The search logic for the IAR linker, librarian and related tools is
moved from `Modules/Compiler/IAR-FindBinUtils.cmake` to
`Modules/CMakeFindBinUtils.cmake` by introducing an new elseif block
for the IAR toolchain.
The search logic was refactored to omit repeating itself.
Fixes: #22425
Revert commit 0665d9092e (CMakeDependentOption: Allow parentheses in the
depends string, 2021-06-13, v3.21.0-rc1~32^2). It broke existing use
cases with parentheses in regular expressions. Also add a test for this.
Fixes: #22447
IAR is available for both Linux and Windows OSes. Moreover,
binaries `*.exe` could be called from a Linux box, or vice versa.
See also commit 01a4eec446 (IAR: Changes required for Linux,
2019-04-25, v3.15.0-rc1~185^2).
Fixes: #22312
Setting `CMAKE_EXECUTABLE_FORMAT` as a normal variable is unnecessary
because setting it as a cache entry already makes the value visible to
the calling scope.
Fixes: #22433
While the flag tables for C and C++ were generated from MSBuild `.xml`
files, the CSharp flag tables were written by hand. Copy the `v142`
flag table to use for the `v143` toolset.
Update the logic that converts a `try_compile` result from a cache
entry to a normal variable to tolerate an existing normal variable
under CMP0126 NEW behavior. Otherwise the `try_compile` result
is ignored because CMake uses the false value of the normal variable,
and CMake incorrectly reports that the compiler does not work.
This went unnoticed for some languages (e.g. C and CXX) because the
check for a working compiler is skipped if ABI detection works.
It does affect other languages (e.g. CSharp).
Fixes: #22423
Given the compiler to use, `CMakeFindBinUtils.cmake` automatically
determines a number of tools including linker (CMAKE_LINKER) and archiver
(CMAKE_AR) and stores them in a generated file `CMakeCCompiler.cmake` as
non-CACHE entries. The compiler-specific ARMClang.cmake then tries to
override CMAKE_LINKER and CMAKE_AR as CACHE entries.
Following the introduction of CMP0126, which is set to NEW in the test
for a working compiler, setting a CACHE entry does not replace a normal
entry of the same name anymore, resulting in a failed test due to wrong
linker and archiver.
To fix this, set CMAKE_LINKER and CMAKE_AR for ARMClang directly in
`CMakeFindBinUtils.cmake` as is done for other compilers. Check
for them in `ARMClang.cmake` to safeguard cases when a project explicitly
includes `ARMClang.cmake` prior to compiler determination (which some
projects do to work around other problems in older CMake versions).
When a user provides `CPACK_DEBIAN_PACKAGE_CONTROL_EXTRA` or
`CPACK_DEBIAN_<comp>_PACKAGE_CONTROL_EXTRA` variables in
`CMakeLists.txt` and the package contains dynamic libraries, the
`CPackDeb.cmake` sets `CPACK_ADD_LDCONFIG_CALL` to `1`. Later it
analyzes if defaulted `postinst`/`postrm` should be generated trying to
check if the user provides any in `CPACK_DEBIAN_PACKAGE_CONTROL_EXTRA`
variable. However, the `foreach` loop uses the invalid variable
`PACKAGE_CONTROL_EXTRA` instead, so these files gonna be overridden.
Fix the variable name.
Fixes: #22410
Logic added by commit 4c74c86f40 (FindBLAS/LAPACK: Add support for the
Fujitsu SSL2 library, 2021-01-27, v3.21.0-rc1~402^2~1) accidentally
expressed a boolean condition without proper grouping. The pattern was
then copied by commit 2c9e623e31 (Find{BLAS,LAPACK}: Add support for the
NVHPC LAPACK library, 2021-05-05, v3.21.0-rc1~192^2). The resulting
logic incorrectly tries Fujitsu and NVHPC vendors even after results are
found from another vendor, and then erases those. Fix the grouping.
Fixes: #22403
Revert commit 74cc2e3326 (FindJPEG: Search for 'turbojpeg' and
'turbojpeg-static' too, 2021-01-09, v3.20.0-rc1~176^2) pending further
investigation. The "turbo" variants are not drop-in replacements on all
platforms.
Fixes: #22333
Update the IAR linker and archiver rules to use the `<CMAKE_LINKER>`
and `<CMAKE_AR>` placeholders instead of hard-coding the tool names.
Fixes: #22395
This was previously fixed by commit d46590910c (Check*CompilerFlag: Do
not set result as a normal variable too, 2020-09-21, v3.18.3~1^2^2), but
was regressed by refactoring in commit 90dead024c (CheckCompilerFlag:
unified way to check compiler flags per language, 2020-09-25,
v3.19.0-rc1~88^2) due to the changes being developed concurrently.
Fix it again, and add a test case.
Fixes: #21207
In commit a90d2a9eed (IntelLLVM: Add support for Intel LLVM-based
compilers, 2020-11-02, v3.20.0-rc1~89^2~20) we accidentally left out
activation of the C/C++ standard level selection logic when IntelLLVM is
targeting the MSVC ABI.
Fixes: #22388
In commit a90d2a9eed (IntelLLVM: Add support for Intel LLVM-based
compilers, 2020-11-02, v3.20.0-rc1~89^2~20) we accidentally left out
activation of the C/C++ standard level selection logic when IntelLLVM is
targeting the MSVC ABI.
Fixes: #22388
Update the change from commit c6aaaf066a (DetermineCompiler: Restore
identification of MSVC with no INCLUDE dirs, 2021-06-29,
v3.21.0-rc2~6^2) to also work with no `LIB` dirs. Add the `-c` flag to
compile an object file without linking.
Since commit bd844387df (ROCMClang: Add the ROCm toolkit derived clang
compiler to CMake, 2020-08-28, v3.21.0-rc1~66^2~6), our generated
compiler id source file contains a preprocessor condition of the form
#elif ... && __has_include(...)
When MSVC is invoked with no include directories in the `INCLUDE`
environment variable or in `-I` flags, its implementation of
`__has_include` errors out immediately. This breaks identification of
the compiler.
Work around the problem by adding a fallback attempt to compile the
identification source using a dummy include directory.
RPM 4.11.3 for el7 contains backported support for the RPM weak dep tags.
It only supports querying those tags, but rpmbuild can not make use of them.
Since CPack relies on rpmbuild --querytags to check for support, this
commit switches to rpm --suggests to check for support of weak
dependencies.
Fixes: #22350
In 3aec26660a (Merge topic 'nasm_add_deps', 2021-03-03) the
ASM_NASM language enabled the generation of GCC compatible
depfiles for NASM. Unfortunatly, those are not supported by YASM.
Fixes: #22341
In 5e941a545b (ExternalProject: Ensure git fetch if updating to hash we
don't have yet, 2021-05-09), the form ${ref}^{commit} was used to ensure
the ref was treated as a commit. When running under a MSYS2 or
Cygwin shell, the {commit} part can be translated by the shell, leaving it as
${ref}^commit, which git doesn't understand. Use the equivalent but
safer form ${ref}^0, since it won't be subject to the shell substitution
issue.
Fixes: #22299
If a project or user sets a variable or cache entry named `C`, `CXX`,
`CUDA`, `Fortran`, `CSharp`, or `ISPC`, we were previously comparing
each enabled language name to the value of that variable, rather than
the name itself. Double-quote the string to take advantage of policy
`CMP0054`, but also add "x" prefixes to support projects that do not set
the policy.
This extends a similar fix in commit bd16a985fc (CompilerId/Features:
Tolerate variables named for languages, 2021-04-28) in more places. In
particular, ensure that we consider the proper list of compiler ids.
Issue: #22125
Brad King