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 Ninja and Makefile generators were adding a space to the value both
before and after expanding rule variables. Only the latter is needed.
While at it, revise some outdated comments since the rule variable
expansion is no longer responsible for inserting the launcher.
Suggested-by: Mate Pek <mate.pek@gmail.com>
Fixes: #18340
The format for the TAG file was changed in 3.12, and the way it was
read caused a regression which changed how the track was decided. This
commit fixes the regression.
Fixes#18347.
If two imported targets in different directories have the same name we
should still be able to propagate transitive usage requirements from
both. Fix the DAG checker to work with target pointers instead of
target names since the pointers will not be duplicated even if the names
are.
Fixes: #18345
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
All callers were constructing with a non-empty target name using the
target whose pointer was passed anyway. Drop this argument. Simplify
logic accordingly. Re-order constructor arguments to match the
cmCompiledGeneratorExpression::Evaluate arguments.
Also remove unnecessary getters.
If two imported targets in different directories have the same name we
should still be able to propagate transitive link dependencies from
both. Fix the target and link dependency analyzers to de-duplicate
targets using target pointers rather than target names since the
pointers will not be duplicated even if the names are.
Issue: #18345
Avoid exposing the item name implicitly as std::string. When the item
is a target, avoid storing a second copy of its name.
Most link item construction is paired with calls to `FindTargetToLink`
to get the possible target pointer. Rename these methods to
`ResolveLinkItem` and refactor them to construct the entire item.
The custom command de-duplication added by commit v3.12.0-rc1~171^2 (VS:
Generate a custom command only in the least dependent target,
2018-03-23) accidentally also applied to the `CMakeLists.txt` file
reference we put in each target. This file reference comes with a
custom command that has no dependencies and that is safe to run
repeatedly across multiple targets (via internal stamp checking).
Therefore it should be excluded from the de-duplication so that
`CMakeLists.txt` references appear in all targets for human reference.
Fixes: #18310
Text inputs for "where is the source code" and "where to build the
binaries" had different amount of left spacing, so the two paths were
not aligned. This could create a feeling that you typed something wrong
in the input even though the paths were identical or had a common root
path.
Replace use of std::string allocation, std::set lookups, and toupper
conversions with explicit manual logic to do case-insensitive
recognition of a specific set of strings.
Brad King