summaryrefslogtreecommitdiff
path: root/eclass
Commit message (Collapse)AuthorAgeFilesLines
* Fix CC, CXX, LD and LINK in all test files, bug 321335.Dror Levin2010-05-241-3/+6
|
* fixed 4.7.0_beta1 regression with -fvisibility=hiddenAlex Alexander2010-05-241-1/+8
|
* Use the new syntax for rake dependencies.Diego Elio Pettenò2010-05-241-5/+5
|
* Avoid touching IUSE in _ruby_wrap_conditions; set the test USE flagDiego Elio Pettenò2010-05-241-3/+3
| | | | | directly into the ruby_add_rdepend function to avoid subshell handling.
* run automake if the configure script requests it via AM_INIT_AUTOMAKE even ↵Mike Frysinger2010-05-231-10/+16
| | | | if there are no .am files #311439
* add new tc-has-openmp helper func #320215 by Samuli SuominenMike Frysinger2010-05-231-1/+25
|
* inform the users that building qt-webkit with -ggdb can be problematic.Alex Alexander2010-05-221-1/+13
| | | | bug 307861
* fix bug 320851.Alex Alexander2010-05-221-20/+12
| | | | | | also apply ayoy's fixes: - Remove changes to mkspecs in favor of setting correct env variables in configure script. - Make compilation output more verbose.
* Set RDEPEND="${RDEPEND}" to avoid depend to filter through rdepend.Diego Elio Pettenò2010-05-221-1/+2
|
* Avoid globbing while parsing the dependency string.Diego Elio Pettenò2010-05-221-1/+3
|
* Use a slightly different syntax for expliciting implementation-dependent ↵Diego Elio Pettenò2010-05-221-9/+15
| | | | | | | | | | | | | | dependencies. Rather than using a single, huge USE-conditional per atom, put the dependendencies under conditional _and_ USE-dependency. This allows for syntax like: USE_RUBY="ruby18 ruby19" ... ruby_add_bdepend "test? ( dev-ruby/first )" USE_RUBY="ruby18" ruby_add_bdepend "test? ( dev-ruby/second )" USE_RUBY="ruby19" ruby_add_bdepend "test? ( dev-ruby/third )"
* Allow for standard depend syntax in the single parameter form of ↵Diego Elio Pettenò2010-05-221-59/+49
| | | | | | | | | | ruby_add_rdepend and ruby_add_bdepend. Make the two parameters form throw a warning for gentoo developers only (for now), and call back the single-parameter function after wrapping. Update documentation to only suggest using the new syntax. This allows dropping _ruby_add_rdepend and _ruby_add_bdepend convenience functions, and rather adds a _ruby_wrap_conditions function.
* Fix missing link warnings in the eclass to print the library's soname again.Diego Elio Pettenò2010-05-221-3/+3
|
* Change processing of atoms in ruby_add_rdepend and ruby_add_bdepend toDiego Elio Pettenò2010-05-211-7/+16
| | | | | | | | | | | | make it possible to process complex dependency strings, such as || ( virtual/ruby-test-unit dev-ruby/test-unit:2 ) which is going to be common for test-unit-2 compatible packages; to have a proper dependency tree. This, theorically, can also drop the need for the 2-ary calls for ruby_add_bdepend and ruby_add_rdepend.
* Solve bug #320545 in the least hacky way: disable --no-undefined forDiego Elio Pettenò2010-05-211-3/+10
| | | | | | | | | the Ruby-Gnome2 bindings. Unfortunately they rely on a partially-undefined behaviour, let's just hope they'll never come giving us trouble for this. Also avoid using cp -a (non-portable) for installing samples, use insinto/doins instead.
* Bug #317297: Some kernel sources packages in overlays have version numbers ↵Robin H. Johnson2010-05-201-2/+14
| | | | like -9999 that does not allow us to work out what the actual version number is for functions like kernel_is. They SHOULD have a CKV variable in this case, but without it, they previously went into an infinite loop during the version detection in the eclass.
* Change ELT-patches for AIX to use import files as soname-emulator (bug#213277).Michael Haubenwallner2010-05-2044-194/+274
| | | | | This requires switching libtool to force -no-undefined behaviour on AIX. Additionally, get_libname() now does return .so and versions like Linux.
* drop unnecessary output padding in unpack_makeself #320313Mike Frysinger2010-05-201-5/+5
|
* Fix handling of PYTHON_USE_WITH with more than 2 USE flags.Arfrever Frehtes Taifersar Arahesis2010-05-171-307/+476
| | | | | | | | | | | | Improve python_pkg_setup(). Rename python_clean_sitedirs() to python_clean_installation_image(). Improve python_clean_installation_image(). Improve handling of -- option in phase functions. Add check for dev-lang/python in validate_PYTHON_ABIS() (bug #319651). Improve error message in python_copy_sources(). Improve python_set_active_version(). Improve error messages in PYTHON(). Support PEP 3147 in python_mod_optimize(), python_mod_cleanup() and python_mod_compile().
* Improve automatic misc documentation files installation (README's, ↵Maciej Mrozowski2010-05-172-17/+10
| | | | ChangeLogs, etc)
* Update qimageblitz categoryMaciej Mrozowski2010-05-151-2/+2
|
* Sync eclasses from kde overlay.Maciej Mrozowski2010-05-153-165/+246
| | | | | | | | - kdevelop ebuils support - cmake-2.6.4 needed for KDE SC 4.5 - crosscompilation fixes - fix README's installation - pre-release koffice support
* Added dev-util/automoc dependency for all KDE4 packagesMaciej Mrozowski2010-05-151-1/+2
|
* Document for eclass-manpages.Ryan Hill2010-05-151-18/+31
|
* Improve documentation.Ryan Hill2010-05-151-21/+14
|
* sgml-catalog.eclass: Add EAPI3/Gentoo Prefix support. bug 310777Jeremy Olexa2010-05-131-15/+20
|
* Bug #318851: USE=cluster also needs the partition engine support in 5.1.46.Robin H. Johnson2010-05-131-3/+4
|
* Replace grep -P with more portable sed -ne equivalentMaciej Mrozowski2010-05-121-2/+2
|
* Fix db-use.eclass to work with sys-libs/db-5.0.Diego Elio Pettenò2010-05-121-2/+6
|
* add -m1 to grep for unpack_makeself for bug #319059Michael Sterrett2010-05-111-2/+2
|
* Bug #318389: check for each examples directory before trying to use it. Also ↵Robin H. Johnson2010-05-111-4/+9
| | | | avoid QA warning on "use java" on ebuilds where java was not available.
* sys-libs/db-5 has changed some of the test scripts slightly, so we need to ↵Robin H. Johnson2010-05-111-4/+11
| | | | look a little wider for parallel.tcl. Also improve the regex we run on it for relative directories.
* Improve binary renaming to NOT rename all elements in the path, and also ↵Robin H. Johnson2010-05-111-4/+7
| | | | pick up binaries that match the prefix of 'db' but not 'db_'.
* change bug notice to report a bug rather than annoy specific peopleMike Frysinger2010-05-111-3/+3
|
* Add helpful comment about where the MY_* vars are defined.Jeremy Olexa2010-05-091-1/+3
|
* Simplified the eclass, since the rules for desklet and control submission ↵Joe Sapp2010-05-081-149/+116
| | | | have become much more standardized recently (http://gdesklets.de/index.php?q=node/2). This also means that we will no longer support sensors and sensor-based desklets. No more desklets that do not heed these rules will be put into portage.
* Tried to fix bug #314653Joe Sapp2010-05-081-8/+2
|
* Add 4.5 -mno-* flags to exception list.Ryan Hill2010-05-081-1/+4
|
* Fix db_src_install_usrbinslot to proper SLOT again (#318367) and add a dieTorsten Veller2010-05-041-2/+3
|
* Bug #313637: Fix versioning of libdb_stl.a. Also clean up the eclass for ↵Robin H. Johnson2010-05-041-24/+28
| | | | quoting and a slight refactor.
* Bug #275652: Install examples with USE=examples.Robin H. Johnson2010-05-031-2/+16
|
* Bug #253312: implement parallel tests for sys-libs/db.Robin H. Johnson2010-05-031-5/+17
|
* Bug #263797: Do not run the sys-libs/db testsuite as root.Robin H. Johnson2010-05-031-1/+6
|
* Use PYTHON_USE_WITH_OPT to only pull in python when actually required. Bug ↵Jim Ramsay2010-05-031-1/+2
| | | | #317939
* Fix typo.Ulrich Müller2010-05-021-2/+2
|
* Bug #318095: Improve K_SECURITY_UNSUPPORTED messages for deblobbing. Based a ↵Robin H. Johnson2010-05-021-4/+9
| | | | patch by Ulrich Müller <ulm@gentoo.org>.
* Don't print errors during unpack because the directory does not exist yet.Diego Elio Pettenò2010-05-011-2/+2
|
* Change variable name (not yet used): CMAKE_MIN_VER -> CMAKE_MIN_VERSIONJonathan Callen2010-04-301-4/+4
|
* Allow changing the minimum required cmake versionJonathan Callen2010-04-301-2/+8
|
* Create a special handling of S variable for github-based packages.Diego Elio Pettenò2010-04-301-2/+11
| | | | | | Since using GitHub is widely common for Ruby packages, especially those not packaging the tests within the gem, we would be reimplementing this so many times that it's worth special casing here.