The CTestTestStopTime test has been failing sporadically because the
stop time causes the first internal test to have a timeout short enough
that we might hit it and start the second test just before the stop time
is reached. Instead we should track when a timeout is shortened in
order to stay within the stop time. If a test times out for this reason
then we should consider the stop time reached and not start any more
tests.
The `ctest --test-load` option is implemented in `StartNextTests` by not
starting any tests when the load is too high and instead sleeping and
then returning. Prior to commit v3.11.0-rc1~117^2 (CTest: Re-implement
test process handling using libuv, 2017-12-10) our outer loop in
`RunTests` would immediately call `StartNextTests` again. However, now
the `uv_run` loop may simply terminate if there are no tests running
because no events are left pending.
Fix this by converting the sleep in `StartNextTests` into a libuv timer
that it starts instead. This avoids leaving `uv_run` with no pending
events. In the case that there are other running tests this also allows
CTest to detect when they finish even if it during the wait period where
we previously slept.
This regression was not caught by the test suite because it only
verified that we do not start new tests when the load was too high and
not that we proceed to start tests when the load drops. Revise the test
suite to cover both.
Fixes: #18338
Sphinx-build warnings look like this:
../path/to/file.rst:105: WARNING: This is not an error.
CTest detects these as errors by this regex:
([^ :]+):([0-9]+): ([^ \\t])
Add an exception as we already have for lower-case `: warning`.
Sanitizers do not create a log file when no defects are found. Therefore,
it is currently impossible for ctest_memcheck to set both
`CAPTURE_CMAKE_ERROR` and `RETURN_VALUE` to zero.
With defects, `CAPTURE_CMAKE_ERROR`=0 and `RETURN_VALUE`=-1, as expected.
With no defects, `CAPTURE_CMAKE_ERROR`=-1 and `RETURN_VALUE`=0.
After changing the ``cmGeneratedFileStream`` methods to accept
``std::string const&`` instead of ``const char*`` we don't
need to call ``std::string::c_str`` anymore when passing
a ``std::string`` to a ``cmGeneratedFileStream`` method.
This patch removes all redundant ``std::string::c_str``
calls when passing a string to a ``cmGeneratedFileStream`` method.
It was generated by building CMake with clang-tidy enabled using
the following options:
-DCMAKE_CXX_CLANG_TIDY=/usr/bin/clang-tidy-4.0;-checks=-*,readability-redundant-string-cstr;-fix;-fix-errors
71f27c04b0 ctest_submit: show headers as debug info
4464ef21da ctest_submit: Check CA file during CDASH_UPLOAD
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2171
Hide any headers specified by `ctest_submit(HTTPHEADER ...)` unless ctest
is run with --debug. This option is used to present bearer tokens to CDash.
We do not want to inadvertently expose this secret data in the console
output log.
This fixes a bug where ctest(SUBMIT CDASH_UPLOAD ...) would fail when
attemping to submit to a site bearing a genuine SSL certificate.
This code path was missing a call to cmCurlSetCAInfo() which loads
the certificate authority file.
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.
While we already support `cmake --build . -- -j`, the options after `--`
are specific to the native build tool. Add new options `--parallel
[<N>]` and `-j [<N>]` to abstract this and map to the proper option
for the native build tool.
This change reworks ctest_start() so that simply calling
ctest_start(APPEND) will read all the information from the TAG file.
On top of that, it relaxes the argument parsing for ctest_start() to
allow greater flexibility in the argument ordering, and the documentation
for ctest_start() has been cleaned up.
Refactoring in commit v3.9.0-rc1~156^2 (c++: prefer vectors over lists,
2017-05-04) switched `cmCTestSVN::Repositories` from `std::list` to
`std::vector`. This can cause re-allocation when svn externals are
processed and break the `RootInfo` pointer that is supposed to point at
the first repository element. Switch back to `std::list` so that the
address remains stable.
Fixes: #17854
5901699672 cmDepends: Remove attempt to change directory that always fails
e60e4dfc88 cmWorkingDirectory: Check success of current dir changes
e654622aee Tests: Add --build-and-test test case
a865f0beb2 Tests: Confirm test working dir set successfully
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !1817
In commit v2.8.0~170 (ENH: Added ctest test options PROCESSORS and
RUN_SERIAL, 2009-09-07) CTest learned to track the number of processors
allocated to running tests in order to balance it against the desired
level of parallelism. Extend this idea by introducing a new
`PROCESSOR_AFFINITY` test property to ask that CTest run a test
with the CPU affinity mask set. This will allow a set of tests
that are running concurrently to use disjoint CPU resources.
* Change some functions to take `std::string` instead of
`const char*` in the following classes: `cmMakeFile`, `cmake`,
`cmCoreTryCompile`, `cmSystemTools`, `cmState`, `cmLocalGenerator`
and a few others.
* Greatly reduce using of `const char*` overloads for
`cmSystemTools::MakeDirectory` and `cmSystemTools::RelativePath`.
* Remove many redundant `c_str()` conversions throughout the code.
This is to avoid scope issues with CTEST_RUN_CURRENT_SCRIPT. If
ctest_start() is called within a function scope, the value of
CTEST_RUN_CURRENT_SCRIPT that it sets doesn't make it to the global
scope. With this change, ctest_start() no longer sets
CTEST_RUN_CURRENT_SCRIPT, and instead sets a field directly in
cmCTestScriptHandler. The old behavior of CTEST_RUN_CURRENT_SCRIPT
has also been kept for projects and tests that rely on setting it.
This commit continues the changes made in CTest to support std::chrono
by
applying it throughout every component where a duration was used.
No functional change intended.
A problem area by recent refactoring of time to std::chrono has been the
unsafe conversion from duration<double> to std::chrono::seconds, which
is of an unspecified integer type.
This commit adds a template function that for a given type provides a
safe conversion, effectively clamping a duration<double> into what fits
safely in that type. A specialisation for int and unsigned int are
provided.
It changes the protential problem areas to use this safe function.
This commit introduces cmDuration as a typedef for
std::chrono::duration<double, std::ratio<1>>. It is less verbose and
provides for a point to put future common functionality for durations.
No functional change intended.
The libuv documentation states that the stream read callback may
be called with `nread == 0` for EAGAIN. Handle this gracefully.
It also states that the callee is responsible for closing the
stream on error. Always close the stream for `nread < 0`.
Brad King