Get Gentoo!
gentoo.org sites
gentoo.org Wiki Bugs Forums Packages
Planet Archives Sources
Infra Status
summaryrefslogtreecommitdiff
path: root/eclass
Commit message (Collapse)AuthorAgeFilesLines
* cmake-utils.eclass: enable CMAKE_WARN_UNUSED_CLI for EAPI 6Maciej Mrozowski2016-05-181-2/+2
|
* l10n.eclass: l10n_find_plocales_changes: make it more friendlyMike Frysinger2016-05-161-5/+8
| | | | | | - Use ebegin/eend to show process messages instead of two einfos. - Fix sorting logic to work in all locales. - Return 1 when things are out of sync so callers can abort if they want.
* eclass/gnat: fix misspelling of 'virtual'Brian Norris2016-05-161-1/+1
| | | | | | Closes: https://github.com/gentoo/gentoo/pull/1466 Signed-off-by: Brian Norris <computersforpeace@gmail.com> Signed-off-by: Patrice Clement <monsieurp@gentoo.org>
* kde5.eclass: Update SRC_URI for kde-frameworks/krunnerJohannes Huber2016-05-151-2/+1
|
* python-utils-r1.eclass: python_wrapper_setup, fix pythonN-config cleanupMichał Górny2016-05-151-1/+1
|
* java-utils-2.eclass: DESTTREE querying -> /usr, #566480Michał Górny2016-05-141-3/+3
|
* java-utils-2.eclass: DESTTREE assignment -> into, #566480Michał Górny2016-05-141-1/+4
|
* java-utils-2.eclass: Use insinto & insopts instead of vars, #566480Michał Górny2016-05-141-9/+17
| | | | | Replace INSDESTTREE and INSOPTIONS internal variable references with correct calls to insinto & insopts, using a subshell to scope them.
* python-utils-r1.eclass: Make python_{script,module}root vars privateMichał Górny2016-05-141-36/+24
| | | | | | | Remove python_scriptroot & python_moduleroot from the public eclass API, in favor of calls to python_scriptinto and python_moduleinto. This feature was seldom used, not listed in the official API guide, and the *into functions is more in line with the PMS.
* python-utils-r1.eclass: stop respecting DESTTREE in python_do*/new*Michał Górny2016-05-141-3/+3
| | | | | | | | | | Replace the code respecting current value of ${DESTTREE} in favor of using /usr unconditionally. The feature was extremely unlikely to be ever used (only if you did 'into ...; python_do*', quite confusing and relied on semi-internal package manager variables. Furthermore, it was potentially broken as it relied on DESTTREE having an initial '/usr' value which is not warranted by the PMS.
* distutils-r1.eclass: Do not rely on INSDESTTREE, #173630Michał Górny2016-05-141-4/+6
|
* eclass: ELT-patches: update cross patches for libtool-2.4+ #572038Mike Frysinger2016-05-133-4/+45
|
* toolchain.eclass: prune more unused libtool archives #573302Mike Frysinger2016-05-111-2/+38
|
* mysql-multilib-r1.eclass: Remove private header and don't install tests wrt ↵Brian Evans2016-05-111-6/+15
| | | | | | | | | | bug 582738 The private include file mysql/private/config.h is not needed by anything and should not be referenced. It will never match in multiple ABIs so just remove it. When USE=-test, don't install tests which would just be removed later anyway.
* sys-devel/gcc: add USE=mpx support for Intel MPX #578396Mike Frysinger2016-05-111-1/+5
|
* eclass/latex-package.eclass: set VARTEXFONTS inside the sandbox when ↵Alexis Ballier2016-05-111-0/+4
| | | | | | possibly compiling latex files. Should help in bugs like #580912
* eclass/ros-catkin.eclass: Add support for nodejs messages.Alexis Ballier2016-05-111-6/+13
|
* multilib.eclass (get_libname): Cygwin allows versioning symlinksMichael Haubenwallner2016-05-091-1/+1
| | | | | | Although useless on Cygwin, some build systems or ebuilds still prefer to create the versioning symlinks. As symlinked import libs do not harm anything on Cygwin, just leave those build systems or ebuilds alone.
* sys-devel/gcc: add USE=vtv support for vtable verification #547040Mike Frysinger2016-05-081-1/+8
|
* l10n.eclass: Bump copyright yearJames Le Cuirot2016-05-081-1/+1
|
* l10n.eclass: Sort and normalize PLOCALES in l10n_find_plocales_changeJames Le Cuirot2016-05-081-1/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | l10n_find_plocales_change assumes that PLOCALES is sorted alphanumerically with a single space between each entry and no surrounding whitespace. This is not a bad assumption but it isn't documented and it's inconvenient in at least one particular case. MakeMKV uses non-standard locale names and I intend to map these using an associative array, which is possible as of EAPI 6. This allows me to do the following, though only with the above change as associative arrays are not ordered. declare -A MY_LOCALES MY_LOCALES=( [zh]=chi [da]=dan … ) PLOCALES="${!MY_LOCALES[@]}" inherit l10n src_prepare() { PLOCALES="${MY_LOCALES[@]}" l10n_find_plocales_changes … } src_install() { for locale in $(l10n_get_locales); do doins makemkv_${MY_LOCALES[${locale}]}.mo.gz done } Fixes bug #513242.
* Revert "fixed einstall vs. emake install for eapi=6"Ryan Hill2016-05-071-5/+1
| | | | This reverts commit 66afcab271f65b97330e610040ad3acc1b812a03.
* fixed einstall vs. emake install for eapi=6Joerg Bornkessel2016-05-071-1/+5
|
* EAPI=6 support; removed support befor media-video/vdr-2Joerg Bornkessel2016-05-071-27/+20
|
* vim-spell.eclass: drop epause usage (bug #581762)Tim Harder2016-05-071-1/+0
|
* ruby-single.eclass: Support EAPI=6Hans de Graaff2016-05-061-2/+2
| | | | | The current implementation is compatible with EAPI=6. Thanks to Poly-C for additional testing.
* font-ebdftopcf.eclass: Move global-scope use calls to phase, #566540Michał Górny2016-05-051-3/+3
|
* kde4-meta.eclass: Support tar.gz snapshot from gitAndreas Sturmlechner2016-05-021-1/+7
| | | | Signed-off-by: Johannes Huber <johu@gentoo.org>
* kde5.eclass: improve handling of nls USE flagMichael Palimaka2016-05-011-2/+2
| | | | Linguas was being removed unconditionally on packages without an nls USE flag.
* eclass/ros-catkin.eclass: use emake tests with nonfatal as nonfatal is still ↵Alexis Ballier2016-04-301-1/+3
| | | | fatal with cmake-utils_src_make and breaks e.g. dev-ros/rviz.
* www-client/firefox: version bump to 46.0Ian Stakenvicius2016-04-281-0/+404
| | | | | | | | | | Cairo-gtk3 is now the upstream default in this version of firefox, and so the USE flag was switched to make cairo-gtk2 optionally selectable via a "force-gtk2" flag. Bug: http://bugs.gentoo.org/555620 Package-Manager: portage-2.2.26
* firefox-45, thunderbird-45: support system harfbuzz and graphite2Ian Stakenvicius2016-04-281-5/+59
| | | | | | | | | | | | This commit introduces changes to mozconfig-v6.45.eclass and related changes to the existing mail-client/thunderbird-45.0 and the new version bump www-client/firefox-45.1.0 to support using system copies of media-libs/harfbuzz and media-gfx/graphite2 instead of the copies bundled inside of the mozilla codebase. www-client/firefox-45.1.0 was additionally bumped to support security bug 581326 Package-Manager: portage-2.2.26
* eclass/games-mods.eclass: use #!/sbin/openrc-run instead of #!/sbin/runscriptAustin English2016-04-251-1/+1
|
* kde5.eclass: Spell iuse correctly.Ryan Hill2016-04-241-1/+1
|
* ssl-cert.eclass: guard against EAPI=0Anthony G. Basile2016-04-231-0/+12
|
* kde5.eclass: Drop translations when nls not wantedJohannes Huber2016-04-231-0/+5
| | | | Gentoo-bug: 580720
* Drop all refrences to eclipse in intel compiler productsJustin Lecher2016-04-221-38/+0
| | | | Signed-off-by: Justin Lecher <jlec@gentoo.org>
* HTTPS for *.freedesktop.orgAnthony Ryan2016-04-203-5/+5
| | | | | | | | | | | | | | | | | | | | | | | | | * Excluding xorg.freedesktop.org * Excluding tango.freedesktop.org The following modified ebuilds were found to have problems after modification but the problems were determined to not be a regression. Upstream tarball has a new hash: * app-misc/evtest-1.29 * app-misc/evtest-1.30 * dev-embedded/scratchbox2-2.0-r1 * dev-ml/cairo-ocaml-1.2.0 * net-libs/libqmi-1.0.0 * sys-auth/libfprint-0.4.0 * sys-auth/libfprint-0.5.0 * sys-auth/libfprint-0.5.1 Upstream tarball has been deleted: * dev-libs/liblazy-0.2 * dev-util/pkgconfig-openbsd-20130507-r1 * x11-libs/xvba-video-0.8.0-r3 * x11-misc/driconf-0.9.1-r1 * x11-misc/xdg-utils-1.1.0_rc2-r1
* kde5.eclass: don't run pkg_pretend for kde-frameworksMichael Palimaka2016-04-211-1/+5
| | | | | | | | Merely exporting the phase function is very expensive, even if nothing useful is done inside it. Frameworks are supposed to support GCC 4.5 at a minimum which is old enough to make it safe enough to omit the check.
* kde5.eclass: merge type is already checked in _check_gcc_versionMichael Palimaka2016-04-211-6/+2
|
* kde5.eclass: drop minimal gcc requirement for kde-frameworksMichael Palimaka2016-04-211-1/+6
| | | | | This will be used in a future commit for reducing the number of time-consuming gcc version checks.
* kde4-base.eclass: Cleanup SRC_URIsAndreas Sturmlechner2016-04-211-11/+2
|
* kde5-functions.eclass: punt_bogus_deps improvementsJohannes Huber2016-04-191-1/+1
| | | | | Syncs eclass with kde overlay. Improvements by Michael Palimaka <kensington@gentoo.org>. Needed for bug #580282.
* java-*.eclass: Bump java-config dependency to avoid bug #580504James Le Cuirot2016-04-193-3/+3
| | | | | People may remove java-config-wrapper without updating java-config, leading to a broken setup.
* ssl-cert.eclass: add libressl supportAnthony G. Basile2016-04-181-4/+8
|
* perl-module.eclass: Fix bash array handling of DIST_EXAMPLESAndreas K. Hüttel2016-04-181-1/+1
|
* qt4-build-multilib.eclass: prepare configure for gcc-6 and beyondDavide Pesavento2016-04-171-0/+4
|
* ssl-cert.eclass: remove = slot operator from DEPENDAnthony G. Basile2016-04-161-2/+2
| | | | | | | | It is common practice to include := in DEPENDs even though slot operators are only meaningful in RDEPENDs because one often sees RDEPEND="${DEPEND} ...". However, the scope of DEPEND does not cross the boundary between eclass and inhering ebuild, so that operator is useless here.
* Revert "ssl-cert.eclass: add libressl support"Anthony G. Basile2016-04-161-8/+4
| | | | This reverts commit ad0c2ab2bdbd34f4550e49c56cfd5974d6a2c07a.
* ssl-cert.eclass: add libressl supportAnthony G. Basile2016-04-161-4/+8
|