It is more consistent with the tex.buildlink3.mk name. Also, if a package
really needs latex, it just has to set TEX_ACCEPTED to latex distributions
altough today, all TEX_ACCEPTED possibilities are latex distributions
Remove GNU_CONFIGURE=no from pkgsrc Makefile.
Add "--mksrc none" to arguments and
add --prefix=${PREFIX} and --sysconfdir=${PKG_SYSCONFDIR}
and remove ${CONFIGURE_ARGS} for boot-strap.
Note that I didn't import the files that only change was a
CVS $Id$ tag.
From ChangeLog:
2005-11-05 Simon J. Gerraty <sjg@void.crufty.net>
* Makefile.in (BMAKE_VERSION): bump to 20051105
* configure.in: always set default_sys_path
default is ${prefix}/share/mk
- remove prefix_sys_path, anyone wanting more than above
needs to set it manually.
2005-11-04 Simon J. Gerraty <sjg@void.crufty.net>
* boot-strap: make this a bit easier for pkgsrc folk.
bootstrap still fails on IRIX64 since MACHINE_ARCH gets set to
'mips' while pkgsrc wants 'mipseb' or 'mipsel'
2005-11-02 Simon J. Gerraty <sjg@void.crufty.net>
* Makefile.in (BMAKE_VERSION): bump to 20051102
* job.c (JobFinish): fix likely ancient merge lossage
fix from Todd Vierling.
* boot-strap (srcdir): allow setting mksrc=none
2005-10-31 Simon J. Gerraty <sjg@void.crufty.net>
* Makefile.in (BMAKE_VERSION): bump to 20051031
* ranlib.h: skip on OSF too.
(NetBSD PR 31864)
Fix major problem brought to the surface by buildlink3 on Interix. For C++
or F77 libraries only, the soname did not always make it into the final
binary, resulting in runtime link failures in some cases and runtime linking
with a missing major version in others.
This is unfortunately a flag day for Interix + pkgsrc. Identifying all the
PKGREVISIONs normally requiring a bump is a little too unwieldy. However,
existing binaries not experiencing the runtime link failure will continue
to run, but will be linking against the unversioned ".so" at runtime until
the binaries have been rebuilt.
While here, fix a couple wrong assumptions in the libtool config for
Interix. shlibvar_overrides_runpath should have been "no", and
hardcode_direct should have been "yes".
XXX: the package is heavily gnome-infested, pulling in half the
known universe via lablgtk2. Only the gnome-canvas bindings
are really needed; there's a local copy of lablgtk2 in the
m-viz source tree now, perhaps use that?
0.12:
- change the way displayed branches are selected: now one can view
any set of branches
- allow the ancestry graph view to be limited by date
- the `query' window now uses monotone selectors
- change the reconnection algorithm
- cert values can be copied to the clipboard (via a context menu)
- DnD support: accept drops from a file manager
- rewrite the sqlite bindings (use parsed statements now)
0.11:
- a new `query' window to search for revision matching some criterion
- some DnD support: drag a node and drop it in a terminal to copy its
revision id
- tweak the reconnection algorithm
- build with a shared sqlite library by default
- add a preference option for the location of the `dot' program
0.10:
- try to reconnect disconnected parts of the graph (by adding grey edges)
- libgnomecanvas hack to speed up drawing
- somehow display the output of monotone diff, even if it is not valid UTF-8
- middle clicking on the graph with a (prefix) of an id selected will
center the view on this id
- support builds with a shared sqlite library
(--with-shared-sqlite option to configure)
+ * src/check.h.in: Added include of stddef.h for NULL definition
+ * doc/tutorial.sgml: Fixed sourceforge bug #1216502
+ Refactored messaging to use the new tmpfile() method all the way,
+ removing the message keys, pipes, pipe entries and pipe list. This
+ makes the messaging work with forking tests, and also with threading
+ tests on linux 2.4 (on 2.6 it already worked). Added check_fork and
+ check_waitpid_and_exit to be used for forking tests.
+ * src/check.h.in: Fixed compatibility with gcc 2.95.3 according
+ to sourceforge patch #1161654.
+ * src/check_run.c, tests/check_check_master.c: Changed timeout
+ error message according to sourceforge feature request #1121452.
+ * tests/ex_xml_output.c, tests/ex_log_output.c, tests/ex_output.c,
+ src/check_log.c:
+ Fixed memory leaks.
+ * check.m4, config.h.in, configure.in, src/check_pack.c: Fixed
+ quoting and added configure test for stdint.h.
+ * src/check_run.c, tests/check_check_master.c: Use strsignal
+ to print describing text for signals.
+ * doc/tutorial.sgml: Documented signals handling and timeouts.
+ * tests/check_check_master.c src/check.h.in:
+ Changed failure message for fail_if.
+ * src/check.c, src/check.h.in, src/check_impl.h, src/check_run.c,
+ tests/check_check_master.c, tests/check_check_sub.c:
+ Added support for timeouts on tests, enabling detection of
+ eternal loops as errors.
+ * src/check.c, src/check.h.in, src/check_impl.h, src/check_run.c,
+ tests/check_check_master.c, tests/check_check_sub.c:
+ Added support for testing on expected signals. Implementation
+ courtesy of Lucas Di Pentima and Cesar Ballardini. Also cleaned
+ up the test verification to simplify merging of new tests.
+ * src/check.c, src/check_list.c, src/check_list.h, src/check_log.c,
+ src/check_msg.c, tests/check_list.c:
+ Changed name on function list_create to check_list_create to avoid
+ name clash.
+ * src/check.c, src/check.h.in, tests/check_check_master.c,
+ tests/check_check_sub.c: Applied ANSI C99 patch (#1047014)
I got few private comments that one should not use buildlink3.mk but the
standard DEPENDS statement because we do not link against libraries
in this case.
From the Changelog:
- 0006273: [security] File Inclusion Vulnerability (vboctor)
- 0006275: [security] SQL injection (vboctor)
- 0006234: [filters] Filter sometimes returns no results (thraxisp)
- 0006295: [filters] Old filters and view_state problems. (thraxisp)
- 0006288: [filters] Patch against CVS HEAD for Saved filter problem with view_state (thraxisp)
- 0006296: [filters] Filter sql includes unnecessary links to custom_field_string_table for date custom fields (thraxisp)
- 0006297: [filters] sorting on custom field, bring MySQL to deadlock loop (thraxisp)
Noteworthy changes in version `1.0.0'
=========================================
* All inspectors are now modeled in .gorm files.
* Added autosizing to form attributes inspector.
* Utilize and maintain parent/child data structure more pervasively
* Reorganized code in palettes for cleaner implementation.
* Removed code to check for user bundles, since bugs in Camaelon
which prompted those changes were fixed long ago.
* Added documentation to GormCore
