summaryrefslogtreecommitdiff
path: root/eclass
Commit message (Collapse)AuthorAgeFilesLines
* eclass/tests: source tests-common.sh || exitMike Gilbert2022-02-1427-27/+27
| | | | Signed-off-by: Mike Gilbert <floppym@gentoo.org>
* eclass/tests/savedconfig.sh: abort when source tests-common.sh failsMike Gilbert2022-02-141-1/+1
| | | | | Closes: https://bugs.gentoo.org/833342 Signed-off-by: Mike Gilbert <floppym@gentoo.org>
* llvm.eclass: enable EAPI 8Stephan Hartmann2022-02-141-2/+2
| | | | Signed-off-by: Stephan Hartmann <sultan@gentoo.org>
* emboss-r2.eclass: last-riteDavid Seifert2022-02-141-1/+4
| | | | | Closes: https://github.com/gentoo/gentoo/pull/24178 Signed-off-by: David Seifert <soap@gentoo.org>
* emboss-r3.eclass: revbump for EAPI 8David Seifert2022-02-141-0/+131
| | | | Signed-off-by: David Seifert <soap@gentoo.org>
* eclass/go-module.eclass: documentation and bdepend adjustmentWilliam Hubbs2022-02-131-6/+8
| | | | | | | | Update the documentation to not refer to golang* eclasses. Bump the go dependency to 1.16 so the behavior for modules is consistent. Signed-off-by: William Hubbs <williamh@gentoo.org>
* eclass/golang-*: fix deprecation noticesWilliam Hubbs2022-02-134-4/+4
| | | | Signed-off-by: William Hubbs <williamh@gentoo.org>
* eclass/dune.eclass: inherit multiprocessingMaciej Barć2022-02-131-3/+5
| | | | | | | in order to run up to makeopts_jobs number of jobs in dune_src_compile and dune_src_test Signed-off-by: Maciej Barć <xgqt@gentoo.org>
* tree-sitter-grammar.eclass: Re-assign maintainerMatthew Smith2022-02-121-3/+3
| | | | Signed-off-by: Matthew Smith <matthew@gentoo.org>
* linux-mod.eclass: Fix EAPI 8 supportMike Pagano2022-02-121-1/+5
| | | | Signed-off-by: Mike Pagano <mpagano@gentoo.org>
* linux-mod.eclass: Remove eutils inherit for EAPI 8Mike Pagano2022-02-121-1/+1
| | | | Signed-off-by: Mike Pagano <mpagano@gentoo.org>
* linux-mod.eclass: enable EAPI 8, remove EAPI 5Mike Pagano2022-02-121-7/+2
| | | | | | Clean-up code needed for deprecated EAPIs (EAPI < 4) Signed-off-by: Mike Pagano <mpagano@gentoo.org>
* llvm.org.eclass: Add LoongArch targetMichał Górny2022-02-101-1/+8
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Silence missing locale(1) errorMichał Górny2022-02-091-1/+1
| | | | | | | | | The python_export_utf8_locale() function checks for missing locale(1) and handles it gracefully. However, it does not redirect stderr probably, so systems without the executable (e.g. musl) get unintended console output. Redirect it to /dev/null. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Add min version to tomli depMichał Górny2022-02-091-1/+1
| | | | | Closes: https://bugs.gentoo.org/832782 Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Add QA check for obsolete PYTHON_COMPATMichał Górny2022-02-091-1/+17
| | | | | | | | Add a QA check that reports obsolete implementation in PYTHON_COMPAT if ebuild has been modified in 2022 (based on copyright year). Requested by sam. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-any-r1.eclass: Do not test EPYTHON twiceMichał Górny2022-02-091-6/+8
| | | | | | | | Avoid checking the implementation from EPYTHON via fallback iteration. While there's no technical harm in doing that, now that we output verbosely the users will notice ;-). Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Support matching impls by stdlib versionMichał Górny2022-02-094-44/+36
| | | | | | | | | | | | | | Update _python_impl_matches() (used to implement python_gen*, python_setup) to support specifying stdlib versions ("3.8", "3.9") in addition to exact implementation names. This makes handling PyPy3 version changes much easier when dealing with backports. For example, if you specify "3.8", then the spec will match python3_8 and pypy3, for as long as we supply PyPy3.8. Once we upgrade to PyPy3.9 completely, it will stop matching pypy3 and we won't have to manually keep updating these deps. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Add status messages to python_optimizeMichał Górny2022-02-091-0/+1
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Remove old phase check from python_optimizeMichał Górny2022-02-091-9/+0
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Use heredoc instead of "python -c"Michał Górny2022-02-091-5/+36
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Use heredoc instead of "python -c"Michał Górny2022-02-091-10/+22
| | | | | | | Use heredocs instead of inlining longish scripts in "python -c", for greater readability. Thanks to arthurzam for the suggestion. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-single-r1.eclass: Remove PYTHON_MULTI_USEDEPMichał Górny2022-02-091-7/+0
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Remove python_is_python3Michał Górny2022-02-093-27/+3
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-r1.eclass: Improve comment for USE-dep generationMichał Górny2022-02-091-3/+9
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-single-r1.eclass: Inline & simplify USE-deps in gen_cond_depMichał Górny2022-02-091-42/+3
| | | | | | | | | | | Generate simpler USE-deps inline in python_gen_cond_dep() and remove _python_gen_usedep(). The original code always repeated USE deps on all targets to aid Portage in giving better suggestions. However, since there always will be exactly one implementation selected, this is unnecessary and we can just have every cond-dep match exactly that one target. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Remove python_wrapper_setupMichał Górny2022-02-091-16/+0
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Remove deprecated python_exportMichał Górny2022-02-091-17/+0
| | | | | | | Remove python_export, as it is no longer used by any ebuilds in ::gentoo. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Fix sphinx_build for non-autodoc caseMichał Górny2022-02-091-4/+18
| | | | | | | | | | | | | | | | | | | Fix the regression in calling sphinx-build for the non-autodoc case that causes the build to fail if dev-python/sphinx isn't built for the newest Python interpreter available. To account for this, we need to call sphinx-build as an executable (i.e. via python-exec). Ideally, build_sphinx would be aware of which case it is used for, and use appropriate invocation. Unfortunately, we cannot do that without breaking backwards compatibility. However, we can simply check if Sphinx is available via ${EPYTHON}, and fall back to calling python-exec directly. This is effectively equivalent to choosing the specific invocation directly, as python-exec would have respected the implementation specified by EPYTHON anyway if sphinx-build executable was available for it. Fixes: f6a17acb8b7c (...: Run sphinx-build via EPYTHON) Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Fix has_version for distutils_enable_sphinxMichał Górny2022-02-091-1/+5
| | | | | | | Fix the has_version calls for distutils_enable_sphinx to use -b option (--host-root in earlier EAPIs). Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Inline python_is_installedMichał Górny2022-02-091-19/+3
| | | | | | | | Inline the python_is_installed function that is used exactly once (in _python_run_check_deps). This helps us avoid having to grab PYTHON_PKG_DEP twice. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Report _python_run_check_deps verboselyMichał Górny2022-02-091-1/+9
| | | | | | | Report dep checking progress verbosely, to help users understand why a particular implementations was rejected or selected. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-any-r1.eclass: Explain the reason for interpreter choiceMichał Górny2022-02-091-3/+3
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-utils-r1.eclass: Add function to run python_check_deps()Michał Górny2022-02-093-34/+25
| | | | | | | Add a function encompassing the common logic to run python_check_deps() from python-any-r1 and python-r1. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-any-r1.eclass: Move EPYTHON validity check to python_setup()Michał Górny2022-02-091-22/+14
| | | | | | | | | Move the EPYTHON validity check from _python_EPYTHON_supported() to python_setup() where it belongs. This avoids unnecessarily retesting implementations taken from PYTHON_COMPAT and paves the way towards moving the common logic to python-utils-r1. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* python-r1.eclass: Remove deprecated python_gen_usedepMichał Górny2022-02-091-42/+1
| | | | | | | This function is deprecated for some time already and there are no more consumers left in ::gentoo. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Create distutils_pep517_install helperMichał Górny2022-02-091-29/+48
| | | | | | Split the wheel build & install logic into a a new helper. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Get wheel name from the backendMichał Górny2022-02-091-9/+5
| | | | | | | Use the wheel name returned by build_wheel() rather than trying to guess it from WHEEL_DIR. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Split backend getting code into a functionMichał Górny2022-02-091-47/+63
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* bazel.eclass: Add SUPPORTED_EAPIS 8 docJason Zaman2022-02-061-1/+1
| | | | Signed-off-by: Jason Zaman <perfinion@gentoo.org>
* llvm.eclass: Allow slot 15Michał Górny2022-02-061-2/+2
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* llvm.org.eclass: Bump main branch to 15.xMichał Górny2022-02-061-1/+1
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* eclass/ruby-fakegem.eclass: updated extensions dir according to formula from ↵Andrew Aladjev2022-02-061-1/+4
| | | | | | | | | | rubygems Closes: https://bugs.gentoo.org/423589 Closes: https://bugs.gentoo.org/832268 Signed-off-by: Andrew Aladjev <aladjev.andrew@gmail.com> Signed-off-by: Hans de Graaff <graaff@gentoo.org>
* scons-utils.eclass: Replace deprecated PYTHON_MULTI_USEDEPMichał Górny2022-02-041-3/+3
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* eclass: deprecate the old go eclassesWilliam Hubbs2022-02-044-0/+4
| | | | | | | | The golang-* eclasses should not be used for new go packages. Modern go packages use go modules and the go-module eclass handles this setup. Signed-off-by: William Hubbs <williamh@gentoo.org>
* dune.eclass: enhance build and test logsAlfredo Tupone2022-02-021-2/+2
| | | | Signed-off-by: Alfredo Tupone <tupone@gentoo.org>
* distutils-r1.eclass: Restore QA warning for DUS + DISTUTILS_OPTIONALMichał Górny2022-02-011-0/+5
| | | | | | | Restore the QA warning (proviously issued as part of install-qa-check.d) for combining DISTUTILS_USE_SETUPTOOLS and DISTUTILS_OPTIONAL. Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Add DISTUTILS_DEPS output var for PEP 517 modeMichał Górny2022-02-011-1/+33
| | | | | | | | | | The PEP 517 build-time deps have gotten more complex, and largely depend on the internal eclass logic used to build and install wheels. Introduce a DISTUTILS_DEPS output variable that contains the correct BDEPEND string for use in DISTUTILS_OPTIONAL=1 ebuilds. Bug: https://bugs.gentoo.org/832337 Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Move DISTUTILS_OPTIONAL check into set_globalsMichał Górny2022-02-011-7/+9
| | | | Signed-off-by: Michał Górny <mgorny@gentoo.org>
* distutils-r1.eclass: Fix subphase return value passthroughMichał Górny2022-02-011-15/+33
| | | | | | | | | | | | | | | | | | | | | | | | Fix distutils-r1 phase functions to correctly pass through the return value from the subphases. This fixes e.g. the mistake of virtx not failing in the following case: src_test() { virtx distutils-r1_src_test } python_test() { epytest } This is because virtx implicitly uses nonfatal and epytest uses `die -n`. However, since the return value was not passed through, virtx never knew that anything has failed. While this covers only trivial cases and this is better solved via dying explicitly in the redefined python_test(), there's no harm in adding this passthrough. Signed-off-by: Michał Górny <mgorny@gentoo.org>