summaryrefslogtreecommitdiff
path: root/eclass
Commit message (Collapse)AuthorAgeFilesLines
* Keep a mostly-internal variable of the latest version of the deblob tool ↵Robin H. Johnson2010-04-261-5/+11
| | | | releases by the upstream libre-sources guys, this saves us updating K_DEBLOB_AVAILABLE quite so often, but we need to remember to redigest sources ebuilds when we DO change the this variable instead.
* Document the new deblob control variables, after renaming them to match the ↵Robin H. Johnson2010-04-261-7/+14
| | | | existing pattern.
* Bug #266157: Deblob support for libre kernels.Robin H. Johnson2010-04-261-3/+46
|
* Refactor debug-print-kernel2-variables to make it faster to add new stuff.Robin H. Johnson2010-04-261-16/+6
|
* Add more variables to the debug output list when it does fire.Robin H. Johnson2010-04-261-1/+4
|
* While working on the libre-sources bug, I noticed that some of the kernel ↵Robin H. Johnson2010-04-261-2/+9
| | | | sources ebuilds out there have OKV set, but NOT KV_*, leading kernel_is to always return false. Check all of the variables we need for safety instead.
* Add support for hardfloat toolchains, acked by vapierRaúl Porcel2010-04-251-1/+2
|
* Add in lto use flag, and cleaning up of python scripts for gcc-4.5Mark Loeser2010-04-241-1/+16
|
* Add EROOT (EAPI3) support to vim-doc & vim-spell eclassesJeremy Olexa2010-04-232-5/+7
|
* Use hardfloat only if hardfloat is trueRaúl Porcel2010-04-231-3/+3
|
* Add tc-is-hardfloat function, rename 'no' of tc-is-softfloat to 'target ↵Raúl Porcel2010-04-221-2/+15
| | | | doesn't support softfloat', acked by vapier, bug #315987
* Drop epatch_user, already executed by base eclass. Thx to sedzimir.Tomas Chvatal2010-04-221-2/+1
|
* Add PERL_EXPORT_PHASE_FUNCTIONS and move perl-helper functions back to ↵Torsten Veller2010-04-222-132/+147
| | | | perl-module
* Enable hardvfp if *eabi, if *hardfloat*, if gcc-4.5 and if arm, bug #315987, ↵Raúl Porcel2010-04-201-1/+7
| | | | acked by vapier
* Add new function font_cleanup_dirs, called in pkg_postrm, to clean up font ↵Ryan Hill2010-04-201-1/+36
| | | | directories that only contain generated files (bug #315369). Based on code taken from xorg-2.eclass. Note that due to portage caching pkg_postrm, currently installed font packages will not trigger this code until the second time they are reinstalled.
* handle yet newer style of gold+bfd parallel installs configure option ↵Mike Frysinger2010-04-191-2/+4
| | | | #315995 by Johannes Nixdorf
* make_desktop_entry: use printf %b to make arg passing simpler since we only ↵Mike Frysinger2010-04-191-2/+2
| | | | want to support embedded escape sequences
* use mozconfig_use_enable where appropriateJory Pratt2010-04-181-11/+4
|
* Add perl-helper.eclass (#302183) and update perl-module.eclass per ↵Torsten Veller2010-04-172-96/+179
| | | | <http://archives.gentoo.org/gentoo-dev/msg_82ff2383bd5836dda541d0b2b4685ec8.xml>. Add support for EAPI=3 (#310513). Removes "AppleDouble encoded Macintosh files" (#268497)
* Add quotes around variables that could contain spaces in paths. Fixes bug ↵Petteri Räty2010-04-171-5/+5
| | | | #315961. Thanks to Michał Górny <gentoo@mgorny.alt.pl>.
* Apply interix *vim workaround to eclass, bug 310991Jeremy Olexa2010-04-151-2/+11
|
* Fix eqmake4 when passed args with spacesJonathan Callen2010-04-151-4/+4
|
* drop spurious arg to popd #314493Mike Frysinger2010-04-141-2/+2
|
* rewrite ELT_walk_patches to handle path name as well as locale issues ↵Mike Frysinger2010-04-141-30/+25
| | | | pointed out by Michał Górny #314493
* Use EPATCH_FORCE=yes. Otherwise patches as dir are not populated.Tomas Chvatal2010-04-121-1/+2
|
* Display compiler and linker flags - a lousy compromise between excessive ↵Maciej Mrozowski2010-04-101-12/+18
| | | | verbosity and hiding relevant information
* Drop accessibility requirement on qt. Per bug #309677Tomas Chvatal2010-04-071-3/+3
|
* Due to portage's environment filtering, variables that start with BASH_ ↵Jeremy Olexa2010-04-072-10/+10
| | | | cannot be used. As such, s/BASH_COMPLETION_NAME/BASHCOMPLETION_NAME/g for bug 312671
* Use split-ebuild patches from one tarballMaciej Mrozowski2010-04-061-21/+23
|
* Add ruby_get_version() and ruby_get_implementation() convenience helper ↵Alex Legler2010-04-051-1/+27
| | | | functions
* Also remove x and opengl from the plugin list. None of the in-tree -bad ↵Mart Raudsepp2010-04-051-4/+2
| | | | package versions have these either. There are X plugins in -base and -good, and a separate opengl related tarball in existence for these things by now.
* Remove old plugin entries that are not present in any -bad packages ↵Mart Raudsepp2010-04-051-17/+4
| | | | remaining in tree anymore. These are safe to remove, as this list is only used during package compilation, never during unmerge. This reduces the amount of QA warnings from unrecognized configure options and just cleans up the list for maintainability. Update comments accordingly.
* Prepare for gst-plugins-bad-0.10.18 - has new flite split plugin and ivorbis ↵Mart Raudsepp2010-04-051-4/+4
| | | | removed
* Fix the plugins list: stop automagically building zbar, cog and rsvg split ↵Mart Raudsepp2010-04-051-7/+14
| | | | plugins as part of the non-split package due to them being forgotten to be added here before (add them now), cog/zbar is not packaged in-tree so not a big problem, but rsvg is; it will be a separate split package. Add missing Windows/OSX sys plugins to the list to avoid useless header/linkage checks that waste configure time for all -bad splits too - directsound, directdraw, osx_video, wininet, acm. Order the list per ./configure --help output. Add comments to old plugins that are not shipped by -bad anymore, for when they were removed, to remove them from the list once safe. Have not tracked down opengl and x removal yet though.
* Add -0 to xargs since input is NULL delimited. Thanks to Dror for catching.Ryan Hill2010-04-031-2/+2
|
* Add deprecation notices to the old ruby eclasses.Hans de Graaff2010-04-033-3/+15
|
* Fully validate and clean up plugin list for current versions. Reorder them ↵Mart Raudsepp2010-04-031-6/+10
| | | | per configure --help output, in order of global/sys/ext switches. Add directsound, osx_audio and osx_video to avoid useless header/linking configure checks for all -good plugins and save some configure time. Remove ladspa and cdio from the list, they have not been part of -good since 0.10.8 and 0.10.10 respectively, so safe to remove and avoid the unrecognized configure option QA warnings.
* ruby-ng.eclass: Add support for targets with different flag and executable ↵Alex Legler2010-04-021-2/+11
| | | | names. Implement this feature for Ruby Enterprise Edition 1.8.
* Allow changing store directory, update maintainers.Krzysztof Pawlik2010-04-021-17/+22
|
* Add append-libs() functionJonathan Callen2010-04-021-1/+16
|
* Allow JAVA_PKG_NV_DEPEND override in java-pkg_needs-vm()Vlastimil Babka2010-04-011-2/+2
|
* Bug #312687: We actually need to consider the automake from the real root, ↵Robin H. Johnson2010-04-011-2/+4
| | | | not the ROOT destination for stage 1 builds.
* Bug #266157: Until such time as we have full deblob support for all sources, ↵Robin H. Johnson2010-04-011-2/+4
| | | | mark that kernels contain firmware blobs that are covered under the freedist license, not GPL-2.
* Followup on bug #284946, #293801, #310615, #307251: In 5.0.87, the mutex ↵Robin H. Johnson2010-04-011-4/+10
| | | | locking code was changed a lot again, so that while 5.0.84 required a newer GCC due to the HAVE_ATOMIC_BUILTINS error, newer releases do NOT need those GCC versions anymore. This also frees up ~hppa to do more stabilization.
* Update to better deal with python3 and multiple python ABIsJim Ramsay2010-04-011-4/+24
| | | | | | Basically we just outright assert that all rox-base/* and rox-extra/* python-based apps will only work with python2. We'll be back to address this again if any of them ever become python3-aware.
* Try to unbreak revision 1.95, at least guessing what Robin tried to do.Diego Elio Pettenò2010-04-011-11/+14
|
* use peardev instead of pear to avoid memory limits (bug #265413)Matti Bickel2010-03-311-5/+5
|
* Bug #312075: The ~ was getting expanded to the optional stuff got duplicated ↵Robin H. Johnson2010-03-311-5/+4
| | | | and that sucked, giving a bad warning sometimes.
* Bug #312315: We must reflect the latest stable automake SLOT for all ↵Robin H. Johnson2010-03-311-3/+24
| | | | architectures, otherwise we get fun circular build dependencies when we already have a should-be-sufficent automake and try to bring in a different one during the stage2->3 transition.
* When using dummy haskell libraries, there is no test suite.Lennart Kolmodin2010-03-301-3/+7
|