The logic added by commit 7808cbd644 (CMakeDetermineCompilerId: support
Intel DPC++ compiler toolset for VS gen, 2020-12-06, v3.20.0-rc1~330^2)
matches a specific toolset known to be the `icx.exe` compiler, and
assumes all other Intel C++ compilers (that are not DPC++) must be
`icl.exe`.
Since `icx.exe` is officially replacing `icl.exe`, use a regex that
matches the now-fixed set of toolsets known to use `icl.exe`. Any other
Intel C++ compiler will be assumed to be `icx.exe`.
Signed-off-by: William R. Dieter <william.r.dieter@intel.com>
Update the list of known versions.
Run the command
cmake -DBOOST_DIR=/path/to/boost_1_78_0 \
-P Utilities/Scripts/BoostScanDeps.cmake
to extract dependencies from the 1.78.0 source tree.
The dependencies differ from those of 1.77:
* The `log` component no longer depends on `date_time`.
Fixes: #23016
The `find_package(OpenMP)` calls added/updated by:
* commit f7f3d8987a (FindBLAS: Add dependency of OpenBLAS on OpenMP for
BLA_STATIC, 2020-11-10, v3.20.0-rc1~492^2)
* commit 9ef82d95d8 (FindBLAS: Fix detection of OpenMP as dependency of
BLA_STATIC, 2021-04-07, v3.20.1~3^2)
were missing the `QUIET` option.
Fixes: #23000
The current regular expression is able to match `/usr/lib/<arch>`,
`/usr/usr/lib/<arch>`, `/usr/usr/usr/lib/<arch>`, ... but not
`/lib/<arch>`.
This behavior ends up causing the detected architecture to
be x86_64-pc-linux-gnu when the Clang compiler is installed on
a "non-system" location (like /opt/llvm-13) which, in turn, makes
almost every 'find_library()' fail because the correct
architecture is x86_64-linux-gnu.
This is due to a typo in commit 764606e256 (CMakeDetermineCompilerABI:
Extract lib arch from implicit object file paths, 2021-04-05,
v3.20.1~10^2), which used `+` instead of `?`.
The upstream `openssl` build system may install libraries to `lib64`
even on platforms whose conventions do not use `lib64` for
distro-packaged libraries.
Fixes: #22945
In commit 7b83ca816a (FindOpenSSL: add target OpenSSL::applink to
support OpenSSL's applink feature, 2020-05-12, v3.18.0-rc1~150^2) the
version check was written as "major.major.fix" instead of
"major.minor.fix".
Since commit 94a84dc0af (FindPkgConfig: add pkgconf to the search list.,
2021-07-02, v3.22.0-rc1~468^2), `pkgconf` is preferred over `pkg-config`
if they appear in the same directory. In some environments,
`pkg-config` may be a wrapper that adds semantics beyond either
`pkgconf` or the normal `pkg-config`. Prefer `pkg-config` over
`pkgconf` in order to preserve the prior behavior in such environments.
Fixes: #22976
Update logic added by commit ecaca8c129 (GNUInstallDirs now aware of
conda lib directory requirements, 2021-09-08, v3.22.0-rc1~142^2).
When it is ambiguous if we are doing a conda install or a system
install prefer using the system library directory.
Fixes: #22962
0fc8b2f61c CompilerId: Restore support for classic C by avoiding C++ style comments
Acked-by: Kitware Robot <kwrobot@kitware.com>
Reviewed-by: Raul Tambre <raul@tambre.ee>
Merge-request: !6759
C++ style comments were added by commit fc3a1cbdd8 (CompilerID: Compiler
extensions default detection, 2021-05-29, v3.22.0-rc1~52^2~2), but they
may not be supported by the default mode of some C compilers. Use
C-style comments instead. For consistency, do this for all languages.
Fixes: #22942
`googlemock` has been absorbed into the
[googletest](https://github.com/google/googletest) project and is built
and installed from the same source tree.
As GTest may be built with or without GMock, skip GMock if it is not
present.
Do not provide result variables for GMock. They are not provided by
upstream GTest's CMake Package Configuration File.
Also update the test case to cover linking to `GTest::gmock`.
When `Visual Studio` and `Xcode` generators are used, directory for depfile
is not implicitely created by CMake when OUTFILE_DIR option is used.
Fixes: #22932
When the HDF5 HL library is not available, FindHDF5 with Fortran does
not work. The problem is because `cmake_hdf5_test.f90` tries to
"use h5lt" and "use h5ds", which comes from the HL library.
Then the basic compiler check fails completely.
The two "use" lines were added by commit 00405af0f1 (HDF5: Refactor the
use of compiler wrappers, 2016-04-04, v3.6.0-rc1~85^2) along with the
rest of the check. Remove them.
Fixes: #22890
* Update generator to use the current variable definition,
not the CACHE definition.
* Avoid overwriting toolchain variables and avoid developer warnings
when setting up the GHS variables.
-- A cache entry is only required when:
a) The toolchain uses set( CACHE ) to set the variable
b) A -D or preset value was used to set the variable
The cache entry is required so that the TYPE gets set properly.
-- Avoid the Policy CMP0126 warnings:
setting cache variable when normal variable exists
* Move GHS_PRIMARY_TARGET back into area so non-GHS generator
toolchain file can trigger the OS search.
Selection of a BSP only needs to be performed if not set by user.
Remove all the logic for printing error and status messages about BSP
selection. These messages also breaks CMake tests.
NOTE: If BSP selection fails then the compiler checks will result in
a build error. The build error will report that the BSP does not exist.
Selection of an RTOS only needs to be performed if not set by user.
Avoid CMake developer warnings for setting GHS_OS_DIR multiple times.
Instead only set it once after searching has been performed.
Remove all the logic for printing error and status messages about RTOS
selection. This was broken and reporting incorrect messages. These
messages also breaks CMake tests.
NOTE: If RTOS selection fails then the compiler checks will result in
a build error. The build error will report that the RTOS
"GHS_OS_DIR-NOT-SPECIFIED" does not exist.
-- Ensure that GHS_TOOLSET_ROOT is used as a path
* Converts directory path slashes to CMake style
-- Use ComparePath() to properly check for path changes of build tool
* Accounts for Windows file-system case insensitivity.
-- Don't print message "defaulting" messages (this causes CMake test failures)
-- Don't force update CMAKE_GENERATOR_TOOLSET back into cache when `-T`
is not used on initial configure. This change avoids an unnessary
error message when accidentally using `-T` in subsequent runs but the same
tools are always used.
-- Use IssueMessage() for error messages.
Changes to ``-A`` handling:
* Don't force CMAKE_GENERATOR_PLATFORM into cache when using default
value (breaks using CMake presets).
* Don't print message when using default value (breaks CMake tests).
Changes to ``GHS_PRIMARY_TARGET`` handling:
* Add as a cache variable so its known to GUI
* Don't always include``GHS_TARGET_PLATFORM``, it's only needed if
``GHS_PRIMARY_TARGET`` wasn't set by the user.
* Set ``GHS_PRIMARY_TARGET`` during platform selection instead of when
a language is enabled. By performing this sooner
``GHS_TARGET_PLATFORM`` is not always required to be set into cache.
Brad King