Cray Fortran uses mangling of the form `my_sub$my_module_` with the
subroutine name first. Teach FortranCInterface to detect this case.
Add `FortranCInterface_MODULE_{,_}ORDER` result variables to report it.
With optimizations on, Cray Fortran inlines the module subroutine into
the calling object, so our symbol with the `INFO` string is not used.
Add a directive to suppress inlining to avoid this.
This was added in commit 98d0f918ba (LFortran: Add support for this
compiler, 2024-01-25, v3.31.0-rc1~303^2~2) because it is needed for
cases covered by CMake's Fortran tests. However, it does not work
with Fortran modules and breaks lfortran's own `examples/project1`.
Move the flag to the test cases that need it, just as the original
commit did with `--implicit-interface`.
Fixes: #26597
Co-authored-by: Brad King <brad.king@kitware.com>
Fat LTO objects contain both traditional object code and the LTO bitcode
IR, but the GNU compiler does not support them on Apple platforms.
A compile error is raised when `-f[no-]fat-lto-objects` flags are used,
so avoid them.
This also implies that static Fortran libraries cannot be built with IPO.
Fixes: #25931
d7c8030541 (FortranCInterface: Fix misuse of IS_NEWER_THAN in
timestamp check, 2021-02-21) updated the IS_NEWER_THAN logic, but it
introduced a couple of errors. 2a00e5072d (FortranCInterface: Fix
regression in timestamp check, 2021-09-30) addressed one of those errors,
but there was still one left behind that wasn't noticed. The Output.cmake
file is in the build directory, but there was still one reference to it that
incorrectly used a path to it in the source directory.
Issue: #22709
GenerateExportHeader had a hidden state requirement that other
modules were included first. Considering include_guard, Modules
should include all they actually use.
`try_compile` and `try_run` now automatically log checks using them to
`CMakeConfigureLog.yaml`.
Add `LOG_DESCRIPTION` arguments to some `try_compile` calls to
replace the description previously written to the old logs.
Issue: #23200
Modify modules that ship with CMake and use the project flavor of
try_compile to use the new signature added by commit 56ae40cc59
(try_compile: Add PROJECT keyword-dispatched signature, 2022-09-14).
The following `module.f90` file
module mymodule
contains
subroutine mysub()
end subroutine
end module
when compiled with `flang-new` (from LLVM 15.0.0) generate the
`_QMmymodulePmysub` symbol.
$ flang-new -c module.f90
$ nm module.o
0000000000000000 T _QMmymodulePmysub
This commit fixes the regular expressions accordingly.
The C flags added by commit 6a0ce19ce1 (FortranCInterface: Fix
compatibility with GCC gfortran 12 LTO, 2022-01-19, v3.22.2~5^2)
should only be added for the GNU C compiler.
Fixes: #23500
Issue: #23123
Since version 12.0 the GCC Fortran compiler has implemented "WG5/N1942",
which causes, if link-time opmization is enabled, obfuscation of hard-coded
string values in the compiler objects and its resulting ELF-binaries.
This causes the CMake-internal detection of the mangling scheme for the
naming of subroutines to fail. Thus we must ensure to have any link-time
optimization features to be disabled on the executable file we perform the
detection on.
The static libraries, however, must be build with LTO and non-LTO objects,
as that will ensure the verify step will operate on IPO objects, if building
those is requested by the system compiler flags.
Fixes: #23123
Signed-off-by: Björn Esser <besser82@fedoraproject.org>
Since commit d7c8030541 (FortranCInterface: Fix misuse of IS_NEWER_THAN
in timestamp check, 2021-02-21, v3.21.0-rc1~631^2~3), FortranCInterface
checks for `Output.cmake.in` in the build tree instead of the source
tree as before. This caused it to always re-run the detection.
Fixes: #22709
When using a file system which only has second resolution timestamps,
there is a reasonably high likelihood of timestamps being the same.
The IS_NEWER_THAN test returns true when timestamps are the same,
so don't redo detection when they match exactly.
Policy CMP0056 determines whether `CMAKE_EXE_LINKER_FLAGS` are passed
into the test project used by the source-file signature of `try_compile`.
That affects how implicit link directories are detected, so we need to
also honor the policy for the source-directory signature of `try_compile`
used in FortranCInterface in order to get matching link directories.
Fixes: #21408
Previously the `find_program` call we used to locate the test executable
but that can be broken by `CMAKE_FIND_ROOT_PATH_MODE_PROGRAM`. Instead
teach the test project to write a file with the location of the
executable it builds. Load that file to get the exact location.
Fixes: #20390
In commit beb991110d (Remove now-unused code once used on IRIX,
2019-01-11, v3.14.0-rc1~167^2) we removed remnants of IRIX support.
Also remove remnants of MIPSpro compiler support.
This change ensures that Intel Fortran's /libs: in
CMAKE_Fortran_FLAGS and Visual C++'s /MT or /MD in the
CMAKE_C_FLAGS_RELEASE do not conflict with each other.
When using a Visual Studio generator with an Intel toolset, such as
-T "Intel C++ Compiler XE 14.0"
the generated FortranCInterface mangling detection project may fail to
build due to `devenv` not working with the `/project ALL_BUILD` option.
This seems to be a bug in `devenv` or the Intel VS integration. Work
around the problem by building with `/project FortranCInterface`
instead. We only need to build this executable and its dependencies
within the detection test project anyway.
Fixes: #16519
When using a Fortran compiler that produces PIC executables by default
with a C compiler that does not produce PIC by default then the mangling
detection executable fails to link. Explicitly enable PIC for the C
compiler just in case the Fortran compiler needs it to link.
Issue: #16405
Per-source copyright/license notice headers that spell out copyright holder
names and years are hard to maintain and often out-of-date or plain wrong.
Precise contributor information is already maintained automatically by the
version control tool. Ultimately it is the receiver of a file who is
responsible for determining its licensing status, and per-source notices are
merely a convenience. Therefore it is simpler and more accurate for
each source to have a generic notice of the license name and references to
more detailed information on copyright holders and full license terms.
Our `Copyright.txt` file now contains a list of Contributors whose names
appeared source-level copyright notices. It also references version control
history for more precise information. Therefore we no longer need to spell
out the list of Contributors in each source file notice.
Replace CMake per-source copyright/license notice headers with a short
description of the license and links to `Copyright.txt` and online information
available from "https://cmake.org/licensing". The online URL also handles
cases of modules being copied out of our source into other projects, so we
can drop our notices about replacing links with full license text.
Run the `Utilities/Scripts/filter-notices.bash` script to perform the majority
of the replacements mechanically. Manually fix up shebang lines and trailing
newlines in a few files. Manually update the notices in a few files that the
script does not handle.
Run the `Utilities/Scripts/clang-format.bash` script to update
all our C++ code to a new style defined by `.clang-format`.
Use `clang-format` version 3.8.
* If you reached this commit for a line in `git blame`, re-run the blame
operation starting at the parent of this commit to see older history
for the content.
* See the parent commit for instructions to rebase a change across this
style transition commit.
When testing CMAKE_<LANG>_COMPILER_ID values, do not explicitly
dereference or quote the variable. We want if() to auto-dereference the
variable and not its value. Also replace MATCHES with STREQUAL where
equivalent.
A few different regular expressions were being used in various
places to extract info strings from binaries. This uses a
consistent regex amongst all of them now. This also fixes the
broken ABI detection for Cray compilers.
Update cmake_minimum_required calls in CMakeLists.txt in Modules and in
CMakeLists.txt generated by other modules, so that they are always in
sync with current CMake version.
The matches have already been calculated and can simply be taken from
CMAKE_MATCH_n variables. This avoids multiple compilations of the same or very
similar regular expressions.
After building the test binary tell find_program to search for it with
the ${CMAKE_EXECUTABLE_SUFFIX} so that the .exe can be found. Since
find_program is normally used to locate host tools while cross-compiling
it needs this hint to find the target binary.
Suggested-by: Denis Barbier <bouzim@gmail.com>
Ancient versions of CMake required else(), endif(), and similar block
termination commands to have arguments matching the command starting the
block. This is no longer the preferred style.
Run the following shell code:
for c in else endif endforeach endfunction endmacro endwhile; do
echo 's/\b'"$c"'\(\s*\)(.\+)/'"$c"'\1()/'
done >convert.sed &&
git ls-files -z -- bootstrap '*.cmake' '*.cmake.in' '*CMakeLists.txt' |
egrep -z -v '^(Utilities/cm|Source/kwsys/)' |
egrep -z -v 'Tests/CMakeTests/While-Endwhile-' |
xargs -0 sed -i -f convert.sed &&
rm convert.sed
To use VS C and Fotran in the same solution, it is required that VS be
able to find the Fortran run time libraries as they will be implicitly
linked by any Fortran library used by VS C programs. This adds a check
into CMakeDetermineCompilerABI using a try-compile to find the correct
PATH.
The Intel Fortran plugin for Visual Studio requires Fortran source files
to be compiled in a separate target from C and C++ code. Compile the
VerifyFortran.f source file in a separate library and link the main
VerifyFortanC executable to it.
The Cray Fortran compiler started using module init symbols in version 7.3.2.
Starting in commit 71287734 (Teach FortranC interface for Intel, PGI, and gcc
4.2, 2009-08-05) we provide C versions of the module init symbols so that the
detection executable can link when the C versions of the module-mangled symbols
are picked up.
If no C module-mangled symbol matches then we cannot let the C module init
symbol appear because it will be duplicated by the Fortran copy that provides
the module-mangled symbol. This was first handled for the PathScale compiler
in commit 21faaa5d (FortranCInterface: Fix PathScale detection, 2010-01-22) and
commit 46858720 (FortranCInterface: Fix PathScale detection again, 2010-02-16).
Handle it now for the Cray compiler too.
PathScale Fortran mangles module symbols as "MY_SUB.in.MY_MODULE" and
also requires "my_module_" when the module is imported. We cannot
provide the symbol with ".in." mangling so we should not provide
"my_module_" because it would duplicate the one in the Fortran-provided
object file.
Commit "FortranCInterface: Fix PathScale detection" (2010-01-22) already
made the same fix for the non-underscore module case.
PathScale Fortran mangles module symbols as "MYSUB.in.MYMODULE" and also
requires "mymodule_" when the module is imported. We cannot provide the
symbol with ".in." mangling so we should not provide "mymodule_" because
it would duplicate the one in the Fortran-provided object file.
The commit "FortranCInterface: Honor language flags in checks" taught
the FortranCInterface module to pass C and Fortran flags into its
detection and verification checks. We improve on the change to allow
the '=' character in the language flags. This requires passing the
cache entry type with the -D options.
We pass CMAKE_C_FLAGS, CMAKE_CXX_FLAGS, and CMAKE_Fortran_FLAGS through
try_compile() for the FortranCInterface Detect and Verify projects.
This honors user-specified compiler flags for each language, thus
supporting flags that affect the Fortran mangling.
This adds copyright/license notification blocks CMake's non-find
modules. Most of the modules had no notices at all. Some had notices
referring to the BSD license already. This commit normalizes existing
notices and adds missing notices.
