The CMP0157 old behavior omitted the output-file-map after the Swift
build split because we stopped calling `EmitSwiftDependencyInfo` when we
could not split the build because we didn't call
`WriteObjectBuildStatement`. If we can't split the build, then we still
need to include the OFM information to not break incremental builds.
In commit 878ae03832 (macOS: IMPORTED framework: Honor SYSTEM target
property in all cases, 2023-08-27, v3.28.0-rc1~162^2) we broke support
for `IMPORTED_LOCATION_<CONFIG>` without `IMPORTED_CONFIGURATIONS`.
Previously it worked if the importing project's configurations match the
set of `IMPORTED_LOCATION_<CONFIG>` properties set. Fix that case.
Fixes: #25506
Issue: #25515
If the working directory didn't exist, the old message just says:
Error running test executable.
Path: '<Path to test executable>'
Result: No such file or directory
Output:
This leads the user to the conclusion that the test executable doesn't
exist, which isn't true. Make the true cause visible by reporting the
working directory in the error message.
The test added by commit 37bc3400cd (CMakePackageConfigHelpers: Add
generate_apple_platform_selection_file(), 2023-11-03) covers importing
only on macOS, and is specific to the xcframework test case. Add a
dedicated test for `generate_apple_platform_selection_file` that covers
export and import on all platforms.
04959e0285 Tests: Verify CUDA::cupti existence on a subset of machines
09a3c8b46b CUDAToolkit: Correctly search all include paths from compiler
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !9072
Splitting the Swift build into an object build and a separate link step,
instead of building and linking in one step. The immediate benefit is
LSP support because we are able to emit compile-commands for Swift files
now. Additionally, it is possible to specify flags to the compile step,
enabling folks to emit C and C++ headers from their Swift builds for
C/C++ interop, without needing custom commands. Eventually, this gives
us a path toward working object libraries.
Object Libraries:
- Object libraries don't work today because CMake doesn't emit targets
for object libraries into the Ninja build file.
- tl;dr: Object libraries work if they aren't WMO. Still need work to
make WMO'd object libraries work.
Object libraries still don't completely work with this patch because,
while we emit the targets, the `TARGET_OBJECTS` generator expression
expansion has a separate mechanism for determining what the names of
the objects are based on the input source files, so targets that
depend on an object library built with a whole-module optimization
will depend on objects based on the name of the source file instead
of the actual emitted object file.
These features require being able to accurately model wholemodule builds
though, because we actually need to track object files and WMO affects
what objects are emitted. For that, we require CMP0157 use the NEW
policy. When it's OLD, we have to fall back on the old behavior and
cannot provide object libraries or the compile-commands for LSP.
Issue: #25308
Previously `RunCMake.CTestTimeout` failed on GNU/Hurd because the
`sleep` command does not set `errno` to `EINTR` when interrupted by a
signal, which is a Linux-specific feature.
Since commit 0f37000304 (try_{compile,run}: add LINKER_LANGUAGE option,
2023-10-12) the test has failed with gfortran < 4.3 because it does
not support the Fortran syntax used by the test. Add a version check.
LLVMFlang 18.0 adds MSVC runtime library selection flags and associated
Fortran runtime library variants. Resolve the corresponding FIXME left
by commit 26bf32cdc6 (LLVMFlang: Add support for targeting MSVC ABI on
Windows, 2023-09-28, v3.28.0-rc1~10^2).
Issue: #24840
As of llvm-project `main` branch commit `86accd4e03` (2023-12-04),
LLVMFlang 18.0.0, when used to drive linking an executable, emits a MSVC
linker flag to use all object files from the `Fortran_main` library.
These object files are meant for use when linking the program entry
point, and so are not implicit link dependencies of Fortran libraries.
The internal helper variable '_GOOGLETEST_DISCOVER_TESTS_SCRIPT' can have gone
out-of-scope when 'gtest_discover_tests()' is called, depending on where the
GoogleTest module is actually included. This leads to a silent failure of
dynamic test discovery, since the custom post-build commands actually does
nothing (it basically invokes 'cmake -P ""'). Ctest will then fail to run the
tests, considering them to be 'not built'.
Fix this by determining the path to the GoogleTest module based on
'${CMAKE_ROOT}' instead, which is always available.
A new test case was added to test suite 'RunCMake/GoogleTest' to ensure that
'gtest_discover_tests()' works correctly when invoked in a different variable
scope.
Fixes: #25477
Previously CMake may generate incomplete transitive requirements in
CMakeFiles/<target>.dir/CXXModules.json and therefore in module mapper
for compiler, when source files were listed in CMakeList.txt in a
certain order.
This commit fixes the problem by correctly tracking unfinished
transitive requirements computation of module units.
There have been a simple circular test case whose circular dependency
was reported by build system. Now with this correct implementation it's
reported by CMake generating module mappers.
Add two test cases for transitive requirements computation, one with
adding source files in hardcoded order, and the other in randomized
order.
Fixes: #25465
40dc13b242 cmNinjaTargetGenerator: PCH files do not need dyndep
f61c64cd1c cmLocalGenerator: prevent scanning of PCH source files
ea8c37b759 Tests/CXXModules: add a test which scans a PCH-using source
Acked-by: Kitware Robot <kwrobot@kitware.com>
Tested-by: buildbot <buildbot@kitware.com>
Merge-request: !9032
This test previously did not *require* that the internal partition be
specified as a transitive usage because nothing from it was exposed.
Plumb through usages such that the internal partitions are required.
Brad King