algaos/algaos-ebuild-tree/eclass/cmake.eclass, branch next

cmake.eclass: only define Fortran compiler if using fortran-2.eclass

2026-09-08T20:29:06Z

One is "not supposed to" use fortran in an ebuild without inheriting this eclass. Unconditionally doing so regardless, means that any ebuild which makes use of this cmake compiler definition likely has an automagic dep on virtual/fortran and definitely isn't running compiler sanity pre-checks which we are led to believe by the fortran eclass maintainers, is very important to check the sanity of. Thanks-to: Eli Schwartz Closes: https://bugs.gentoo.org/981677 Signed-off-by: Andreas Sturmlechner

cmake.eclass: save build tool's exit code in case of running under nonfatal

2026-06-08T20:01:48Z

Thanks-to: Fat-Zer Bug: https://bugs.gentoo.org/970798 Signed-off-by: Andreas Sturmlechner

cmake.eclass: Unconditionally set CMAKE_SYSTEM_NAME in toolchain file

2026-06-08T20:01:48Z

Also explicitly add CMAKE_CROSSCOMPILING=(TRUE|FALSE). Bug: https://bugs.gentoo.org/975603 Signed-off-by: Andreas Sturmlechner

cmake.eclass: cmake_src_configure: Move conditionals into heredoc, write once

2026-06-08T20:01:48Z

Signed-off-by: Andreas Sturmlechner

cmake.eclass: Add Hurd support

2026-03-21T18:31:56Z

On CMake, `GNU` is the sysname for Hurd. https://cmake.org/cmake/help/latest/variable/CMAKE_SYSTEM_NAME.html#system-names-known-to-cmake Signed-off-by: Christopher Fore Signed-off-by: Sam James

cmake.eclass: fix removal of FindBLAS.cmake during cmake-4.2.1 upgrade

2026-01-11T09:16:58Z

Change done in 3f8ff6b53f4c9d048cb677b0e41a4f5089951d5d had a logic mistake. Original logic was: if CMAKE_REMOVE_MODULES_LIST is not empty, then validate that it is array. Otherwise (emty string or unset), set it to remove FindBLAS.cmake and FindLAPACK.cmake. After change if effectively always set any non-empty CMAKE_REMOVE_MODULES_LIST array to ( FindBLAS FindLAPACK ), if Part-of: https://github.com/gentoo/gentoo/pull/45321 Signed-off-by: Sam James

cmake.eclass: Restrict default rm Find{BLAS,LAPACK}.cmake to CMake <4.2.1

2026-01-05T16:59:46Z

This goes back to ancient cmake-utils.eclass,v 1.70 in CVS of 2011, or historical.git commit 70230d8d15c0d287b5e81b09a6e78b6aff01b1d6, and it was done to ensure use of downstream-modified versions of these modules shipped by Gentoo's cmake package over consumer packages' own copy overriding them. After commit 04ce4a03554e3541c587db60013b9b796613c2c7 finally dropped this downstream modification for CMake >=4.2.1, we can also stop removing these modules by default. Closes: https://bugs.gentoo.org/736547 Signed-off-by: Andreas Sturmlechner

cmake.eclass: cmake_comment_add_subdirectory: Work w/ macro args

2026-01-05T16:59:46Z

As of 3.25, CMake allows 3 optional arguments to pass in addition to source_dir: [binary_dir] [EXCLUDE_FROM_ALL] [SYSTEM] The eclass function so far did not cope with that. See also: https://cmake.org/cmake/help/latest/command/add_subdirectory.html Signed-off-by: Andreas Sturmlechner

cmake.eclass: fix MYCMAKEARGS not being listed under variables

2025-11-23T18:42:08Z

This variable is a user-variable similar to ECONF_EXTRA in nature. Document it as such. Signed-off-by: Arsen Arsenović Signed-off-by: Sam James

cmake.eclass: Introduce cmake_prepare-per-cmakelists

2025-11-16T19:00:15Z

Override this to be provided with a hook into the cmake_src_prepare loop over all CMakeLists.txt below CMAKE_USE_DIR. Will be called from inside that loop with as single argument. Used for recursive CMakeLists.txt detections and modifications. Signed-off-by: Andreas Sturmlechner