- Add upstream testsuite tests/Regression/upstream-testsuite
to the CI testplan. However, this one isn't a good measure
for gating, because it is too complex and flaky. But it's
good to have test results at hand. This testcase reports a
failure only in case of an infrastructure problem. Individual
.exp test failures are not a show stopper from the persp of
the CI gating.
- Add upstream systemtap build and install pseudo testcase.
tests/Install/upstream-systemtap to the testplan.
This one is disabled by default (in its main.fmf file). The
point here is to be able to test with upstream version of
systemtap in case a working fedora build isn't available.
This is a helper for manual testing, although if enabled, it
will do its job in the automated CI pipeline too.
- Use FMF tag instead of tier attributes to define the testplan.
- Testcase updates:
* tests/Regression/second-command-not-captured
* tests/Regression/python-probing
Tests are supposed to be executed within defined context, such as e.g.
tmt --root . -c arch=x86_64 -c distro=fedora-38 -c trigger=build run ...
FTR: https://tmt.readthedocs.io/en/stable/spec/plans.html#context
This update sets DEBUGINFOD_URLS based on the provided distro
via the ci.fmf file.
tests/Regression/RHEL6Feature-cpp-methods:
Fix locale/LANG dependent sort collation prob
tests/Regression/python-probing:
Make sure needed debuginfo is in place
tests/Regression/wrong-data-returned-by-nfs-commit-done:
Refactor, improve clranup, simplify
Find new home for downstream RHEL tests. Upstream them. The set of
tests used for fedora gating stays intact: The gating tests are only
those having the tier1 tag set in their main.fmf file. The testplan
plans/ci.fmf filters the others out from gating.
The set of Fedora gating tests stays the same as it was before this
change.
