Sumit Bhardwaj authored 4 years ago and Brad King committed 4 years ago

Previously, cm::static_reference_cast used invoke_result_t and took the
address of O::get. This is not in complete conformance with standard.

This MR changes the implementation to use std::declval<O>.get() which is
always well-defined.

Merge-request: !5105

Jörg Bornemann authored 4 years ago and Brad King committed 4 years ago

Change the default value of `CMAKE_AUTOMOC_PATH_PREFIX` to `OFF` to
restore compatibility with behavior of CMake 3.15 and below.

C++ source files that are generated by Qt's meta object compiler (moc)
include the header file that was passed as input argument to moc. This
is usually a path relative to the source directory, for example

    #include "../../source/dir/myobject.h"

That is problematic for reproducible builds as described in #18815.
To cope with that, the target property AUTOMOC_PATH_PREFIX was
introduced in CMake 3.16 by commit d018d27c (Autogen: Add moc path
prefix generation (AUTOMOC_PATH_PREFIX), 2019-09-13, v3.16.0-rc1~94^2~4).
The property is default-initialized from the variable
`CMAKE_AUTOMOC_PATH_PREFIX`, which defaults to `ON`.

If this property is ON, and myobject.h is located in an include
directory of the target, moc-generated C++ files include the file
without the "path prefix":

    #include "myobject.h"

This behavior, however, can break projects that have equally named
header files in different include directories.  As "not breaking
existing projects" trumps "have reproducible builds by default" we
change the default of `CMAKE_AUTOMOC_PATH_PREFIX` to `OFF`.

Also, it is now possible to pass `-DCMAKE_AUTOMOC_PATH_PREFIX=ON` on the
CMake command line.  Before, it was overridden in `CMakeGenericSystem`.

Fixes: #20598
Issue: #18815

Merge-request: !5108

Avoid interference from `CMAKE_PREFIX_PATH` and friends if they
happen to be set in the environment from which the tests are run.

Merge-request: !5085

When using Qt 5.15.0 or above together with Ninja, check that touching
a source file of a dependency does not needlessly re-run AUTOMOC for
the dependee target.

The AutoMoc timestamp creating custom command explicitly depended
on all dependencies of the origin target (associated to the AutoGen
target).

When an origin target depended on a shared library 'libfoo.so',
if it was re-linked, the AutoMoc custom command would touch its
output timestamp file, and thus cause needless rebuilding of sources,
despite the shared library not having any influence on the AutoMoc
generated files.

Introduce a new '<target>_autogen_timestamp_deps' utility target,
which will serve as an 'order-only' dependency for the custom command.

This will prevent needless rebuilding, because touching 'libfoo.so'
will not cause the custom command to be re-executed.

The new AutoMoc dependency tree looks like:
    '_autogen_timestamp_deps (serves as order-only dep)'
 <- '<target_autogen>/timestamp' file ( + moc deps file)
 <- '<target>_autogen' target.

Fixes: #21020

Merge-request: !5094

Kyle Edwards authored 4 years ago and Brad King committed 4 years ago

In commit 99ed39b0 (Ninja Multi-Config: Make link response files
per-config, 2020-07-15, v3.17.4~3^2), we added the target directory to
the response file under the mistaken assumption that two different
targets with the same name could be in different directories. However,
this causes the path to the response file to be too long to fit on a
command line.  Take the path back out, while leaving in the per-config
split.

Fixes: #21050

Merge-request: !5057

On systems with older GNU system compilers, the Intel C++ compiler does
not define `__cplusplus` to any version newer than C++11.  This
prevented `bootstrap` from detecting that a given C++ standard flag has
enabled C++17 mode in the compiler.  In commit 033a4b12 (bootstrap:
Extend C++17 check for our cast functions, 2019-12-14,
v3.17.0-rc1~291^2) we added a preprocessor condition to attempt to
detect C++17 mode in the Intel compiler on such systems by looking
for `__cpp_if_constexpr`.  However, on systems with a modern GNU
system compiler, that definition is available even in C++11 mode.

Switch to using `__cpp_deduction_guides` to detect C++17 mode for the
Intel C++ compiler.  That seems to be defined exclusively in C++17 mode
regardless of the version of the system compiler.

Fixes: #21013

Merge-request: !5039

Stephan Rohmen authored 4 years ago and Brad King committed 4 years ago

Issue: #20928

Stephan Rohmen authored 4 years ago and Brad King committed 4 years ago

The behaviors controlled by options `GRAPHVIZ_GENERATE_PER_TARGET` and
`GRAPHVIZ_GENERATE_DEPENDERS` were broken by commit 55365839 (Graphviz:
added test suite, fixes, enhancements, 2019-10-08, v3.17.0-rc1~615^2).
It had not been covered in the test suite previously, and those changes
left out checks for these features from the `default_options` case.

Implement the previously-existing behavior in the new graphviz
generation engine added by the above-mentioned commit.

Fixes: #20928

Merge-request: !5020

Fixes: #20961

Merge-request: !4860

Billy Brumley authored 4 years ago and Brad King committed 4 years ago

Fix the regex syntax added by commit 61d746e5 (FindOpenSSL: Detect
OpenSSL 3.0.0, 2020-05-27, v3.17.3~1^2).  Add missing escapes.
Test with `openssl-3.0.0-alpha3`.

While at it, also unset a temporary variable after use.

Brad King authored 4 years ago and Kitware Robot committed 4 years ago

10c88c43 PCH: Do not enable GNU or Intel PCH settings for Fortran

Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !4843

Brad King authored 4 years ago and Kitware Robot committed 4 years ago

b69010b7 VS: Fix .sln support for VS Version Selector with VS 2019

Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !4844

VS 2019 changed the naming pattern used by 2015 and 2017.

Fixes: #20783

The PCH settings are shared by C and CXX languages but do not make sense
for Fortran.  In particular, `CMAKE_PCH_EXTENSION` should not be set
because it can overwrite the value set for C/C++ languages, which may
have a different compiler vendor than the Fortran compiler.

Fixes: #20752

Brad King authored 4 years ago and Kitware Robot committed 4 years ago

90285836 Merge branch 'backport-3.16-pch-fix-bad-ClearSourcesCache'
fa7b041e PCH: Fix logic error that incorrectly clears sources during VS generation

Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !4815

Brad King authored 4 years ago and Kitware Robot committed 4 years ago

6fc4bfa1 Ninja Multi-Config: Fix bug in CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS

Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Alex Reinking <alex_reinking@berkeley.edu>
Merge-request: !4825

Fixes: #20775

Brad King authored 4 years ago and Kitware Robot committed 4 years ago

2c0db404 FindSubversion: Do not accept macOS stub without Xcode implementation

Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !4813

Merge-request: !4815

Since commit 729d997f (Precompile Headers: Add REUSE_FROM signature,
2019-08-30, v3.16.0-rc1~101^2), `GetPchFileObject` handles the case that
it is called first for another target's `REUSE_FROM` by calling
`AddSource` to make sure `GetObjectName` can produce the correct object
name.  However, `AddSource` causes `ClearSourcesCache` to be called,
which since commit a9f4f58f (cmGeneratorTarget: Clear AllConfigSources
in ClearSourcesCache, 2020-05-15, v3.16.7~2^2) now correctly erases the
`AllConfigSources` structure.  This is okay during `AddPchDependencies`,
but there is another code path in which it is problematic.

When the Visual Studio generator's `WriteAllSources` method is looping
over the sources, the `cmake_pch.cxx` source is encountered first.  This
causes `OutputSourceSpecificFlags` to call `GetPchCreateCompileOptions`,
which calls `GetPchFile`, which under MSVC with `CMAKE_LINK_PCH` calls
`GetPchFileObject`.  That leads to `ClearSourcesCache` erasing the
structure over which `WriteAllSources` is iterating!

This bug is caught by our `RunCMake.PrecompileHeaders` test when run
with the VS generator as of the commit that exposed it by fixing
`ClearSourcesCache`.  However, that change was backported to the CMake
3.16 series after testing only with later versions versions that contain
commit a55df204 (Multi-Ninja: Add precompile headers support,
2020-01-10, v3.17.0-rc1~136^2).  By adding proper multi-config support
for PCH, that commit taught `cmLocalGenerator::AddPchDependencies` to
call `GetPchFile` with the real set of configurations instead of just
the empty string.  This allows the `GetPchFile` cache of PCH sources to
be populated up front so that the later calls to it in the
`WriteAllSources` loop as described above do not actually call
`GetPchFileObject` or `ClearSourcesCache`.  That hid the problem.

Fix this by re-ordering calls to `AddPchDependencies` to handle
`REUSE_FROM` targets only after the targets whose PCH they re-use.
Remove the now-unnecessary call to `AddSource` from `GetPchFileObject`
so that `ClearSourcesCache` is never called during `WriteAllSources`.
Update the PchReuseFrom test case to cover an ordering of targets that
causes generators to encounter a `REUSE_FROM` target before the target
whose PCH it re-uses.

Fixes: #20770

Xcode no longer provides a `svn` implementation, but the `/usr/bin/svn`
stub may still exist.

Brad King authored 4 years ago and Kitware Robot committed 4 years ago

61d746e5 FindOpenSSL: Detect OpenSSL 3.0.0

Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !4805

Brad King authored 4 years ago and Kitware Robot committed 4 years ago

6ba84216 CPack-deb: don't add a line with a dot to pkg desc

Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !4806

Currently, if the package description ends with a newline
(typically if it is read from a file) cpack -deb adds a single line
with a dot at the end which leads to a violation of the
`extended-description-contains-empty-paragraph` debian policy.

This commit fixes the above behaviour.

Fixes: #20763

Vitezslav Cizek authored 4 years ago and Brad King committed 4 years ago

The OpenSSL versioning is changing with the upcoming 3.0.0 release.
https://www.openssl.org/blog/blog/2018/11/28/version/
Since 3.0.0, the patch letters are being dropped. The new format is:
MAJOR.MINOR.PATCH

The OPENSSL_VERSION variable can now be directly derived from the new
OPENSSL_VERSION_STR macro.
https://www.openssl.org/docs/manmaster/man3/OPENSSL_VERSION_NUMBER.html