Since commit 8f64df0a7c (CUDA: Generic all and all-major support,
2021-12-19, v3.23.0-rc1~23^2), setting `CUDA_ARCHITECTURES` to a value
other than `all` or `all-major` is ignored with NVCC 11.5+. Fix the
logic to return early only when actually using an "all" value.
Fixes: #23243
When given two source paths via `-S` or just directory paths prefer
the last one. When the paths are mixed always prefer the last `-S`
entry.
Fixes: #23238
f73457ca2e cmake: Ignore any empty "" command line arguments
67f97f5478 Tests: Add RunCMake helper to run cmake with raw execute_process args
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !6980
Prior to commit 1d709ea2f5 (cmGeneratorTarget: Propagate backtraces from
INTERFACE_LINK_LIBRARIES, 2021-12-15, v3.23.0-rc1~228^2), the value of
`INTERFACE_LINK_LIBRARIES` was a single string. If an unquoted
generator expression passed to `target_link_libraries` contained `;` and
became multiple arguments, they would all accumulate as a single
`;`-separated list in the value of `INTERFACE_LINK_LIBRARIES`. Since
that commit, each argument to `target_link_libraries` is placed in
`INTERFACE_LINK_LIBRARIES` as a separate entry, as has long been the
case for `LINK_LIBRARIES`. That behavior change broke projects that were
accidentally relying on accumulation in `INTERFACE_LINK_LIBRARIES` to
produce complete generator expressions containing `;`.
Teach `target_link_libraries` to accumulate consecutive non-keyword
arguments into a single entry before placing them in `LINK_LIBRARIES` or
`INTERFACE_LINK_LIBRARIES`. For example, treat `a b c` as if they were
written as `"a;b;c"`. This restores the previously accidental support
for unquoted generator expressions in `INTERFACE_LINK_LIBRARIES`, and
also enables it for `LINK_LIBRARIES`.
For now, do not drop the `target_link_libraries` documentation that
recommends quoting generator expressions. Quoting is still important to
populate `LINK_LIBRARIES` in CMake 3.22 and below, and is also good
practice to keep generator expressions whole.
Fixes: #23203
Since commit daa6d2bc04 (ENH: updated handling of debug and optimized
target link libraries, 2006-11-29, v2.6.0~2471) the ancient
`<lib>_LINK_TYPE` compatibility lookup was done using the name of the
dependent target for which `target_link_libraries` is called, rather
than the name of the library dependency being considered. This code
probably does nothing. Remove it.
The argument name used in the class declaration didn't match the one
used in the implementation file. The one in the header appears to be a
cut-n-paste from GetDebugFindOutput(), but the implementation makes
it clear that "pkg" better infers what the argument represents than "var".
Extend the feature added by commit d7b18895bc (cmake: Add filtered
debug-find options, 2021-12-07, v3.23.0-rc1~217^2) to enable debug
output for `find_*` calls within a find module or cmake package
configuration file.
Fixes: #23211
The overload for `--debug-find-pkg` has the same signature as the base
class method for `--debug-find-var`. To avoid confusion, drop the
overload and inline it its only call site.
In commit 55e4753bbb (Refactor cmTargetExport removing
InterfaceIncludeDirecories, 2021-07-20, v3.22.0-rc1~337^2~1) the storage
of `INCLUDES DESTINATION` was moved into each target. However, a target
may be installed in multiple exports, and their `INCLUDES DESTINATION`
should not be mixed.
Convert the IncludeDirectoriesEntries vector to a map and modify access
function to store the directories lists with respect to cmExportTarget
object. This fixes error when the same target is exported more than once
via different exports and each for consequent export its include
directories list grows. Add a test for this case.
Fixes: #23183
5cdd774d51 VS: Handle build target correct for .NET SDK style projects with Any CPU
309191052c VS: Set Visual Studio versions read out from solution file
f7791698cb VS: Allow setting output directory in .NET SDK style projects
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !6944
* Extend Visual Studio solution parser for reading build target
* Map solution build target to project build target (especially for Any CPU)
* Use C++ <optional> template instead of pointer return value for cmSlnData::GetProjectByGUID
This was accidentally left out of commit 736663deed (Begin 3.23 release
versioning, 2022-02-03). The step is documented as part of the release
branching process in the CMake Maintainer Guide `Help/dev/maint.rst`.
Avoid looking up the PCH create/use flags just to map them through
flag tables back to the `.vcxproj` settings. Instead, simply
generate the PCH settings directly for each source file.
Since commit 9df1f33c9a (VisualStudio: move PCH rules to projects when
possible., 2020-10-15, v3.20.0-rc1~638^2) we already do this for the
target-wide PCH settings.
Commit 14d8a276 (CUDA: Support nvcc 11.5 new -arch=all|all-major flags,
2021-08-17) added all and all-major options to CUDA_ARCHITECTURES. These are
fairly generic and likely to see real-world use by distributors. Thus it's
desirable to support these also for Clang and older NVCC versions.
The supported architectures are dependent on the toolkit version. We determine
the toolkit version prior to compiler detection. For NVCC we get the version
from the vendor identification output, but for Clang we need to invoke NVCC
separately.
The architecture information is mostly based on the Wikipedia list with the
earliest supported version being CUDA 7.0. This could be documented and
expanded in the future to allow projects to query CUDA toolkit version and
architecture information.
For Clang we additionally constrain based on its support.
Additionally the architecture mismatch detection logic is fixed, improved and
updated for generic support:
* Commit 01428c55 (CUDA: Fail fast if CMAKE_CUDA_ARCHITECTURES doesn't work
during detection, 2020-08-29) enabled CMAKE_CUDA_COMPILER_ID_REQUIRE_SUCCESS
if CMAKE_CUDA_ARCHITECTURES is specified. This results in
CMakeDetermineCompilerID.cmake printing the compiler error and our code for
presenting the mismatch in a user-friendly way being useless. The custom
logic seems preferable so go back to not enabling it.
* Commit 14d8a276 (CUDA: Support nvcc 11.5 new -arch=all|all-major flags,
2021-08-17) tried to support CMP0054 but forgot to add x to the interpolated
result. Thus the conditions would always evaluate to false. This is fixed as
a byproduct of removing NVIDIA specific checks, improving the error message
and replacing architectures_mode with a simpler architectures_explicit.
Visual Studio support omits testing the flags during detection due to
complexities in determining the toolkit version when using it.
A long-term proper implementation would be #23161.
Implements #22860.
When support for `DOTNET_SDK` was added, only a minimal set of options
were built in. Based on user feedback, support for reference to dlls
(not just projects) is needed. That support is added here.
Fixes: #23166
f3ad061858 Add usage requirements to update direct link dependencies
193a999cd5 cmTarget: Add INTERFACE_LINK_LIBRARIES_DIRECT{,_EXCLUDE} backtrace storage
22d5427aa6 cmGeneratorTarget: Add LookupLinkItem option to consider own target name
f3d2eab36a cmGeneratorTarget: Fix link interface caching of partial results
d75ab9d066 cmGeneratorTarget: Clarify CMP0022 logic in ComputeLinkInterfaceLibraries
f3e9e03fe0 cmGeneratorTarget: Simplify CMP0022 warning check
216aa14997 cmGeneratorTarget: Return early from ExpandLinkItems with no items
1bc98371d1 Tests: Remove unnecessary policy setting from ObjectLibrary test
...
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !6886
Brad King