Commit graph

3104 commits

Author SHA1 Message Date
jlam
1a98a9895b sjg recommends escaping the bare left curly brace in the make variable
modifer expression.
2004-02-23 09:15:33 +00:00
jlam
00b950724b Back out previous because that brace is not supposed to be there. 2004-02-23 08:58:22 +00:00
dogcow
48adda54db add a missing } to the end of the line; fixes PR/24525. 2004-02-23 08:18:59 +00:00
jlam
911e6a0991 Allow setting CCACHE_DIR in the environment or in /etc/mk.conf. 2004-02-22 12:31:10 +00:00
jlam
2db74558c6 Use BUILD_ENV instead of just {CONFIGURE,MAKE}_ENV to pass CCACHE_HASHCC
to ccache.  This now correctly passes CCACHE_HASHCC to ccache if the
package has a custom do-build target that invokes ${CC} directly.
2004-02-21 12:31:38 +00:00
jlam
82d8d5199f 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 12:29:23 +00:00
jlam
0af79f88fa Pass LDFLAGS (minus any libraries) to the compiler when creating the
trivial libtool archive.
2004-02-21 10:35:57 +00:00
jlam
2849792ae1 Symlink the libtool-1.4.m4 file into any directory in which there's a
configure script under ${WRKSRC}.  The symlink is called "libtool.m4",
which is the name expected by the autoconf tools.  The symlinking is
only done if ${AUTOCONF_REQD} == 2.13 and USE_LIBTOOL is defined.  This
allows autoconf-2.13, autoreconf-2.13 and aclocal-2.13 to use the older
libtool.m4 file when regenerating files.  This is okay, because we later
override the generated libtool file anyway.

This fixes problems when libtool-1.5.x is installed and a package tries
to invoke autoconf-2.13 programs at pre-configure time, e.g., gnome-games,
xpuyopuyo, faad2, libwww, etc.  These packages no longer need to be
worked around by bumping AUTOCONF_REQD to 2.50.
2004-02-20 09:58:49 +00:00
jlam
8132b7353c Provide a much more complete list of version number patterns <= 2.95.3. 2004-02-20 07:18:19 +00:00
jlam
75bd3d729a If for some reason, _GCC_REQD doesn't match any of the version pattern
lists, then just assume that we need gcc3.  The rest of the code in the
file assumes that we either need gcc2 or gcc3, so this provides a
fall-through, default value.
2004-02-20 07:10:06 +00:00
jlam
65af2d1fe6 According to the documentation on version numbering of shared libraries at:
http://fink.sourceforge.net/doc/porting/shared.php

It's okay to link against a name like "libqt.2.3.0.dylib" using
"-lqt.2.3.0", which means we never need to do anything more than just
strip the trailing ".dylib" from shared library names when converting from
a full path to "-L... -l...".  This should fix PR 24402.
2004-02-19 18:39:27 +00:00
jlam
3d05ef90b1 Use the mipspro-ucode bits if we're using the mipspro-ucode compiler. 2004-02-19 13:17:13 +00:00
jlam
b4081e6300 Resolve ${LOCALBASE} to its phyiscal path and mangle it back into
${LOCALBASE}.  Some packages' configure scripts resolve all paths to
physical paths, and since buildlink3 suppresses references outside of
${LOCALBASE}, it can break the build of those packages.

This should fix the problem noted by Nathan Williams in the thread
titled "x11/tk build failure" at:

	http://mail-index.netbsd.org/tech-pkg/2004/02/17/0004.html
2004-02-19 12:50:47 +00:00
jlam
2a766c79fa Don't spew files in the package directory if the BUILDLINK_DEPENDS.<pkg>
uses {} to list several "or" dependencies.
2004-02-19 07:41:44 +00:00
jlam
5c40aefbbe Output a more informative/desriptive error than:
cd: cd BUILDLINK_PREFIX.foo_not_found

We now say that the package wasn't installed at the time that we were
trying to buildlink its files.
2004-02-18 19:16:52 +00:00
jlam
3bc670ced3 Remove some checks for PREPEND_PATH as we don't need them. 2004-02-18 14:23:42 +00:00
jlam
c40ee95964 * 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 13:32:38 +00:00
jlam
455aa176d8 Initial support for the MIPSpro ucode compilers found on IRIX 5.3. The
ucode compilers only support building the "o32" ABI.
2004-02-18 11:18:42 +00:00
jlam
78080df597 Allow overriding CC_VERSION_STRING and CC_VERSION. 2004-02-18 11:13:54 +00:00
jlam
990623913c Typo in comment. 2004-02-18 11:10:35 +00:00
jlam
192e73f9e3 Typo in comment. 2004-02-18 05:23:37 +00:00
jlam
bedc09eef8 Require x11-links-0.23 for the correct set of links with XFree86-4.3.x. 2004-02-17 16:56:47 +00:00
jlam
350ca28970 Fix the glob pattern to match >= dependencies correctly. 2004-02-17 16:07:37 +00:00
jlam
50e37dd0bd Do "dependency reduction" where dependency lists of the form "foo>=0.2
foo>=0.3 foo>=0.4" are collapsed into a single "foo>=0.4" dependency.
2004-02-17 14:49:17 +00:00
jlam
2eb8293c09 G/C the unused GNU_MISSING_OVERRIDE logic, as we don't need it anymore
after the changes in revision 1.27 of tools.mk to hide the auto* tools
between pre-configure and do-configure.
2004-02-17 12:16:39 +00:00
jlam
59ed73505b Move up the creation of the auto* stubs to just before the do-configure
target.  This still allows any autoconf/automake we execute at
pre-configure time to find its own scripts, but forces the configure script
find our stubs instead.  This should prevent the need to constantly
override the missing script using GNU_MISSING_OVERRIDE.
2004-02-17 12:09:16 +00:00
jlam
bad42f1f3f +less obfuscated errors 2004-02-17 08:59:44 +00:00
grant
1c81e8e6a1 also ignore -pedantic, SunPro doesn't need to be told ;-) 2004-02-16 22:22:38 +00:00
jlam
e95593a479 Symlink any *.pc files into ${BUILDLINK_DIR} for dependencies and force
pkg-config to only look in ${BUILDLINK_DIR}/lib/pkgconfig for *.pc files.
This will correctly hide the presence of software from configure scripts
that query pkg-config for that information.

Idea suggested by Julio M. Merino Vidal.
2004-02-16 21:01:17 +00:00
jlam
2a3f166b38 Add some TODO items for me. 2004-02-16 20:50:50 +00:00
hubertf
4b44008c84 Add RCS ID 2004-02-16 20:25:35 +00:00
jlam
f72a85a532 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 20:25:18 +00:00
seb
28ddfa550f Handle compressed info files in PLIST handling. 2004-02-16 13:21:34 +00:00
grant
0e24610e47 ignore -ansi, SunPro doesn't need to be told to support ANSI C. 2004-02-16 12:18:15 +00:00
seb
b6b6fefbf6 Quote '+' in awk regexp related to PLIST and INFO_FILES.
As seen in devel/libg++ with its libg++.info file...
2004-02-16 11:28:13 +00:00
jlam
bacf725a9d Move section that checks for PREFER_PKGSRC a little higher so that it
actually has some effect.
2004-02-16 03:08:52 +00:00
jmmv
3653afd74f ml locale goes after mk. 2004-02-15 23:31:49 +00:00
jlam
d48d056fe7 Whitespace nits. 2004-02-15 14:50:11 +00:00
jlam
d859a79bb1 Aside from everything under ${PREFIX}/include, grab all *.h files for
symlinking into ${BUILDLINK_DIR} for a more complete set of headers.
2004-02-15 14:37:00 +00:00
jlam
f9efcd5148 Create wrappers for the compilers for every language supported by the
compiler set.  This will cause the libtool configuration found in several
packages to use the correct C++ compiler even though the package doesn't
use C++.  This was causing bugs when CXXFLAGS contained flags not
understood by the system gcc-2.95.3.
2004-02-15 13:34:33 +00:00
grant
b56215fa53 avoid another long command line calling tar(1). 2004-02-15 06:17:10 +00:00
lukem
7f5e976c4d Set CONF_FILES_MODE from @CONF_FILES_MODE@. 2004-02-15 03:01:08 +00:00
seb
daa975015f Use relative to ${WRKSRC} pathnames in do-gnu-missing-override target
by cd'ing into it. Also symlink the missing script instead of copying it.
2004-02-14 18:29:17 +00:00
jlam
c24ff672f7 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 17:18:49 +00:00
jlam
9822f69382 Make CONFIG_{GUESS,SUB}_OVERRIDE slighly more powerful by allowing them to
be lists of shell globs relative to ${WRKSRC}.
2004-02-14 15:22:42 +00:00
jlam
e694fd214d Change into ${WRKSRC} first so that shell globs expand properly. 2004-02-14 15:16:31 +00:00
jlam
9ae3d0260c Some shells barf if the for loop list is empty, so avoid that. 2004-02-14 13:58:34 +00:00
jlam
f598249781 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 13:54:28 +00:00
jlam
01d241c9c1 Convert a debugging definition back into the real thing. 2004-02-14 11:28:28 +00:00
grant
5fdae71af5 zoularis is deprecated, don't backup zoularis.mk. 2004-02-14 04:06:29 +00:00