The test added in commit 50e90e2828 (try_compile: Honor CMP0128 setting
in test project, 2022-10-18, v3.25.0-rc2~4^2) requires that the compiler
support C++11 mode, so do not run it on GNU compilers that are too old.
Some projects pass a raw `-std=` flag to the compiler in the
`try_compile` project. If they do not set CMP0128 to NEW,
we should not append a `-std=` flag where we did not before
the policy was added.
Fixes: #24063
When `--debug-trycompile` is enabled, issue a log (with trace) for every
`try_compile` noting what directory is being used for the compilation.
This will make it easier to find the corresponding artifacts.
Closes: #24022
Change the SOURCE_FROM_ARG keyword to try_compile to SOURCE_FROM_CONTENT
(which we can do because it was recently added and hasn't been in a
release yet). The new name should be clearer as to what it does, and
also more consistent with the CONTENT arguments to some other commands.
Also, fix a typo in an error message.
Add ability to copy try_compile (and try_run) source files from
arbitrary locations into the operation directory. This is included for
the sake of completion and consolidation, although use cases which
actually require this may be rare.
Add ability to "feed" try_compile (and try_run) sources more directly,
either from literal content, or from a CMake variable which contains
literal content. This saves the user from needing a separate step to
write the content to a file, and allows for the sources to only exist in
the scratch directory.
Introduce a new signature for the project flavor of try_compile (and
try_run) which removes the `bindir` argument and adds a required PROJECT
tag. This is similar to the SOURCES flavor added by commit aa9220d3
(try_compile: Add keyword-dispatched signature, 2022-09-02).
Fix the tests added by commit 0c141b03 (try_compile: Record output
location instead of reverse computing it, 2022-08-31), which were added
as old-and-new-signature tests but only used the old signature.
Introduce a new signature for try_compile (and try_run) which removes
the `bindir` argument and requires the SOURCES tag. This will eventually
allow us to add other ways of providing sources, but also allows us to
change the behavior without breaking compatibility.
The old signature uses a special, but non-unique temporary location
inside the specified `bindir`, which conventionally is just the
project's build directory. The new signature unconditionally uses the a
unique temporary directory which is unconditionally within the project's
build directory (which is no longer separately specified). This ensures
that successive runs do not overwrite previous runs, will simplify
debugging, and should also, eventually, allow us to execute multiple
trials in parallel.
Modify cmCoreTryCompile to only recognize try_run arguments when
actually parsing try_run. (The old behavior was to recognize try_run
arguments for try_compile also and then complain.)
This has a small chance that a try_run keyword will be consumed as part
of a multi-valued keyword argument. However, this behavior is more
consistent with other commands, as we don't normally treat keywords as
universally reserved. Also, the code is noticeably simplified.
C++ modules have two variants which are of importance to CMake:
- `CXX_MODULES`: interface modules (those using `export module M;`,
`export module M:part;`, or `module M:internal_part;`)
- `CXX_MODULE_HEADER_UNITS`: importable header units
Creating C++ modules or partitions are *not* supported in any other
source listing. This is because the source files must be installed (so
their scope matters), but not part of usage requirements (what it means
for a module source to be injected into a consumer is not clear at this
moment). Due to the way `FILE_SET` works with scopes, they are a perfect
fit as long as `INTERFACE` is not allowed (which it is not).
Add policy CMP0137 to propagate both our builtin variables and those
listed by `CMAKE_TRY_COMPILE_PLATFORM_VARIABLES` to `try_compile`
whole-project builds.
Inspired-by: Alexander Neumann <Alexander.Neumann@hamburg.de>
Fixes: #23219
Since:
* commit e216b9bbd3 (cmake: Allow CMAKE_BUILD_TYPE to be set by
environment variable, 2021-06-29, v3.22.0-rc1~503^2~1)
* commit ef56eefc9b (cmake: Allow CMAKE_CONFIGURATION_TYPES to be set by
environment variable, 2021-06-29, v3.22.0-rc1~503^2)
the environment variables are supposed to provide defaults for settings
the user otherwise can control via cache entries. However, they
accidentally affect `try_compile` projects too, which are supposed to be
programmatically controlled.
Fixes: #22935
Due to MCST LCC compiler identification is now changed to LCC,
there should be a way for old projects to still identify it as GNU,
as it was before.
This commits adds the policy:
CMP0129: Compiler id for MCST LCC compilers is now LCC, not GNU.
This policy controls such a behavior.
OLD behaivior is to treat LCC as GNU, NEW is to treat is as LCC.
Divert LCC compiler as a new one, instead of treating it as GNU.
Since old times, Elbrus C/C++/Fortran Compiler (LCC) by MCST has been
passing checks for GNU compilers, so it has been identified as GNU.
Now, with intent of seriously upstreaming its support, it has been
added as a separate LCC compiler, and its version displays not a
supported GCC version, but LCC version itself (e.g. LCC 1.25.19 instead
of GNU 7.3.0).
This commit adds its support for detection, and also converts basically
every check like 'is this compiler GNU?' to 'is this compiler GNU or
LCC?'. The only places where this check is untouched, is where it
regards other platforms where LCC is unavailable (primarily non-Linux),
and where it REALLY differs from GNU compiler.
Note: this transition may break software that are already ported to
Elbrus, but hardly relies that LCC will be detected as GNU; still such
software is not known.
Upstream ninja commit `ad3d29fb53` (Put builder output through status
interface) from ninja PR 1899 changed the status output from stdout to
stderr. In particular, `ninja: no work to do` is now printed on stderr.
Update our RunCMake tests to accept this difference.
A few RunCMake test cases check for `ninja: no work to do`. For those,
move the message to stdout using `RunCMake_TEST_OUTPUT_MERGE`. The rest
of the test cases do not care about the message, so remove it from the
actual stderr content before comparing against that expected.
The cmLocalGenerator::AddCompilerRequirementFlag was another
source of truth on what standard levels existed per language,
so fold that into cmStandardLevelResolver to make updating
CMake easier
On low-resolution filesystems, `Makefile` can end up with the same
time stamp as `CMakeCache.txt`. CMake must re-run in this case to
be conservative, but that invalidates the `-nowork` cases.
Rather than trying to detect 1s filesystems, just add an option
called `CMake_TEST_FILESYSTEM_1S` that external scripts can use
to tell CMake's test suite about this.
The OLD behaviors of all policies are deprecated, but only by
documentation. Add an explicit deprecation diagnostic for policies
introduced in CMake 3.8 and below to encourage projects to port away
from setting policies to OLD.
Root causes were:
* Using incorrect conditions (assuming MSVC-like command line mode)
* Trying to compile the MSVC STL in C++11 mode, when parts of it require
C++14 or enabling MS extensions in clang.
* Missing flush in a testcase using stdout in a dll and a main part
with static crt
The OLD behaviors of all policies are deprecated, but only by
documentation. Add an explicit deprecation diagnostic for policies
introduced in CMake 3.7 and below to encourage projects to port away
from setting policies to OLD.
* Disable the system include unused variable test in ExportImport when
clang is in MSVC compatible mode.
* Disable CxxDialect testcase when clang is in MSVC compatible mode, as
it doesn't support `typeof`.
* Teach Module.WriteCompilerDetectionHeader to treat clang-cl as MSVC.
* Disable the SystemIncludeDirectories testcase within
IncludeDirectories when clang is in MSVC compatible mode.
* Disable the CMakeOnly.CheckCXXCompilerFlag testcase when clang is in
MSVC compatible mode.
* Treat clang-cl as MSVC in LinkOptions.cmake in the try_run and
try_compile testcases.
The OLD behaviors of all policies are deprecated, but only by
documentation. Add an explicit deprecation diagnostic for some policies
to encourage projects to port away from setting policies to OLD.
Run the `clang-format.bash` script to update all our C and C++ code to a
new style defined by `.clang-format`. Use `clang-format` version 6.0.
* 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.
Brad King