1851aa49be FetchContent: Pass through networking-related CMAKE_... variables
96937438b7 Help: Clean up how TLS and NETRC variables are discussed
2a82bd85b6 Help: Add documentation for CMAKE_TLS_CAINFO
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !6589
CMAKE_TLS_VERIFY, CMAKE_TLS_CAINFO, CMAKE_NETRC and
CMAKE_NETRC_FILE are now passed through to the underlying
ExternalProject sub-build. Previously, they were silently ignored.
Fixes: #22144
The minimum CMake version for Qt6 is 3.16, so all the calls to
cmake_minimum_required() are updated here to enforce that
minimum. This will avoid any CMake version-related warnings
from Qt.
Avoid hard-coding Qt5 where the tests could now be using
Qt5 or Qt6.
Fixes: #22188
The changes are part of CMP0128.
When the standard level is unset:
* Flags are added if extension mode doesn't match the compiler's default.
Previously logic only worked if LANG_EXTENSIONS was ON. Fixes#22224.
* The full flag is used. Previously CMAKE_LANG_EXTENSION_COMPILE_OPTION was
used. This was only supported for IAR.
Otherwise:
* Avoid adding flags if not necessary per the detected compiler defaults.
* Fixed check for when the requested standard is older. It now matches the
nearby comments.
I reworded the fallback comment as its logic was a bit difficult to wrap my
head around.
On Windows, imported shared libraries which only have an
`IMPORTED_IMPLIB` set fail if they are depended upon by a target using
automoc. Add a test for the behavior of depending upon all imported
target types from an automoc-using target.
The generated path with the packages uses $CPACK_TOPLEVEL_TAG, which
by default is $CPACK_SYSTEM_NAME, thus the OS name.
To make the expected stderr match also non-Linux OSes, accept any
non-slash characters in place of "Linux", so it works also on other
Debian OSes (e.g. Debian/Hurd).
The pch object file could cause problems when the reused pch is passed
through an OBJECT library, which would use INTERFACE_LINK_LIBRARIES to
link the pch object file.
Fixes: #22630
Extend the feature added by commit 2ce42f281f (VS: Add VS_SETTINGS
source file property, 2020-03-18, v3.18.0-rc1~449^2~3) to support
all source file types.
b9698f89df cmake_host_system_information: Make it available for all systems
5469c71a82 Refactor: Simplify `GetValue()` function calls
6c92f80f2e cmake_host_system_information: Also set `USED_FALLBACK_SCRIPT`
efe139d1b8 cmake_host_system_information: Can run fallback scripts
1e65e4a6e5 cmake_host_system_information: Can read `/etc/os-release` file
e808cbb1dd Testing: Convert `cmake_host_system_information` tests into `RunCMake`
9e831284e5 Documentation: Use definition list instead of tables
346f3de005 Refactor: Deduplicate code for `VS_nn_DIR` keys processing
...
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: buildbot <buildbot@kitware.com>
Merge-request: !6410
eae125ace5 Refactor: Get rid of `isTrue` variable in the `while` block execution
4c1cdfd8f0 Refactor: Keep `cmWhileFunctionBlocker` members private
d22f68d019 Refactor: Transform `while` loop into `for`
e97e714f0d Fix: `while()` reports an error the same way as `if()`
880ca66b51 Fix: `while()` can silently ignore incorrect condition
61b33c3f4e Fix: Regression in the `cmConditionEvaluator::HandleLevel0`
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !6442
As reported in the BUG #22524, mismatched parenthesis reported differently
for `while()` and `if()`.
The problem was in the double loop (over "handlers" and the arguments),
where the outer loop didn't check the result of the running handler.
54a70b3988 Refactor: Move version compare op table out of the function
f37d6a3170 Refactor: Remove redundant checks for `0` and `1` literals
00961a4782 Refactor: Copy exactly required count of args and avoid `pop_back()`
2a72cad9be Style: Better names of local vars for `IN_LIST` handler
0dd7795706 Refactor: Extract the logic of testing for special variable to func
866b0595f6 Refactor: Introduce `cmArgumentList` container class
51d9194a96 Refactor: Reduce one more condition checking on handling math compare
46810235e3 Refactor: Avoid `if` → `else if` → … for compare operators
...
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !6392
The `TEST_FILTER` argument can be used to filter tests during the
discovery phase. It combines `--gtest_filter=<expr>` with the
`--gtest_list_tests` argument when invoking the test excutable for
listing defined tests.
Fixes: #17493
In commit d7159f9591, (Ninja: Fix LINK_WHAT_YOU_USE link flag placement,
2021-06-24), we fixed usage of -Wl,--no-as-needed to override
-Wl,--as-needed from LDFLAGS. However, that change also passes
-Wl,--no-as-needed to `ar`, which is incorrect flag for `ar`.
Let's check the target type before passing down the flags.
Fixes: #22487
aafa392c12 string(TIMESTAMP): Add %V specifier for ISO 8601 week number
a915f691ad Help: Format string(TIMESTAMP) format specifiers as a definition list
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !6374
The `cmArgumentList` has been turned into a class (forward declared in
the header). It inherits from the `std::list` (yeah, but we don't intend
to store polymorphic classes in it). In addition to the standard methods,
now it's possible to move `HandlePredicate` (renamed to `ReduceOneArg`)
and `HandleBinaryOp` (renamed to `ReduceTwoArgs`) as its members.
Additionally, iterators managements (`IncrementArguments`) have been
refactored into two separate classes mimicking iterators. This also
allows having a uniform `for` loop and concentrates the logic of
iterators advancing in it instead of the loop's body. The arguments
processing algorithms operate with "windows" over a collection of
arguments. Hence there are two kinds of "iteration windows" -- allowing
to observe 2 or 3 elements per loop iteration. These iteration "windows"
also passed to reducers.
