace
2003-03-30 07:06:49 +02:00
# dependencies are all installed.
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: print-pkg-size-depends
Tune the print-pkg-size targets to not re-calculate the size for all the depends, but use pkg_info -sS instead. Some code-restructuring and renaming of the targets while there: print-pkg-size-this prints the size of the current pkg, and print-pkg-size-depends prints the size of all the depends ONLY (i.e. NO LONGER including the size of the current pkg). Speeds up "make install" on a pkg requiring teTeX from ~3.2s to ~1.7s.
2000-06-25 05:58:17 +02:00
print-pkg-size-depends:
Do not run xargs if its input is an empty string; xargs in Interix executes the utility even in such a case. Based on the patch provided by Hiramatsu Yoshifumi in PR pkg/25777.
2004-08-04 08:22:39 +02:00
@pkglist=`${MAKE} ${MAKEFLAGS} run-depends-list`; \
if [ "X$${pkglist}" != "X" ]; then \
${ECHO} $${pkglist} \
| ${XARGS} -n 1 ${SETENV} ${PKG_BEST_EXISTS} \
| ${SORT} -u \
| ${XARGS} -n 256 ${SETENV} ${PKG_INFO} -qs \
Don't make pkgsrc depend on dc(1) to be present in the system. Since we already use awk so pervasively in pkgsrc, simply use awk in place of dc for simple computations.
2005-01-28 22:05:59 +01:00
| ${AWK} '/^[0-9]+$$/ { s += $$1; } END { print s; }'; \
Do not run xargs if its input is an empty string; xargs in Interix executes the utility even in such a case. Based on the patch provided by Hiramatsu Yoshifumi in PR pkg/25777.
2004-08-04 08:22:39 +02:00
else \
${ECHO} "0"; \
fi
Split print-pkg-depend-sizes into two pkgs, to work make processing of $DEPENDS with wildcards and relational depends possible. (The previous solution with :Q was not useful as this also quoted the seperator between the single depends, making the DEPENDS useless).
2000-01-10 20:17:27 +01:00
Add code to include the size of a package into both the installed version and binary packages. The size can be queried then via pkg_info: xfeyrer @ noon% pkg_info -s xv Information for xv-3.10apl1: Size of this package in bytes: 4670692 xfeyrer @ noon% pkg_info -S xv Information for xv-3.10apl1: Size in bytes including required pkgs: 14610165 While doing work on the size code, support for the @src directive was removed (formerly enabled with the pkg_create -s switch, but unused in our pkg system). The new pkg_info -s and -S switches were tested on installed, local (file) and remote (ftp) packages. In bsd.pkg.mk, take special care for pkg_* versions that do not have the pkg_create -s and -S switches and do not record size information there.
1999-11-29 20:48:15 +01:00
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
###
### Automatic PLIST generation
### - files & symlinks first
### - @dirrm statements last
### - empty directories are handled properly
### - dirs from mtree files are excluded
Improve, move around and bring up to date comments about print-PLIST target.
2003-12-13 23:15:11 +01:00
### - substitute for platform or package specifics substrings
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
###
### Usage:
### - make install
lose some more pkg/ references
2001-11-28 11:21:46 +01:00
### - make print-PLIST | brain >PLIST
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
###
print-PLIST target pipeline modifications: - reduce the number of components with pervasive awk usage. - handle Info directory file when it is not `info/dir'. Notice that print-PLIST substitutions are now made globally (awk's gsub()) when previously the sed substitution commands did not have the 'g'lobal flag. While here fix PKGVERSION substitution just like PLIST_SUBST's PKGVERSION element is handled in ${PLIST} target i.e. without considering the 'nb${PKGREVISION}' suffix. While here give COMMON_DIRS variable a more "internal type" name: _PRINT_PLIST_COMMON_DIRS.
2003-12-26 18:43:24 +01:00
_PRINT_PLIST_AWK_SUBST={ \
gsub(/${OPSYS}/, "$${OPSYS}"); \
gsub(/${OS_VERSION:S/./\./g}/, "$${OS_VERSION}"); \
gsub(/${MACHINE_GNU_PLATFORM}/, "$${MACHINE_GNU_PLATFORM}"); \
gsub(/${MACHINE_ARCH}/, "$${MACHINE_ARCH}"); \
gsub(/${MACHINE_GNU_ARCH}/, "$${MACHINE_GNU_ARCH}");
.if !empty(LOWER_VENDOR)
_PRINT_PLIST_AWK_SUBST+= gsub(/${LOWER_VENDOR}/, "$${LOWER_VENDOR}");
.endif
_PRINT_PLIST_AWK_SUBST+= \
gsub(/${LOWER_OS_VERSION:S/./\./g}/, "$${LOWER_OS_VERSION}"); \
gsub(/${LOWER_OPSYS}/, "$${LOWER_OPSYS}"); \
gsub(/${PKGNAME_NOREV}/, "$${PKGNAME}"); \
gsub(/${PKGVERSION:S/./\./g:C/nb[0-9]*$$//}/, "$${PKGVERSION}");\
gsub(/${PKGLOCALEDIR}\/locale/, "$${PKGLOCALEDIR}/locale"); \
}
_PRINT_PLIST_AWK_IGNORE= ($$0 ~ /emul\/linux\/proc/)
_PRINT_PLIST_AWK_IGNORE+= || ($$0 ~ /^info\/dir$$/)
.if defined(INFO_DIR) && empty(INFO_DIR:Minfo)
_PRINT_PLIST_AWK_IGNORE+= || ($$0 ~ /^${INFO_DIR:S|/|\\/|g}\/dir$$/)
.endif
Introduce dynamic registration of info files in PLIST. For each file listed in INFO_FILES the installed files matching ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)? are registered as belonging to the package. Also ignore the same files in print-PLIST target. Info files should not be listed in PLIST files anymore but the awk code snippet (_PLIST_AWK_INFO) in charge of actually dealing with info files for the plist target will strip any it finds in them. Now for the reason behind this... Older makeinfo command (before pre-release version 4.5.90 or before release version 4.6) had SPLIT_SIZE_THRESHOLD: when an info file size was >= SPLIT_SIZE_THRESHOLD then the info file was split in chunks of DEFAULT_SPLIT_SIZE size. DEFAULT_SPLIT_SIZE in older versions of makeinfo was 50000 (bytes) and SPLIT_SIZE_THRESHOLD was 70000. So if an info file has a size < 70000 and is processed by makeinfo < 4.6 it is not split in chunks. But if it is processed by makeinfo >= 4.6 texinfo.mk to emulate older version automatically adds `--split-size=50000' to makeinfo's command line and this time the info file is split in chunks... This can lead to PLIST inconsistency from platform to platform or more exactly from version to version of makeinfo. Now we don't have to care about this problem and the info files removal from PLIST files can happen.
2004-01-02 13:28:05 +01:00
.if !empty(INFO_FILES)
. for _f_ in ${INFO_FILES}
Handle compressed info files in PLIST handling.
2004-02-16 14:21:34 +01:00
_PRINT_PLIST_AWK_IGNORE+= || ($$0 ~ /^${INFO_DIR:S|/|\\/|g}\/${_f_:S|+|\+|g}(-[0-9]+)?(\.gz)?$$/)
Introduce dynamic registration of info files in PLIST. For each file listed in INFO_FILES the installed files matching ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)? are registered as belonging to the package. Also ignore the same files in print-PLIST target. Info files should not be listed in PLIST files anymore but the awk code snippet (_PLIST_AWK_INFO) in charge of actually dealing with info files for the plist target will strip any it finds in them. Now for the reason behind this... Older makeinfo command (before pre-release version 4.5.90 or before release version 4.6) had SPLIT_SIZE_THRESHOLD: when an info file size was >= SPLIT_SIZE_THRESHOLD then the info file was split in chunks of DEFAULT_SPLIT_SIZE size. DEFAULT_SPLIT_SIZE in older versions of makeinfo was 50000 (bytes) and SPLIT_SIZE_THRESHOLD was 70000. So if an info file has a size < 70000 and is processed by makeinfo < 4.6 it is not split in chunks. But if it is processed by makeinfo >= 4.6 texinfo.mk to emulate older version automatically adds `--split-size=50000' to makeinfo's command line and this time the info file is split in chunks... This can lead to PLIST inconsistency from platform to platform or more exactly from version to version of makeinfo. Now we don't have to care about this problem and the info files removal from PLIST files can happen.
2004-01-02 13:28:05 +01:00
. endfor
.endif
print-PLIST: * move redundant sed expression into variable to only maintain it once * properly catch grep return code in *both* cases (makes substitution of pkg variables in @dirrm statements / directory names work)
2002-08-03 02:15:30 +02:00
Improve, move around and bring up to date comments about print-PLIST target.
2003-12-13 23:15:11 +01:00
# Common (system) directories not to generate @dirrm statements for
print-PLIST target pipeline modifications: - reduce the number of components with pervasive awk usage. - handle Info directory file when it is not `info/dir'. Notice that print-PLIST substitutions are now made globally (awk's gsub()) when previously the sed substitution commands did not have the 'g'lobal flag. While here fix PKGVERSION substitution just like PLIST_SUBST's PKGVERSION element is handled in ${PLIST} target i.e. without considering the 'nb${PKGREVISION}' suffix. While here give COMMON_DIRS variable a more "internal type" name: _PRINT_PLIST_COMMON_DIRS.
2003-12-26 18:43:24 +01:00
# Reads MTREE_FILE and generate awk statements that will
Improve, move around and bring up to date comments about print-PLIST target.
2003-12-13 23:15:11 +01:00
# sort out which directories NOT to include into the PLIST @dirrm list
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
.if make(print-PLIST)
print-PLIST target pipeline modifications: - reduce the number of components with pervasive awk usage. - handle Info directory file when it is not `info/dir'. Notice that print-PLIST substitutions are now made globally (awk's gsub()) when previously the sed substitution commands did not have the 'g'lobal flag. While here fix PKGVERSION substitution just like PLIST_SUBST's PKGVERSION element is handled in ${PLIST} target i.e. without considering the 'nb${PKGREVISION}' suffix. While here give COMMON_DIRS variable a more "internal type" name: _PRINT_PLIST_COMMON_DIRS.
2003-12-26 18:43:24 +01:00
_PRINT_PLIST_COMMON_DIRS!= ${AWK} 'BEGIN { \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
i=0; \
stack[i]="${PREFIX}" ; \
cwd=""; \
} \
! ( /^\// || /^\#/ || /^$$/ ) { \
if ( $$1 == ".." ){ \
i=i-1; \
cwd = stack[i]; \
} else if ( $$1 == "." ){ \
} else { \
stack[i] = cwd ; \
if ( i == 0 ){ \
cwd = $$1 ; \
} else { \
print-PLIST target change: Make COMMON_DIRS generating an awk script instead of sed expressions hence use awk instead of sed to deal with the mtree file. This get ride of the annoying 'sed: Too many commands' message making print-PLIST unusable with Solaris' seds.
2003-08-12 17:51:29 +02:00
cwd = cwd "\\/" $$1 ; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
} \
print-PLIST target change: Make COMMON_DIRS generating an awk script instead of sed expressions hence use awk instead of sed to deal with the mtree file. This get ride of the annoying 'sed: Too many commands' message making print-PLIST unusable with Solaris' seds.
2003-08-12 17:51:29 +02:00
print "/^" cwd "$$$$/ { next; }"; \
Be consistent throughout the file with line continuations and whitespace. No functional changes.
2002-05-03 14:44:01 +02:00
i=i+1 ; \
} \
} \
print-PLIST target change: Make COMMON_DIRS generating an awk script instead of sed expressions hence use awk instead of sed to deal with the mtree file. This get ride of the annoying 'sed: Too many commands' message making print-PLIST unusable with Solaris' seds.
2003-08-12 17:51:29 +02:00
END { print "{ print $$$$0; }"; } \
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
' <${MTREE_FILE}
.endif
# scan $PREFIX for any files/dirs modified since the package was extracted
# will emit "@exec mkdir"-statements for empty directories
# XXX will fail for data files that were copied using tar (e.g. emacs)!
# XXX should check $LOCALBASE and $X11BASE, and add @cwd statements
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
_PRINT_PLIST_FILES_CMD= \
${FIND} ${PREFIX}/. -xdev -newer ${EXTRACT_COOKIE} \! -type d -print
_PRINT_PLIST_DIRS_CMD= \
${FIND} ${PREFIX}/. -xdev -newer ${EXTRACT_COOKIE} -type d -print
Propagate $ECHO, $GREP, $SORT to print-la-libnames; fixes this on IRIX. Noted by Georg Schwarz <geos@epost.de>.
2004-10-17 00:37:29 +02:00
_PRINT_LA_LIBNAMES= ${SETENV} ECHO=${ECHO:Q} GREP=${GREP:Q} SORT=${SORT:Q} \
${SH} ${.CURDIR}/../../mk/scripts/print-la-libnames
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
Put back LIBTOOLIZE_PLIST, which I meant to do but didn't commit. This is needed by some packages currently which install broken .la files or otherwise don't work with the transform logic yet.
2004-10-21 19:00:10 +02:00
.if !empty(LIBTOOLIZE_PLIST:M[yY][eE][sS])
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
_PRINT_PLIST_LIBTOOLIZE_FILTER?= \
( \
if ${TEST} -d ${WRKDIR}; then \
tmpdir="${WRKDIR}"; \
else \
tmpdir="$${TMPDIR-/tmp}"; \
fi; \
fileslist="$$tmpdir/print.plist.files.$$$$"; \
libslist="$$tmpdir/print.plist.libs.$$$$"; \
while read file; do \
case $$file in \
*.la) \
Propagate $ECHO, $GREP, $SORT to print-la-libnames; fixes this on IRIX. Noted by Georg Schwarz <geos@epost.de>.
2004-10-17 00:37:29 +02:00
${_PRINT_LA_LIBNAMES} $$file >> $$libslist; \
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
;; \
esac; \
${ECHO} "$$file"; \
done > $$fileslist; \
Make print-PLIST work if there are no *.la files.
2004-09-24 17:00:10 +02:00
if ${TEST} -f "$$libslist"; then \
${GREP} -hvxF "`${SORT} -u $$libslist`" "$$fileslist"; \
else \
${CAT} "$$fileslist"; \
fi; \
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
${RM} -f "$$fileslist" "$$libslist"; \
)
Put back LIBTOOLIZE_PLIST, which I meant to do but didn't commit. This is needed by some packages currently which install broken .la files or otherwise don't work with the transform logic yet.
2004-10-21 19:00:10 +02:00
.else
_PRINT_PLIST_LIBTOOLIZE_FILTER?= ${CAT}
.endif
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: print-PLIST
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
.if !target(print-PLIST)
print-PLIST:
Make print-PLIST work again if no linux=procfs mounted. Patch by hubertf.
2001-09-11 18:41:36 +02:00
${_PKG_SILENT}${_PKG_DEBUG}\
${ECHO} '@comment $$'NetBSD'$$'
${_PKG_SILENT}${_PKG_DEBUG}\
backout previous until it can be done a better way.
2004-01-10 10:52:56 +01:00
shlib_type=`${MAKE} ${MAKEFLAGS} show-shlib-type`; \
case $$shlib_type in \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
"a.out") genlinks=1 ;; \
*) genlinks=0 ;; \
esac; \
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
${_PRINT_PLIST_FILES_CMD} \
| ${_PRINT_PLIST_LIBTOOLIZE_FILTER} \
Somewhat tune the generation of "PLIST file" by print-PLIST target: - remove unwanted entries earlier by using ${GREP} -v - substitute more specific bits before lesser ones: ${MACHINE_GNU_PLATFORM} before ${MACHINE_ARCH} and ${MACHINE_GNU_ARCH} - substitute *after* the sort so substitution does not move the pathnames around - add substitution of ${PKGVERSION} - substitute ${PKGLOCALEDIR}/locale (this can help actually catching PLIST, configuration, build, ... problems with respect to the correct locale directory varying from platform to platform)
2002-04-10 14:03:10 +02:00
| ${SORT} \
* Fix processing of @dirrm commands: use the proper RE in sed, and move it after the processing, so empty dirs are still handled properly * Align trailing \s
2000-07-04 13:58:49 +02:00
| ${AWK} ' \
print-PLIST target pipeline modifications: - reduce the number of components with pervasive awk usage. - handle Info directory file when it is not `info/dir'. Notice that print-PLIST substitutions are now made globally (awk's gsub()) when previously the sed substitution commands did not have the 'g'lobal flag. While here fix PKGVERSION substitution just like PLIST_SUBST's PKGVERSION element is handled in ${PLIST} target i.e. without considering the 'nb${PKGREVISION}' suffix. While here give COMMON_DIRS variable a more "internal type" name: _PRINT_PLIST_COMMON_DIRS.
2003-12-26 18:43:24 +01:00
{ sub("${PREFIX}/\\./", ""); } \
${_PRINT_PLIST_AWK_IGNORE} { next; } \
${_PRINT_PLIST_AWK_SUBST} \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
/^@/ { print $$0; next } \
/.*\/lib[^\/]+\.so\.[0-9]+\.[0-9]+\.[0-9]+$$/ { \
print $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
if ('$$genlinks') print $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
if ('$$genlinks') print $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
if ('$$genlinks') print $$0; \
next; \
} \
/.*\/lib[^\/]+\.so\.[0-9]+\.[0-9]+$$/ { \
print $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
if ('$$genlinks') print $$0; \
Properly fix meta-characters quoting in awk regexp specified as strings: `\' needs to be quoted in a string. Also remove some unnecessary `\' in awk strings.
2003-12-13 01:34:39 +01:00
sub("\\.[0-9]+$$", ""); \
The way that shared objects were handled in the PLISTs and bsd.pkg.mk was out of date - it was based on a.out OBJECT_FMT, and added entries in the generated PLISTs to reflect the symlinks that ELF packages uses. It also tried to be clever, and removed and recreated any symbolic links that were created, which has resulted in some fun, especially with packages which use dlopen(3) to load modules. Some recent changes to our ld.so to bring it more into line with other Operating Systems also exposed some cracks. + Modify bsd.pkg.mk and its shared object handling, so that PLISTs now contain the ELF symlinks. + Don't mess about with file system entries when handling shared objects in bsd.pkg.mk, since it's likely that libtool and the BSD *.mk processing will have got it right, and have a much better idea than we do. + Modify PLISTs to contain "ELF symlinks" + On a.out platforms, delete any "ELF symlinks" from the generated PLISTs + On ELF platforms, no extra processing needs to be done in bsd.pkg.mk + Modify print-PLIST target in bsd.pkg.mk to add dummy symlink entries on a.out platforms + Update the documentation in Packages.txt With many thanks to Thomas Klausner for keeping me honest with this.
2001-01-04 16:10:17 +01:00
if ('$$genlinks') print $$0; \
next; \
} \
Introduce support for the PRINT_PLIST_AWK variable. This can be used to provide extra awk expressions that are used while generating the PLIST with the print-PLIST target (before the final 'print' is done), so that it is closer to reality. Ok'ed by wiz@ and hubertf@.
2004-04-12 22:48:07 +02:00
${PRINT_PLIST_AWK} \
print-PLIST target pipeline modifications: - reduce the number of components with pervasive awk usage. - handle Info directory file when it is not `info/dir'. Notice that print-PLIST substitutions are now made globally (awk's gsub()) when previously the sed substitution commands did not have the 'g'lobal flag. While here fix PKGVERSION substitution just like PLIST_SUBST's PKGVERSION element is handled in ${PLIST} target i.e. without considering the 'nb${PKGREVISION}' suffix. While here give COMMON_DIRS variable a more "internal type" name: _PRINT_PLIST_COMMON_DIRS.
2003-12-26 18:43:24 +01:00
{ print $$0; }'
Make print-PLIST work again if no linux=procfs mounted. Patch by hubertf.
2001-09-11 18:41:36 +02:00
${_PKG_SILENT}${_PKG_DEBUG}\
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
for i in `${_PRINT_PLIST_DIRS_CMD} \
print-PLIST target pipeline modifications: - reduce the number of components with pervasive awk usage. - handle Info directory file when it is not `info/dir'. Notice that print-PLIST substitutions are now made globally (awk's gsub()) when previously the sed substitution commands did not have the 'g'lobal flag. While here fix PKGVERSION substitution just like PLIST_SUBST's PKGVERSION element is handled in ${PLIST} target i.e. without considering the 'nb${PKGREVISION}' suffix. While here give COMMON_DIRS variable a more "internal type" name: _PRINT_PLIST_COMMON_DIRS.
2003-12-26 18:43:24 +01:00
| ${SORT} -r \
| ${AWK} ' \
/emul\/linux\/proc/ { next; } \
/${PREFIX:S|/|\\/|g}\/\.$$/ { next; } \
{ sub("${PREFIX}/\\\\./", ""); } \
${_PRINT_PLIST_COMMON_DIRS}'` ; \
* Fix processing of @dirrm commands: use the proper RE in sed, and move it after the processing, so empty dirs are still handled properly * Align trailing \s
2000-07-04 13:58:49 +02:00
do \
Use ${SORT} and ${WC} instead of sort and wc. Also some minor whitespace fixes.
2001-06-12 15:17:01 +02:00
if [ `${LS} -la ${PREFIX}/$$i | ${WC} -l` = 3 ]; then \
In print-PLIST, handle '@exec ${MKDIR}' by PRINT_PLIST_AWK as well.
2004-09-27 02:27:45 +02:00
${ECHO} @exec \$${MKDIR} %D/$$i | ${AWK} ' \
${PRINT_PLIST_AWK} \
{ print $$0; }' ; \
* Fix processing of @dirrm commands: use the proper RE in sed, and move it after the processing, so empty dirs are still handled properly * Align trailing \s
2000-07-04 13:58:49 +02:00
fi ; \
Introduce support for the PRINT_PLIST_AWK variable. This can be used to provide extra awk expressions that are used while generating the PLIST with the print-PLIST target (before the final 'print' is done), so that it is closer to reality. Ok'ed by wiz@ and hubertf@.
2004-04-12 22:48:07 +02:00
${ECHO} @dirrm $$i | ${AWK} ' \
${PRINT_PLIST_AWK} \
{ print $$0; }' ; \
* Fix processing of @dirrm commands: use the proper RE in sed, and move it after the processing, so empty dirs are still handled properly * Align trailing \s
2000-07-04 13:58:49 +02:00
done \
print-PLIST target pipeline modifications: - reduce the number of components with pervasive awk usage. - handle Info directory file when it is not `info/dir'. Notice that print-PLIST substitutions are now made globally (awk's gsub()) when previously the sed substitution commands did not have the 'g'lobal flag. While here fix PKGVERSION substitution just like PLIST_SUBST's PKGVERSION element is handled in ${PLIST} target i.e. without considering the 'nb${PKGREVISION}' suffix. While here give COMMON_DIRS variable a more "internal type" name: _PRINT_PLIST_COMMON_DIRS.
2003-12-26 18:43:24 +01:00
| ${AWK} '${_PRINT_PLIST_AWK_SUBST} { print $$0; }'
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
.endif # target(print-PLIST)
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
# By default, all packages attempt to link into the views.
.if ${PKG_INSTALLATION_TYPE} == "pkgviews"
BUILD_VIEWS?= yes
.endif
Create two new variables: DEFAULT_VIEW.${PKGBASE} default view for ${PKGBASE} VIEWBASE base location of files at run-time The first is meant to be set by package Makefiles or in /etc/mk.conf, while the latter is meant to read/used only. VIEWBASE is either ${LOCALBASE} or or ${LOCALBASE}/${DEFAULT_VIEW.${PKGBASE}} depending on whether the first variable is the empty string or not.
2003-09-12 12:54:49 +02:00
# DEFAULT_VIEW.${PKGBASE} default view for ${PKGBASE}
# VIEWBASE base location of files at run-time
#
DEFAULT_VIEW.${PKGBASE}?= ${DEFAULT_VIEW}
.if ${PKG_INSTALLATION_TYPE} == "pkgviews"
. if empty(DEFAULT_VIEW.${PKGBASE}:M"")
VIEWBASE= ${LOCALBASE}/${DEFAULT_VIEW.${PKGBASE}}
. else
VIEWBASE= ${LOCALBASE}
. endif
.elif ${PKG_INSTALLATION_TYPE} == "overwrite"
VIEWBASE= ${PREFIX}
.endif
Make the value of VIEWBASE available in more settings.
2003-11-23 08:31:18 +01:00
MAKE_ENV+= VIEWBASE=${VIEWBASE}
Create two new variables: DEFAULT_VIEW.${PKGBASE} default view for ${PKGBASE} VIEWBASE base location of files at run-time The first is meant to be set by package Makefiles or in /etc/mk.conf, while the latter is meant to read/used only. VIEWBASE is either ${LOCALBASE} or or ${LOCALBASE}/${DEFAULT_VIEW.${PKGBASE}} depending on whether the first variable is the empty string or not.
2003-09-12 12:54:49 +02:00
PKGVIEWS+= ${DEFAULT_VIEW.${PKGBASE}}
Add print-PLIST target for automatic PLIST generation: - files & symlinks first - @exec/@unexec calls are added for info files - @dirrm statements last - empty directories are handled properly - dirs from mtree files are excluded Usage: - make install - make print-PLIST | brain >pkg/PLIST Moved over from bsd.hf-pkg.mk, after it's lived there for quite some time now.
2000-06-16 11:18:26 +02:00
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
.PHONY: pre-install-fake-pkg
.if !target(pre-install-fake-pkg)
pre-install-fake-pkg:
Add the PKG_PRESERVE capability to bsd.pkg.mk. A package which has PKG_PRESERVE definied in its Makefile will not be able to be deleted, and the capability is carried into binary packages.
2003-01-10 12:57:09 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${RM} -f ${PRESERVE_FILE}
. if defined(PKG_PRESERVE)
${_PKG_SILENT}${_PKG_DEBUG}${DATE} > ${PRESERVE_FILE}
. endif
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${RM} -f ${BUILD_VERSION_FILE}
When constructing the build version information, avoid problems when there is a filename with an embedded space in the directory. Fixes PR 15755 from Simon Burge (simonb@wasabisystems.com). Whilst I'm here, only include the build information for the package Makefile once.
2002-02-28 12:08:55 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
files=""; \
Partitally backout revision 1.932: We *must* record the version of "${.CURDIR}/Makefile" because we otherwise only record the version of the wrong "Makefile" e.g. "x11/xscreensaver-gnome/../xscreensaver/Makefile" for the "xscreensaver-gnome" package. This will break the update target horribly.
2002-03-15 08:16:23 +01:00
for f in ${.CURDIR}/Makefile ${FILESDIR}/* ${PKGDIR}/*; do \
Fix PR 15755 from Simon Burge (simonb@wasabisystems.com) properly. When constructing the build version information, avoid problems when there is a filename with an embedded space in the directory. With thanks to Stoned Elipot for the practical help.
2002-03-12 16:40:24 +01:00
if [ -f "$$f" ]; then \
files="$$files \"$$f\""; \
fi; \
done; \
Deprecate DIGEST_FILE and PATCH_SUM_FILE (replaced by DIGEST_FILE, and not used anymore in pkgsrc).
2001-06-09 14:15:59 +02:00
if [ -f ${DISTINFO_FILE} ]; then \
for f in `${AWK} 'NF == 4 && $$3 == "=" { gsub("[()]", "", $$2); print $$2 }' < ${DISTINFO_FILE}`; do \
Make sure we record the RCS Ids of all of the patch files in the BUILD_VERSION information (pkg_info -b) Correct the English in the licence gripe.
2001-02-15 14:49:04 +01:00
if [ -f ${PATCHDIR}/$$f ]; then \
Fix PR 15755 from Simon Burge (simonb@wasabisystems.com) properly. When constructing the build version information, avoid problems when there is a filename with an embedded space in the directory. With thanks to Stoned Elipot for the practical help.
2002-03-12 16:40:24 +01:00
files="$$files \"${PATCHDIR}/$$f\""; \
Don't try to record non existing patch files in "BuildVersion".
1999-09-03 23:05:58 +02:00
fi; \
Record the names of the patches which were applied to the package (taken from the patch-sum file), plus any local patches, in the BuildVersion file (which can be viewed with the -b option to pkg_info(1)), rather than simply recording the contents of any patches directory.
1999-08-18 12:02:01 +02:00
done; \
fi; \
Use the new flags for pkg_create(1) for the build information and versioning information. Re-arrange the fake-pkg target so that the files exist before they're used.
1998-10-26 18:40:57 +01:00
if [ -d ${PATCHDIR} ]; then \
for f in ${PATCHDIR}/patch-*; do \
case $$f in \
*.orig|*.rej|*~) ;; \
Record the names of the patches which were applied to the package (taken from the patch-sum file), plus any local patches, in the BuildVersion file (which can be viewed with the -b option to pkg_info(1)), rather than simply recording the contents of any patches directory.
1999-08-18 12:02:01 +02:00
${PATCHDIR}/patch-local-*) \
Fix PR 15755 from Simon Burge (simonb@wasabisystems.com) properly. When constructing the build version information, avoid problems when there is a filename with an embedded space in the directory. With thanks to Stoned Elipot for the practical help.
2002-03-12 16:40:24 +01:00
files="$$files \"$$f\"" ;; \
Use the new flags for pkg_create(1) for the build information and versioning information. Re-arrange the fake-pkg target so that the files exist before they're used.
1998-10-26 18:40:57 +01:00
esac; \
done; \
fi; \
* Make PKGSRC_TOPDIR a private variable by renaming it to _PKGSRC_TOPDIR, as it's only used internally by bsd.prefs.mk. * Make _PKGSRCDIR a public variable by renaming it to PKGSRCDIR. Also, generate its value from ${_PKGSRC_TOPDIR} so it's less fragile than the old method of stripping off the last two components of ${.CURDIR}. PKGSRCDIR may now be used after bsd.prefs.mk is defined. * Change all references to _PKGSRCDIR to PKGSRCDIR.
2004-10-07 04:01:37 +02:00
eval ${GREP} '\$$NetBSD' $$files | ${SED} -e 's|^${PKGSRCDIR}/||' > ${BUILD_VERSION_FILE}
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${RM} -f ${BUILD_INFO_FILE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. for def in ${BUILD_DEFS}
remove whitespace after '=' in the generated BUILD_INFO file, making it more sane for parsing.
2003-06-10 07:21:42 +02:00
@${ECHO} ${def}=${${def}:Q} | ${SED} -e 's|^PATH=[^ ]*|PATH=...|' >> ${BUILD_INFO_FILE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endfor
The values of CC are CC_VERSION are redundant. Just use CC_VERSION. We record it only if the package actually used the compiler.
2004-02-10 23:33:47 +01:00
. if !empty(USE_LANGUAGES)
set CC_VERSION for non-gcc and use it in the BUILD_INFO file.
2003-10-11 05:43:14 +02:00
@${ECHO} "CC_VERSION=${CC_VERSION}" >> ${BUILD_INFO_FILE}
The values of CC are CC_VERSION are redundant. Just use CC_VERSION. We record it only if the package actually used the compiler.
2004-02-10 23:33:47 +01:00
. endif
* Change the semantics of the USE_PERL5 variable. USE_PERL5 may now take an optional value of "build" to represent that the package only wants a build dependency on perl. * Add a new public variable "PERL5_PKGSRCDIR" that is either "../../lang/perl58" or "../../lang/perl5" depending on whether we need perl>=5.8.0 or not.
2002-07-24 21:42:21 +02:00
. if defined(USE_PERL5) && (${USE_PERL5} == "run")
remove whitespace after '=' in the generated BUILD_INFO file, making it more sane for parsing.
2003-06-10 07:21:42 +02:00
@${ECHO} "PERL=`${PERL5} --version 2>/dev/null | ${GREP} 'This is perl'`" >> ${BUILD_INFO_FILE}
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
Create a symlink from the "make" requested by the package into ${TOOLS_DIR}/bin/make. If a package wants GNU make, then it should set: USE_GNU_TOOLS+= make in the package Makefile. Allow for USE_GMAKE to continue to work until we get a chance to sweep through pkgsrc and remove the usage of this now-redundant variable. This change allows us to avoid patching makefiles that use a bare "make" command to invoke sub-make processes. Idea suggested by salo@netbsd.org in pkg/22509.
2003-08-26 22:15:23 +02:00
. ifdef _USE_GMAKE
remove whitespace after '=' in the generated BUILD_INFO file, making it more sane for parsing.
2003-06-10 07:21:42 +02:00
@${ECHO} "GMAKE=`${GMAKE} --version | ${GREP} Make`" >> ${BUILD_INFO_FILE}
As part of the build information, record the full pathnames of the shared object "provides" and "requires" information. This is only turned on just now if ${CHECK_SHLIBS} is set to "YES". Example output: % pkg_info -B libutf | grep '^PROVIDES' PROVIDES=/usr/pkg/lib/libutf.so.2 % pkg_info -B ssam | grep '^REQUIRES' REQUIRES=/usr/lib/libc.so.12 REQUIRES=/usr/pkg/lib/libutf.so.2 %
2003-07-22 12:12:47 +02:00
. endif
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
${ECHO} "_PKGTOOLS_VER=${PKGTOOLS_VERSION}" >> ${BUILD_INFO_FILE}
.endif
.PHONY: post-install-fake-pkg
.if !target(post-install-fake-pkg)
post-install-fake-pkg: ${PLIST} ${DESCR} ${MESSAGE}
${_PKG_SILENT}${_PKG_DEBUG} \
if [ ! -f ${PLIST} -o ! -f ${DESCR} ]; then \
${ECHO} "** Missing package files for ${PKGNAME} - installation not recorded."; \
exit 1; \
fi
${_PKG_SILENT}${_PKG_DEBUG} \
if [ ! -d ${_PKG_DBDIR} ]; then \
${RM} -f ${_PKG_DBDIR}; \
${MKDIR} ${_PKG_DBDIR}; \
fi
. if defined(FORCE_PKG_REGISTER)
${_PKG_SILENT}${_PKG_DEBUG}${PKG_ADMIN} delete ${PKGNAME}
. endif
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Fix breakage in previous, installed pkg db files should be installed into ${PKG_DB_TMPDIR} not ${_PKG_DB_TMPDIR}, patch from Geoff Wing.
2005-01-24 10:31:06 +01:00
${ECHO} ${COMMENT:Q} > ${PKG_DB_TMPDIR}/+COMMENT
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
case "${DESCR}" in \
${PKG_DB_TMPDIR}/*|"") ;; \
*) if ${TEST} -f ${DESCR}; then \
Fix breakage in previous, installed pkg db files should be installed into ${PKG_DB_TMPDIR} not ${_PKG_DB_TMPDIR}, patch from Geoff Wing.
2005-01-24 10:31:06 +01:00
${CP} ${DESCR} ${PKG_DB_TMPDIR}/+DESC; \
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
fi ;; \
esac
${_PKG_SILENT}${_PKG_DEBUG} \
case "${MESSAGE}" in \
${PKG_DB_TMPDIR}/*|"") ;; \
*) if ${TEST} -f ${MESSAGE}; then \
Fix breakage in previous, installed pkg db files should be installed into ${PKG_DB_TMPDIR} not ${_PKG_DB_TMPDIR}, patch from Geoff Wing.
2005-01-24 10:31:06 +01:00
${CP} ${MESSAGE} ${PKG_DB_TMPDIR}/+DISPLAY; \
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
fi ;; \
esac
${_PKG_SILENT}${_PKG_DEBUG} \
case ${PKG_INSTALLATION_TYPE} in \
Fix breakage in previous, installed pkg db files should be installed into ${PKG_DB_TMPDIR} not ${_PKG_DB_TMPDIR}, patch from Geoff Wing.
2005-01-24 10:31:06 +01:00
pkgview) ${TOUCH} ${PKG_DB_TMPDIR}/+VIEWS ;; \
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
esac
${_PKG_SILENT}${_PKG_DEBUG} \
case "${INSTALL_FILE}" in \
${PKG_DB_TMPDIR}/*|"") ;; \
*) if ${TEST} -f ${INSTALL_FILE}; then \
Fix breakage in previous, installed pkg db files should be installed into ${PKG_DB_TMPDIR} not ${_PKG_DB_TMPDIR}, patch from Geoff Wing.
2005-01-24 10:31:06 +01:00
${CP} ${INSTALL_FILE} ${PKG_DB_TMPDIR}/+INSTALL; \
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
fi ;; \
esac
${_PKG_SILENT}${_PKG_DEBUG} \
case "${DEINSTALL_FILE}" in \
${PKG_DB_TMPDIR}/*|"") ;; \
*) if ${TEST} -f ${DEINSTALL_FILE}; then \
Fix breakage in previous, installed pkg db files should be installed into ${PKG_DB_TMPDIR} not ${_PKG_DB_TMPDIR}, patch from Geoff Wing.
2005-01-24 10:31:06 +01:00
${CP} ${DEINSTALL_FILE} ${PKG_DB_TMPDIR}/+DEINSTALL; \
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
fi ;; \
esac
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
${_PKG_SILENT}${_PKG_DEBUG}${RM} -f ${SIZE_PKG_FILE} ${SIZE_ALL_FILE}
Don't chmod shlibs that are symlinks; fixes issue noted by <grant@netbsd.org>. Also don't do the ldd-REQUIRES dance unless SHLIB_HANDLING=YES (because you don't want to run a native ldd on a foreign binary).
2004-10-27 15:47:41 +02:00
. if ${SHLIB_HANDLING} == "YES" && ${CHECK_SHLIBS} == "YES"
As part of the build information, record the full pathnames of the shared object "provides" and "requires" information. This is only turned on just now if ${CHECK_SHLIBS} is set to "YES". Example output: % pkg_info -B libutf | grep '^PROVIDES' PROVIDES=/usr/pkg/lib/libutf.so.2 % pkg_info -B ssam | grep '^REQUIRES' REQUIRES=/usr/lib/libc.so.12 REQUIRES=/usr/pkg/lib/libutf.so.2 %
2003-07-22 12:12:47 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
add support for extracting binary/library requires/provides on Darwin, and simplify the logic to reduce code duplication. fixes PR pkg/22492 from Julien T. Letessier.
2003-08-23 11:11:19 +02:00
case "${LDD}" in \
"") ldd=`${TYPE} ldd 2>/dev/null | ${AWK} '{ print $$NF }'`;; \
*) ldd="${LDD}"; \
esac; \
As part of the build information, record the full pathnames of the shared object "provides" and "requires" information. This is only turned on just now if ${CHECK_SHLIBS} is set to "YES". Example output: % pkg_info -B libutf | grep '^PROVIDES' PROVIDES=/usr/pkg/lib/libutf.so.2 % pkg_info -B ssam | grep '^REQUIRES' REQUIRES=/usr/lib/libc.so.12 REQUIRES=/usr/pkg/lib/libutf.so.2 %
2003-07-22 12:12:47 +02:00
case "${OBJECT_FMT}" in \
ELF) bins=`${SETENV} PREFIX=${PREFIX} ${AWK} '/^(bin|sbin|libexec)\// { print ENVIRON["PREFIX"] "/" $$0 }' ${PLIST} || ${TRUE}`; \
libs=`${SETENV} PREFIX=${PREFIX} ${AWK} '/^lib\/lib.*\.so\.[0-9]+$$/ { print ENVIRON["PREFIX"] "/" $$0 }' ${PLIST} || ${TRUE}`; \
if ${TEST} "$$bins" != "" -o "$$libs" != ""; then \
Ignore any errors from ldd when calculating reuires and provides
2003-07-22 15:48:48 +02:00
requires=`($$ldd $$bins $$libs 2>/dev/null || ${TRUE}) | ${AWK} 'NF == 3 { print $$3 }' | ${SORT} -u`; \
As part of the build information, record the full pathnames of the shared object "provides" and "requires" information. This is only turned on just now if ${CHECK_SHLIBS} is set to "YES". Example output: % pkg_info -B libutf | grep '^PROVIDES' PROVIDES=/usr/pkg/lib/libutf.so.2 % pkg_info -B ssam | grep '^REQUIRES' REQUIRES=/usr/lib/libc.so.12 REQUIRES=/usr/pkg/lib/libutf.so.2 %
2003-07-22 12:12:47 +02:00
fi; \
Fix shlib execute perm pattern (was broken after 1.1537). Also, to address concerns raised on tech-pkg, only issue the warning about execute permission if PKG_DEVELOPER is set.
2004-10-28 16:05:56 +02:00
linklibs=`${SETENV} PREFIX=${PREFIX} ${AWK} '/^[^@].*\.so\.[0-9\.]+$$/ { print ENVIRON["PREFIX"] "/" $$0 }' ${PLIST} || ${TRUE}`; \
For the ELF form, make sure shlibs are installed +x. If not, issue a warning, fix it, and move along.
2004-10-13 19:52:46 +02:00
for i in $${linklibs}; do \
Don't chmod shlibs that are symlinks; fixes issue noted by <grant@netbsd.org>. Also don't do the ldd-REQUIRES dance unless SHLIB_HANDLING=YES (because you don't want to run a native ldd on a foreign binary).
2004-10-27 15:47:41 +02:00
if ${TEST} -r $$i -a ! -x $$i -a ! -h $$i; then \
Fix shlib execute perm pattern (was broken after 1.1537). Also, to address concerns raised on tech-pkg, only issue the warning about execute permission if PKG_DEVELOPER is set.
2004-10-28 16:05:56 +02:00
${TEST} "${PKG_DEVELOPER:Uno}" = "no" || \
${ECHO} "$$i: installed without execute permission; fixing (should use [BSD_]INSTALL_LIB)"; \
For the ELF form, make sure shlibs are installed +x. If not, issue a warning, fix it, and move along.
2004-10-13 19:52:46 +02:00
${CHMOD} +x $$i; \
fi; \
done; \
add support for extracting binary/library requires/provides on Darwin, and simplify the logic to reduce code duplication. fixes PR pkg/22492 from Julien T. Letessier.
2003-08-23 11:11:19 +02:00
;; \
Mach-O) bins=`${SETENV} PREFIX=${PREFIX} ${AWK} '/^(bin|sbin|libexec)\// { print ENVIRON["PREFIX"] "/" $$0 }' ${PLIST} || ${TRUE}`; \
libs=`${SETENV} PREFIX=${PREFIX} ${AWK} '/^lib\/lib.*\.dylib/ { print ENVIRON["PREFIX"] "/" $$0 }' ${PLIST} || ${TRUE}`; \
if ${TEST} "$$bins" != "" -o "$$libs" != ""; then \
fix a bug in the extraction of shlib dependencies on Darwin which caused non-shared objects to be mistakenly added to the buildinfo REQUIRES.
2004-01-05 08:15:39 +01:00
requires=`($$ldd $$bins $$libs 2>/dev/null || ${TRUE}) | ${AWK} '/compatibility version/ { print $$1 }' | ${SORT} -u`; \
add support for extracting binary/library requires/provides on Darwin, and simplify the logic to reduce code duplication. fixes PR pkg/22492 from Julien T. Letessier.
2003-08-23 11:11:19 +02:00
fi; \
;; \
esac; \
for i in "" $$libs; do \
${TEST} "$$i" = "" && continue; \
${ECHO} "PROVIDES=$$i" >> ${BUILD_INFO_FILE}; \
done; \
for req in "" $$requires; do \
${TEST} "$$req" = "" && continue; \
${ECHO} "REQUIRES=$$req" >> ${BUILD_INFO_FILE}; \
done
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. endif
fake-pkg: * print registered dependencies only once - no functional change. * some whitespace changes
2000-11-22 11:40:21 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Tune the print-pkg-size targets to not re-calculate the size for all the depends, but use pkg_info -sS instead. Some code-restructuring and renaming of the targets while there: print-pkg-size-this prints the size of the current pkg, and print-pkg-size-depends prints the size of all the depends ONLY (i.e. NO LONGER including the size of the current pkg). Speeds up "make install" on a pkg requiring teTeX from ~3.2s to ~1.7s.
2000-06-25 05:58:17 +02:00
size_this=`${MAKE} ${MAKEFLAGS} print-pkg-size-this`; \
size_depends=`${MAKE} ${MAKEFLAGS} print-pkg-size-depends`; \
* some whitespace fix * s/.MAKEFLAGS/MAKEFLAGS/ in one place * add check-shlibs target. Need to at least document that later. :) (Should be added at end of root-install target, to check if the pkg's binaries/libs will find the shared libs they need)
2000-06-27 22:14:45 +02:00
${ECHO} $$size_this >${SIZE_PKG_FILE}; \
Don't make pkgsrc depend on dc(1) to be present in the system. Since we already use awk so pervasively in pkgsrc, simply use awk in place of dc for simple computations.
2005-01-28 22:05:59 +01:00
${ECHO} $$size_this $$size_depends \
| ${AWK} '{ print $$1 + $$2; }' >${SIZE_ALL_FILE}
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
.endif
# Fake installation of package so that user can pkg_delete it later.
# Also, make sure that an installed package is recognized correctly in
# accordance to the @pkgdep directive in the packing lists.
#
.PHONY: register-pkg
.if !target(register-pkg)
register-pkg: post-install-fake-pkg
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
${_PKG_SILENT}${_PKG_DEBUG} \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
doit=yes; \
case ${PKG_INSTALLATION_TYPE} in \
overwrite) if [ -d ${_PKG_DBDIR}/${PKGNAME} ]; then \
doit=no; \
fi ;; \
esac; \
case $$doit in \
yes) \
Fix PACKAGE_DEPENDS_QUICK when installing on Solaris - escape the argument from the shell, so that we can install packages on Solaris again, and use canonical ${GREP} and ${AWK} programs.
2000-03-13 15:19:15 +01:00
${ECHO_MSG} "${_PKGSRC_IN}> Registering installation for ${PKGNAME}"; \
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
${PKG_CREATE} ${PKG_ARGS_INSTALL} -O ${PKGFILE} > ${PKG_DB_TMPDIR}/+CONTENTS; \
Undo implicit recursive package registration. The buildlink[23] system is already taking care of this where it's truly needed. OK'd by agc.
2004-04-02 21:09:18 +02:00
list="`${MAKE} ${MAKEFLAGS} run-depends-list ECHO_MSG=${TRUE} | ${SORT} -u`" ; \
Change a few more uses of PKG_INFO into PKG_BEST_EXISTS. This corrects the package size calculation and finding out for which packages we need to add to +REQUIRED_BY.
2003-09-13 13:23:35 +02:00
for realdep in `${ECHO} $$list | ${XARGS} -n 1 ${SETENV} ${PKG_BEST_EXISTS} | ${SORT} -u`; do \
Add quotes round one place where it may be useful, and add a "plist" target to create the PLIST on the fly. (This is just an alias for the existing ${PLIST} target, but with that you have to type the full path which is inconvenient).
2000-01-05 17:55:32 +01:00
if ${TEST} -z "$$realdep"; then \
Improve wording of warning if required package is not found, which causes a dependency to be not registered, to minimize confusion like in pkg/9043.
2000-09-10 14:11:25 +02:00
${ECHO} "$$dep not installed - dependency NOT registered" ; \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
elif [ -d ${_PKG_DBDIR}/$$realdep ]; then \
replace use of ${TEST} -e with -f or -d as appropriate. use consistent shell syntax. this allows the pkgsrc framework to work with IRIX 5.3 and Solaris' sh(1) built-in test(1).
2004-07-03 23:59:04 +02:00
if ${TEST} ! -f ${_PKG_DBDIR}/$$realdep/+REQUIRED_BY; then \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
${TOUCH} ${_PKG_DBDIR}/$$realdep/+REQUIRED_BY; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
fi; \
${AWK} 'BEGIN { found = 0; } \
$$0 == "${PKGNAME}" { found = 1; } \
{ print $$0; } \
END { if (!found) { printf("%s\n", "${PKGNAME}"); }}' \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
< ${_PKG_DBDIR}/$$realdep/+REQUIRED_BY > ${_PKG_DBDIR}/$$realdep/reqby.$$$$; \
${MV} ${_PKG_DBDIR}/$$realdep/reqby.$$$$ ${_PKG_DBDIR}/$$realdep/+REQUIRED_BY; \
Add Solaris support to pkgsrc, using Christos Zoulas' NetBSD emulation package. Tested for some packages on Solaris 2.6 on Sparc.
1999-03-08 18:26:54 +01:00
${ECHO} "${PKGNAME} requires installed package $$realdep"; \
+ explain "Unacceptable license" message a bit better. + create a +BUILD_VERSION file when creating the package, containing the RCS Id strings of files pertinent to the package build. + re-format one block of code + s/echo/${ECHO}/ + don't give out unnecessary internal pkg information to users (the exact location of registration of the package does not matter).
1998-10-12 21:40:33 +02:00
fi; \
Rearrange the register-pkg and post-install-fake-pkg code. All of the meta-data files are generated and stored in ${PKG_DB_TMPDIR} by the *-install-fake-pkg targets, and then all of the contents of that directory are simply copied into ${PKG_DBDIR}/${PKGNAME} by register-pkg. This is intended to make (DE)INSTALL scripts be more powerful. Currently, they have no way to keep state on their own. Now, they can keep state in the current working directory. When invoked by pkg_add(1), the current working directory is ${PKG_DBDIR}/${PKGNAME}, so the state files are already stored in the correct place. When invoked by bsd.pkg.mk, the current working directory is ${PKG_DB_TMPDIR}, and any state files generated by the (DE)INSTALL scripts are copied into the correct place by the register-pkg target.
2005-01-23 22:57:38 +01:00
done; \
case ${PKG_INSTALLATION_TYPE} in \
overwrite) ${RM} -rf ${_PKG_DBDIR}/${PKGNAME} ;; \
esac; \
${MKDIR} ${_PKG_DBDIR}/${PKGNAME}; \
${CP} ${PKG_DB_TMPDIR}/+* ${_PKG_DBDIR}/${PKGNAME}; \
;; \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
esac
Allow BUILD_VIEWS to be "yes" or "YES".
2003-09-08 03:13:02 +02:00
. if (${PKG_INSTALLATION_TYPE} == "pkgviews") && \
!empty(BUILD_VIEWS:M[yY][eE][sS])
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
${_PKG_SILENT}${_PKG_DEBUG}${MAKE} ${MAKEFLAGS} build-views
. endif # pkgviews
Rework the way that packages are registered during a "make install". We split the fake-pkg target into three smaller pieces: pre-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that don't depend on the installed files. post-install-fake-pkg - Create files in ${WRKSRC}/.pkgdb meant to be installed into ${PKG_DBDIR}/${PKGNAME} that may depend on the installed files. register-pkg - Copy the files in ${WRKSRC}/.pkgdb and create a few additional meta-files in ${PKG_DBDIR}/${PKGNAME}. The idea is that ${WRKSRC}/.pkgdb contains as much of the state as possible that will be copied into ${PKG_DBDIR}/${PKGNAME}. The INSTALL and DEINSTALL scripts generated by bsd.pkg.install.mk are now invoked with ${WRKSRC}/.pkgdb as the working directory, so the scripts may use the working directory to store temporary files, regardless of whether they're invoked from bsd.pkg.mk or by pkg_add.
2005-01-23 21:45:22 +01:00
.endif
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.PHONY: build-views
build-views: do-su-build-views
.PHONY: do-su-build-views
do-su-build-views:
@${ECHO_MSG} "${_PKGSRC_IN}> Building views for ${PKGNAME}"
${_PKG_SILENT}${_PKG_DEBUG} \
realtarget="real-su-build-views"; \
action="build-views"; \
${_SU_TARGET}
.PHONY: real-su-build-views
.if !target(real-su-build-views)
real-su-build-views:
. if (${PKG_INSTALLATION_TYPE} == "pkgviews") && defined(PKGVIEWS)
${_PKG_SILENT}${_PKG_DEBUG} \
${MKDIR} ${LOCALBASE}; \
for v in ${PKGVIEWS}; do \
case "$$v" in \
"") dbdir=${PKG_DBDIR}; viewname=standard ;; \
*) dbdir=${LOCALBASE}/$$v/.dbdir; viewname=$$v ;; \
esac; \
${ECHO} "=> Performing package view clash check for ${PKGNAME} in $$viewname view"; \
pkg=`${PKG_INFO_CMD} -K $$dbdir -e ${PKGBASE} || ${TRUE}`; \
case "$$pkg" in \
"") ;; \
*) ${ECHO} "*** ${PKGBASE} exists in $$viewname view - package $$pkg ***"; \
${ECHO} "*** Not hoisting ${PKGNAME} into $$viewname view"; \
continue ;; \
esac; \
${ECHO} "=> Performing package view overwrite check for ${PKGNAME} in $$viewname view"; \
dups=`${SETENV} PLIST_IGNORE_FILES="${_PLIST_IGNORE_FILES}" ${PKG_VIEW} --view=$$v check ${PKGNAME} || ${TRUE}`; \
case "$$dups" in \
"") ;; \
*) ${ECHO} "***********************************************************"; \
${ECHO} "**** The following symbolic links will be overwritten *****"; \
for f in $$dups; do \
${LS} -l ${LOCALBASE}/$$v/$$f; \
done; \
${ECHO} "***********************************************************"; \
;; \
esac; \
${ECHO} "=> Linking package into $$viewname view"; \
${SETENV} PLIST_IGNORE_FILES="${_PLIST_IGNORE_FILES}" ${PKG_VIEW} --view=$$v add ${PKGNAME}; \
done
. else
${_PKG_SILENT}${_PKG_DEBUG}${DO_NADA}
. endif
.endif
.PHONY: remove-views
remove-views: do-su-remove-views
.PHONY: do-su-remove-views
do-su-remove-views:
@${ECHO_MSG} "${_PKGSRC_IN}> Removing ${PKGNAME} from views"
${_PKG_SILENT}${_PKG_DEBUG} \
realtarget="real-su-remove-views"; \
action="remove-views"; \
${_SU_TARGET}
.PHONY: real-su-remove-views
.if !target(real-su-remove-views)
real-su-remove-views:
. if (${PKG_INSTALLATION_TYPE} == "pkgviews") && defined(PKGVIEWS)
${_PKG_SILENT}${_PKG_DEBUG} \
for v in ${PKGVIEWS}; do \
case "$$v" in \
"") dbdir=${PKG_DBDIR}; viewname=standard ;; \
*) dbdir=${LOCALBASE}/$$v/.dbdir; viewname=$$v ;; \
esac; \
${ECHO} "=> Removing package from $$viewname view"; \
${SETENV} PLIST_IGNORE_FILES="${_PLIST_IGNORE_FILES}" ${PKG_VIEW} --view=$$v delete ${PKGNAME}; \
done
.else
${_PKG_SILENT}${_PKG_DEBUG}${DO_NADA}
. endif
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.endif
This is NetBSD. s/port/package/g
1999-08-10 12:48:23 +02:00
# Depend is generally meaningless for arbitrary packages, but if someone wants
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
# one they can override this. This is just to catch people who've gotten into
make -> ${MAKE}, as per FreeBSD's PR 6341.
1998-04-19 14:48:07 +02:00
# the habit of typing `${MAKE} depend all install' as a matter of course.
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
#
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: depend
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(depend)
depend:
.endif
# Same goes for tags
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: tags
Initial import of FreeBSD's ports makefiles.
1997-08-20 12:21:05 +02:00
.if !target(tags)
tags:
.endif
- Add OpenBSD and FreeBSD Ids in the header - Changes from FreeBSD's V1.164: * Warn if mtree-file not found * On 'make reinstall' gets DEPENDS_TARGET set to 'reinstall' instead of 'install' - Changes from OpenBSD's V1.14: * Documentation and implementation for ONLY_FOR_ARCHS * Documentation and implementation for WRKOBJDIR * Documentation on HAVE_MOTIF and MOTIF_STATIC: set in /etc/mk.conf, not /etc/make.conf * Set NOMANCOMPRESS=no to still compress man-pages (should probably be set to 'yes') * DEF_UMASK=022 * add ${MACHINE} suffix to WRKDIR if OBJMACHINE is set * .include <bsd.own.mk>
1997-09-28 02:44:15 +02:00
Some packages use bsd-style .mk files when building, and so any manual pages that are installed will be gzip-compressed, if MANZ is set, or not if MANZ is not set. If the package uses bsd-style .mk files, the variable MANCOMPRESSED_IF_MANZ should be set to a value of "yes" in the package Makefile. This replaces the previous method of specific inclusion of bsd.prefs.mk, followed by a check for MANZ and conditional assignment of MANCOMPRESSED. Add appropriate documentation, and change all necessary ocurrences in package Makefiles.
1999-07-02 10:37:20 +02:00
# if automatic manual page compression is done by the package according
# to MANZ's value, set MANCOMPRESSED if MANZ is set
Add a "show-distfiles" target, used to show the distfiles and patchfiles (including any subdir) for a package. If ${IGNORE} is set, for any reason, do not print anything. The output is relative to ${DISTDIR}. This target can be used to identify out of date distfiles and patchfiles in ${DISTDIR}, in conjunction with suitable filters.
1999-06-28 13:42:28 +02:00
.if defined(MANCOMPRESSED_IF_MANZ) && defined(MANZ)
MANCOMPRESSED= yes
Correctly handle the common case where "MANZ" setting is protected by ".ifdef BSD_PKG_MK" in /etc/mk.conf, but "MANZ" handling, in the package, relys on the system's "bsd.man.mk", which _CAN_ _NOT_ see variables protected by ".ifdef BSD_PKG_MK". By passing "MANZ" in through "MAKE_ENV", only if "MANZ" is defined, we ensure that the PLIST handling and the package's own handling are on the same wavelength.
2001-03-26 19:04:50 +02:00
MAKE_ENV+= MANZ="${MANZ}"
Add a "show-distfiles" target, used to show the distfiles and patchfiles (including any subdir) for a package. If ${IGNORE} is set, for any reason, do not print anything. The output is relative to ${DISTDIR}. This target can be used to identify out of date distfiles and patchfiles in ${DISTDIR}, in conjunction with suitable filters.
1999-06-28 13:42:28 +02:00
.endif
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
# generate ${PLIST} from ${PLIST_SRC} by:
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
# - substituting for PLIST_SUBST entries
# - fixing list of man-pages according to MANZ, MANINSTALL.
* Rename some targets to fit into the usual "do-XXX calls real-XXX" scheme * In real-su-install, do not call "make ${PLIST}" manually, but rather depend on the ${PLIST} file being there for the do-su-install target. * Break out shlib-handlink from real-su-install, and put it into target do-shlib-handling, which will either touch then PLIST (when called via the ${PLIST} target) or do the necessary steps to setup shared library handling (creating symlinks on ELF, running ldconfig on a.out, etc., when called via real-su-install) * Removed some unnecessary tests (check if $PLIST is there when it can be assumed to be there, ...)
2000-11-29 14:18:22 +01:00
# - adding symlinks for shared libs (ELF) or ldconfig calls (a.out).
move documentation to packages.7. also, remove ts=4 crap.
1998-06-05 14:45:53 +02:00
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
# plist awk pattern-action statement to handle MANINSTALL
_PLIST_AWK_MANINSTALL= \
{ \
if (!"${MANINSTALL:Mmaninstall}" && \
match($$0, "^([^/]*/)*man/([^/]*/)?man[1-9ln]") ) { \
next; \
Major changes in manual page handling - deal with imake's man pages in Solaris, and consolidate other manual page handling.
1999-03-12 11:19:14 +01:00
} \
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
if (!"${MANINSTALL:Mcatinstall}" && \
match($$0, "^([^/]*/)*man/([^/]*/)?cat[1-9ln]") ) { \
next; \
Major changes in manual page handling - deal with imake's man pages in Solaris, and consolidate other manual page handling.
1999-03-12 11:19:14 +01:00
} \
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
}
# plist awk pattern-action statement to strip '.gz' from man
# entries
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
_PLIST_AWK_STRIP_MANZ= \
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
/^([^\/]*\/)*man\/([^\/]*\/)?(man[1-9ln]\/.*[1-9ln]|cat[1-9ln]\/.*0)\.gz$$/ { \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
$$0 = substr($$0, 1, length($$0) - 3); \
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
}
# plist awk pattern-action statement to add '.gz' to man entries
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
_PLIST_AWK_ADD_MANZ= \
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
/^([^\/]*\/)*man\/([^\/]*\/)?(man[1-9ln]\/.*[1-9ln]|cat[1-9ln]\/.*0)$$/ { \
Whitespace police - no functional change
2004-12-21 18:30:07 +01:00
$$0 = $$0 ".gz"; \
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
}
fix a file descriptor leak in the darwin PLIST so->dylib translation script that broke packages that installed too many shared libraries (eg libwww).
2004-03-17 17:36:28 +01:00
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
# plist awk pattern-action statement to handle PLIST_SUBST substitutions
# BEWARE: the awk script quote is closed and reopened around the
# string argument of gsub() calls so historic quoting semantic of
# PLIST_SUBST is preserved.
# XXX `_str_quote_{start,end}_' is a gross hack to work around weird word
# splitting.
_PLIST_AWK_SUBST= { ${PLIST_SUBST:S|=|\\}/,_str_quote_start_|:S|$|_str_quote_end_);|:S|^|gsub(/\\\$\\{|:S|_str_quote_start_|"'|g:S|_str_quote_end_|'"|g} }
# plist awk pattern-action statement to rewrite "imake installed" catman pages
# as plain manpages.
_PLIST_AWK_IMAKE_MAN= \
/^([^\/]*\/)*man\/([^\/]*\/)?cat[1-9ln]\/.*0$$/ { \
n = match($$0, "/cat[1-9ln]"); \
sect = sprintf(".%s", substr($$0, n + 4, 1)); \
sub("/cat", "/man"); \
sub("\\.0$$", sect); \
}
Introduce dynamic registration of info files in PLIST. For each file listed in INFO_FILES the installed files matching ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)? are registered as belonging to the package. Also ignore the same files in print-PLIST target. Info files should not be listed in PLIST files anymore but the awk code snippet (_PLIST_AWK_INFO) in charge of actually dealing with info files for the plist target will strip any it finds in them. Now for the reason behind this... Older makeinfo command (before pre-release version 4.5.90 or before release version 4.6) had SPLIT_SIZE_THRESHOLD: when an info file size was >= SPLIT_SIZE_THRESHOLD then the info file was split in chunks of DEFAULT_SPLIT_SIZE size. DEFAULT_SPLIT_SIZE in older versions of makeinfo was 50000 (bytes) and SPLIT_SIZE_THRESHOLD was 70000. So if an info file has a size < 70000 and is processed by makeinfo < 4.6 it is not split in chunks. But if it is processed by makeinfo >= 4.6 texinfo.mk to emulate older version automatically adds `--split-size=50000' to makeinfo's command line and this time the info file is split in chunks... This can lead to PLIST inconsistency from platform to platform or more exactly from version to version of makeinfo. Now we don't have to care about this problem and the info files removal from PLIST files can happen.
2004-01-02 13:28:05 +01:00
# plist awk pattern-action statement to handle info files:
Handle compressed info files in PLIST handling.
2004-02-16 14:21:34 +01:00
# generate list of files matching
# ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)?(.gz)?
Introduce dynamic registration of info files in PLIST. For each file listed in INFO_FILES the installed files matching ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)? are registered as belonging to the package. Also ignore the same files in print-PLIST target. Info files should not be listed in PLIST files anymore but the awk code snippet (_PLIST_AWK_INFO) in charge of actually dealing with info files for the plist target will strip any it finds in them. Now for the reason behind this... Older makeinfo command (before pre-release version 4.5.90 or before release version 4.6) had SPLIT_SIZE_THRESHOLD: when an info file size was >= SPLIT_SIZE_THRESHOLD then the info file was split in chunks of DEFAULT_SPLIT_SIZE size. DEFAULT_SPLIT_SIZE in older versions of makeinfo was 50000 (bytes) and SPLIT_SIZE_THRESHOLD was 70000. So if an info file has a size < 70000 and is processed by makeinfo < 4.6 it is not split in chunks. But if it is processed by makeinfo >= 4.6 texinfo.mk to emulate older version automatically adds `--split-size=50000' to makeinfo's command line and this time the info file is split in chunks... This can lead to PLIST inconsistency from platform to platform or more exactly from version to version of makeinfo. Now we don't have to care about this problem and the info files removal from PLIST files can happen.
2004-01-02 13:28:05 +01:00
# for `filename' being each word of INFO_FILES in turn.
# Notes:
# - first the filenames matching ${PREFIX}/${INFO_DIR}/filename*
# are generated with ls then they are filtered by the exact pattern.
# - ${PREFIX}/${INFO_DIR}/filename is single quoted and single quote
# escaped
# XXX When all info file entries will be removed from PLIST files
# the non-BEGIN pattern-action statements generated below will be retired.
_PLIST_AWK_INFO=
We only need the special PLIST info-file handling if we're using static PLISTs. Dynamic PLISTs automatically list the installed info files in the PLIST.
2004-01-18 01:51:30 +01:00
.if ${PLIST_TYPE} == "static"
. if !empty(INFO_FILES)
. for _f_ in ${INFO_FILES}
Introduce dynamic registration of info files in PLIST. For each file listed in INFO_FILES the installed files matching ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)? are registered as belonging to the package. Also ignore the same files in print-PLIST target. Info files should not be listed in PLIST files anymore but the awk code snippet (_PLIST_AWK_INFO) in charge of actually dealing with info files for the plist target will strip any it finds in them. Now for the reason behind this... Older makeinfo command (before pre-release version 4.5.90 or before release version 4.6) had SPLIT_SIZE_THRESHOLD: when an info file size was >= SPLIT_SIZE_THRESHOLD then the info file was split in chunks of DEFAULT_SPLIT_SIZE size. DEFAULT_SPLIT_SIZE in older versions of makeinfo was 50000 (bytes) and SPLIT_SIZE_THRESHOLD was 70000. So if an info file has a size < 70000 and is processed by makeinfo < 4.6 it is not split in chunks. But if it is processed by makeinfo >= 4.6 texinfo.mk to emulate older version automatically adds `--split-size=50000' to makeinfo's command line and this time the info file is split in chunks... This can lead to PLIST inconsistency from platform to platform or more exactly from version to version of makeinfo. Now we don't have to care about this problem and the info files removal from PLIST files can happen.
2004-01-02 13:28:05 +01:00
_PLIST_AWK_INFO+= \
BEGIN { \
cmd="${_f_}"; gsub("'\''", "\\'\''", cmd); \
sub("^", "${LS} '\''${PREFIX}/${INFO_DIR}/", cmd); \
sub("$$", "'\''*", cmd); \
while ((cmd | getline l) > 0) { \
Handle compressed info files in PLIST handling.
2004-02-16 14:21:34 +01:00
if (match(l, ".*/${_f_:S|+|\\\+|g}(-[0-9]+)?(\\.gz)?$$")) { \
Introduce dynamic registration of info files in PLIST. For each file listed in INFO_FILES the installed files matching ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)? are registered as belonging to the package. Also ignore the same files in print-PLIST target. Info files should not be listed in PLIST files anymore but the awk code snippet (_PLIST_AWK_INFO) in charge of actually dealing with info files for the plist target will strip any it finds in them. Now for the reason behind this... Older makeinfo command (before pre-release version 4.5.90 or before release version 4.6) had SPLIT_SIZE_THRESHOLD: when an info file size was >= SPLIT_SIZE_THRESHOLD then the info file was split in chunks of DEFAULT_SPLIT_SIZE size. DEFAULT_SPLIT_SIZE in older versions of makeinfo was 50000 (bytes) and SPLIT_SIZE_THRESHOLD was 70000. So if an info file has a size < 70000 and is processed by makeinfo < 4.6 it is not split in chunks. But if it is processed by makeinfo >= 4.6 texinfo.mk to emulate older version automatically adds `--split-size=50000' to makeinfo's command line and this time the info file is split in chunks... This can lead to PLIST inconsistency from platform to platform or more exactly from version to version of makeinfo. Now we don't have to care about this problem and the info files removal from PLIST files can happen.
2004-01-02 13:28:05 +01:00
sub("^${PREFIX}/", "", l); \
print l; \
} \
} \
close(cmd); \
} \
/^${INFO_DIR:S|/|\\/|g}\/${_f_}(-[0-9]+)?$$/ { next; }
We only need the special PLIST info-file handling if we're using static PLISTs. Dynamic PLISTs automatically list the installed info files in the PLIST.
2004-01-18 01:51:30 +01:00
. endfor
. endif
Introduce dynamic registration of info files in PLIST. For each file listed in INFO_FILES the installed files matching ${PREFIX}/${INFO_DIR}/filename(-[0-9]+)? are registered as belonging to the package. Also ignore the same files in print-PLIST target. Info files should not be listed in PLIST files anymore but the awk code snippet (_PLIST_AWK_INFO) in charge of actually dealing with info files for the plist target will strip any it finds in them. Now for the reason behind this... Older makeinfo command (before pre-release version 4.5.90 or before release version 4.6) had SPLIT_SIZE_THRESHOLD: when an info file size was >= SPLIT_SIZE_THRESHOLD then the info file was split in chunks of DEFAULT_SPLIT_SIZE size. DEFAULT_SPLIT_SIZE in older versions of makeinfo was 50000 (bytes) and SPLIT_SIZE_THRESHOLD was 70000. So if an info file has a size < 70000 and is processed by makeinfo < 4.6 it is not split in chunks. But if it is processed by makeinfo >= 4.6 texinfo.mk to emulate older version automatically adds `--split-size=50000' to makeinfo's command line and this time the info file is split in chunks... This can lead to PLIST inconsistency from platform to platform or more exactly from version to version of makeinfo. Now we don't have to care about this problem and the info files removal from PLIST files can happen.
2004-01-02 13:28:05 +01:00
.endif
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
# plist awk pattern-action statement to expand libtool archives into
# shared and/or static libraries.
#
.if ${PLIST_TYPE} == "dynamic"
_PLIST_AWK_LIBTOOL?= # empty
Only allow LIBTOOLIZE_PLIST=no if USE_LIBTOOL is *not* used, setting BROKEN if both are specified.
2004-10-21 19:18:44 +02:00
.elif empty(LIBTOOLIZE_PLIST:M[yY][eE][sS])
BROKEN+= "USE_LIBTOOL requires LIBTOOLIZE_PLIST=yes for proper PLIST generation"
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
.else
_PLIST_AWK_LIBTOOL?= \
Fix _PLIST_AWK_LIBTOOL not to pass @comment lines to print-la-libnames.
2004-10-12 16:34:57 +02:00
/^[^@].*\.la$$/ { \
Propagate $ECHO, $GREP, $SORT to print-la-libnames; fixes this on IRIX. Noted by Georg Schwarz <geos@epost.de>.
2004-10-17 00:37:29 +02:00
system("cd ${PREFIX} && ${_PRINT_LA_LIBNAMES} " $$0) \
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
}
.endif
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
# _PLIST_AWK_SCRIPT hold the complete awk script for plist target.
#
_PLIST_AWK_SCRIPT= '
# Do the substitutions
# See comments above about _PLIST_AWK_SUBST: it contains single quotes!
# So _PLIST_AWK_SCRIPT is intended to be single quoted.
_PLIST_AWK_SCRIPT+= ${_PLIST_AWK_SUBST}
Remove trailing whitespace.
2004-02-12 14:16:02 +01:00
# Generated entries for info files
.if !empty(INFO_FILES)
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
_PLIST_AWK_SCRIPT+= ${_PLIST_AWK_INFO}
Remove trailing whitespace.
2004-02-12 14:16:02 +01:00
.endif
* Do the *.la expansion within the current _PLIST_AWK_SCRIPT framework. We no longer require that LIBTOOL_LA_FILES be defined in the package Makefile, and the libtool archives should once again be listed in the PLIST. * Add a new yes/no variable "LIBTOOLIZE_PLIST" to control whether to have bsd.pkg.mk automatically expand *.la files in PLISTs into the true library names represented by the libtool archives. * Rename the "transform-la" script to "print-la-libnames" which more correctly reflects its function. Many thanks to Todd Vierling for the original implementation and for his contructive comments on how to improve the changes in this commit.
2004-09-10 21:51:50 +02:00
# Expand libtool archives
_PLIST_AWK_SCRIPT+= ${_PLIST_AWK_LIBTOOL}
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
# Strip the '.gz' suffixes on man entries
_PLIST_AWK_SCRIPT+= ${_PLIST_AWK_STRIP_MANZ}
# Deal with MANINSTALL and man entries
_PLIST_AWK_SCRIPT+= ${_PLIST_AWK_MANINSTALL}
# Deal with "imake installed" catman pages
* Push the imake- and xmkmf-handling into the new tools framework. * Get rid of an explicit check for ${_IMAKE_MAKE} == ${GMAKE} in bsd.pkg.mk to check for whether we need to depend on gmake or not. Instead, we now note in Linux.mk that packages that need imake will also need to use gmake by setting _IMAKE_TOOLS+=gmake. * Push the definition of MAKE_PROGRAM from bsd.pkg.mk into make.mk where it's closer to related code.
2005-05-02 23:10:02 +02:00
.if defined(USE_IMAKE) || !empty(USE_TOOLS:Mimake) && ${_PREFORMATTED_MAN_DIR} == "man"
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
_PLIST_AWK_SCRIPT+= ${_PLIST_AWK_IMAKE_MAN}
.endif
# Add '.gz' suffixes on man entries if needed
.if defined(MANZ)
_PLIST_AWK_SCRIPT+= ${_PLIST_AWK_ADD_MANZ}
.endif
# Print the entry
_PLIST_AWK_SCRIPT+= { print $$0; }
#
_PLIST_AWK_SCRIPT+= '
Major changes in manual page handling - deal with imake's man pages in Solaris, and consolidate other manual page handling.
1999-03-12 11:19:14 +01:00
Generalize how the PLIST contents are generated: GENERATE_PLIST is a sequence of commands, terminating in a semicolon, that outputs contents for a PLIST to stdout and is appended to the contents of ${PLIST_SRC}. Convert the perl5 packlist handling to use GENERATE_PLIST.
2002-04-16 06:10:08 +02:00
# GENERATE_PLIST is a sequence of commands, terminating in a semicolon,
# that outputs contents for a PLIST to stdout and is appended to
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
# the contents of ${PLIST_SRC}.
Generalize how the PLIST contents are generated: GENERATE_PLIST is a sequence of commands, terminating in a semicolon, that outputs contents for a PLIST to stdout and is appended to the contents of ${PLIST_SRC}. Convert the perl5 packlist handling to use GENERATE_PLIST.
2002-04-16 06:10:08 +02:00
#
GENERATE_PLIST?= ${TRUE};
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.if ${PLIST_TYPE} == "dynamic"
_PLIST_IGNORE_CMD= \
( while read i; do \
ignore=no; \
for p in ${_PLIST_IGNORE_FILES}; do \
case "$$i" in \
$$p) ignore=yes; break ;; \
esac; \
done; \
[ "$$ignore" = "yes" ] || ${ECHO} "$$i"; \
done )
_GENERATE_PLIST= \
${FIND} ${PREFIX} \! -type d -print | ${SORT} | \
${SED} -e "s|^${PREFIX}/||" | \
${_PLIST_IGNORE_CMD}; \
${FIND} ${PREFIX} -type d -print | ${SORT} -r | \
${GREP} -v "^${PREFIX}$$" | \
${_PLIST_IGNORE_CMD} | \
${SED} -e "s|^${PREFIX}/|@unexec ${RMDIR} -p %D/|" \
-e "s,$$, 2>/dev/null || ${TRUE},";
.else
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
_GENERATE_PLIST= ${CAT} ${PLIST_SRC}; ${GENERATE_PLIST}
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
.endif
Generalize how the PLIST contents are generated: GENERATE_PLIST is a sequence of commands, terminating in a semicolon, that outputs contents for a PLIST to stdout and is appended to the contents of ${PLIST_SRC}. Convert the perl5 packlist handling to use GENERATE_PLIST.
2002-04-16 06:10:08 +02:00
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: plist
Add quotes round one place where it may be useful, and add a "plist" target to create the PLIST on the fly. (This is just an alias for the existing ${PLIST} target, but with that you have to type the full path which is inconvenient).
2000-01-05 17:55:32 +01:00
plist: ${PLIST}
Consolidate PLIST_SRC default generation and the plethora of failsafe error checks into labelled blocks toward the top of the file. While here, remove the historic (now completely empty; it was replaced by PKG_FAIL_REASON some time ago) uptodate-pkgtools target and the references to it.
2005-02-11 17:57:45 +01:00
.if ${PLIST_TYPE} == "static"
${PLIST}: ${PLIST_SRC}
.endif
${PLIST}:
Major changes in manual page handling - deal with imake's man pages in Solaris, and consolidate other manual page handling.
1999-03-12 11:19:14 +01:00
${_PKG_SILENT}${_PKG_DEBUG} \
Rewrite most of the pipeline generating ${PLIST} as a single awk script.
2003-12-16 13:36:54 +01:00
{ ${_GENERATE_PLIST} } | ${AWK} ${_PLIST_AWK_SCRIPT} \
* Remove old PLIST-mi/md.shared/md.static that was only used for perl in a previous (pkgsrc) life. * add "lint" target to run pkglint * Add commented out direction on how to use bzip2 compressed binary packages * Some cleanup (clean target, ...) Reviewed by Thomas Klauser <wiz@netbsd.org>.
2000-12-30 12:19:04 +01:00
> ${PLIST}; \
${MAKE} ${MAKEFLAGS} do-shlib-handling \
Merge pkgviews-mk branch into the HEAD by running: cd pkgsrc/mk cvs update -Pd -A cvs update -Pd -j pkgviews-mk-base -j pkgviews-mk
2003-09-02 08:59:37 +02:00
SHLIB_PLIST_MODE=1
Automatically append HOMEPAge to DESCR files, and remove homepage URLs from individual DESCR files that had them.
1998-08-26 18:50:48 +02:00
Move message/${MESSAGE} target around, out of plist target and its supporting code. While here comment the message target a bit.
2003-12-23 19:27:24 +01:00
# generate ${MESSAGE} from ${MESSAGE_SRC} by substituting
Remove trailing whitespace.
2004-02-12 14:16:02 +01:00
# for MESSAGE_SUBST entries
Move message/${MESSAGE} target around, out of plist target and its supporting code. While here comment the message target a bit.
2003-12-23 19:27:24 +01:00
.PHONY: message
message: ${MESSAGE}
.ifdef MESSAGE
${MESSAGE}: ${MESSAGE_SRC}
${_PKG_SILENT}${_PKG_DEBUG} \
if [ -z "${MESSAGE_SRC}" ]; then \
${ECHO} "${MESSAGE_SRC} not found."; \
${ECHO} "Please set MESSAGE_SRC correctly."; \
else \
${CAT} ${MESSAGE_SRC} | \
${SED} ${MESSAGE_SUBST_SED} \
> ${MESSAGE}; \
fi
.endif
Automatically append HOMEPAge to DESCR files, and remove homepage URLs from individual DESCR files that had them.
1998-08-26 18:50:48 +02:00
# generate ${DESCR} from ${DESCR_SRC} by:
# - Appending the homepage URL, if any
Mark all targets that do not correspond to real files with the .PHONY attribute.
2003-07-24 00:02:48 +02:00
.PHONY: descr
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
descr: ${DESCR}
Automatically append HOMEPAge to DESCR files, and remove homepage URLs from individual DESCR files that had them.
1998-08-26 18:50:48 +02:00
${DESCR}: ${DESCR_SRC}
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
@${CAT} ${DESCR_SRC} >${DESCR}
Automatically append HOMEPAge to DESCR files, and remove homepage URLs from individual DESCR files that had them.
1998-08-26 18:50:48 +02:00
.if defined(HOMEPAGE)
Speedups! Major changes: * Avoid a big bunch of subshells/forks/execs by using make(1) features; should be good for some minor speedups. * Use shell builtins for echo, true, false, test. Using full paths here causes fork&execs, making things slow. Only changed for NetBSD. * Rename all references to ${.MAKEFLAGS} to ${MAKEFLAGS}. See make(1). * Evaluate all !=s only once, and add them to MAKEFLAGS * Pass ${MAKEFLAGS} to ALL ${MAKE} calls Now PKGTOOLS_VERSION is really only set once in a whole "make package" cycle, even across dependencies. Minor changes: * SORTED_MASTER_SITES was renamed SORTED_MASTER_SITES_CMD and is now (only) execuded in the places where it's needed. * Don't check for /usr/bin/fetch - we never had it, and if someone depends on it he can set FETCH_CMD instead. * Don't fall back to README.port * Replace some commands by their uppercase equivalents (sed -> ${SED}, same for grep, awk, ...) * Some whitespace changes to improve readability => With these changes, times for a "make clean" in x11/kde go down to 20%, time for "make readme" goes down to 33%!
2000-06-03 09:15:28 +02:00
@\
${ECHO} >>${DESCR} ; \
${ECHO} "Homepage:" >>${DESCR} ; \
${ECHO} '${HOMEPAGE}' >>${DESCR}
Automatically append HOMEPAge to DESCR files, and remove homepage URLs from individual DESCR files that had them.
1998-08-26 18:50:48 +02:00
.endif
Pull in bulk-build targets if BATCH is set
2000-07-28 03:16:27 +02:00
(*) Nuke the _BUILDLINK_SUBST_USE target now that subst.mk does the same thing. (*) Move the code to adjust Makefiles to refer to the correct PKGLOCALEDIR into bsd.pkg.mk and reimplement it using subst.mk so that it's non-USE_BUILDLINK2 packages can use it, too. (*) Reimplement the PKGCONFIG_OVERRIDE code in bsd.pkg.mk using subst.mk. (*) Reimplement the unbuildlink code in bsd.buildlink2.mk using subst.mk.
2003-08-27 13:29:56 +02:00
.include "../../mk/subst.mk"
Pull in bulk-build targets if BATCH is set
2000-07-28 03:16:27 +02:00
#
# For bulk build targets (bulk-install, bulk-package), the
* Remove RUN_DEPENDS (left in packages.7 intentionally) * Renamed misc-depends to run-depends, and drop (old) run-depends altogether Code reviewed by wiz.
2000-08-01 04:16:58 +02:00
# BATCH variable must be set in /etc/mk.conf:
Pull in bulk-build targets if BATCH is set
2000-07-28 03:16:27 +02:00
#
.if defined(BATCH)
Change indentation (whitespace changes only!): .if .if .if . if .else => . else .endif . endif .endif .endif Patch contributed by Marc Espie <espie@liafa.jussieu.fr> in preparation of merging latest changes with the OpenPackages project. Thanks Marc!
2001-07-19 00:43:32 +02:00
. include "../../mk/bulk/bsd.bulk-pkg.mk"
Pull in bulk-build targets if BATCH is set
2000-07-28 03:16:27 +02:00
.endif
Allow the creation of error handlers for shell commands in make targets. PKG_ERROR_CLASSES is a list of error handlers, PKG_ERROR_HANDLER.<class>, to create. The default error handler will rethrow the exit code after emitting an error message that may be specified per-class with PKG_ERROR_MSG.<class>. Create error messages for the configure and build phases that inform the user of places to go looking for why the build didn't complete.
2005-01-10 19:34:48 +01:00
# Create a PKG_ERROR_HANDLER shell command for each class listed in
# PKG_ERROR_CLASSES. The error handler is meant to be invoked within
# a make target.
#
.for _class_ in ${PKG_ERROR_CLASSES}
PKG_ERROR_HANDLER.${_class_}?= { \
ec=$$?; \
for str in ${PKG_ERROR_MSG.${_class_}}; do \
${ECHO} "${_PKGSRC_IN}> $$str"; \
done; \
exit $$ec; \
}
.endfor
Reference in a new issue Copy permalink