Closed Flamefire closed 3 months ago
Test report by @Flamefire
Build succeeded for 14 out of 14 (14 easyconfigs in total) n1625 - Linux RHEL 8.9 (Ootpa), x86_64, Intel(R) Xeon(R) Platinum 8470 (icelake), Python 3.8.17 See https://gist.github.com/Flamefire/a68a4257fd9a053a76a855671f68157e for a full test report.
Was a bit too quick here... If we use tempfile.mkdtemp outside of builddir we need to delete it afterwards. So should unconditionally remove the basedir of impi_testexe after.
Was a bit too quick here... If we use tempfile.mkdtemp outside of builddir we need to delete it afterwards. So should unconditionally remove the basedir of impi_testexe after.
The base directory for that is easybuild-tmp
which IIRC will be deleted after the build or at least after the eb run. Isn't that enough?
If it isn't we should only delete the tmp dir, not the builddir as one might want to inspect that as before via --disable-cleanup-build
Ah yes, forgot that TMPDIR is set to eb controlled space. Blaming first week after vacation for brain not being fully engaged.
Going in, thanks @Flamefire!
(created using
eb --new-pr
)During sanity-check-only the
prepare_step
is skipped which sets up the toolchain including toolchain and environment variables such as $CCWithout that step (i.e. in sanity-check-only mode)
os.getenv('CC')
will fail because the variable is not set. Or even worse: It is set but to a wrong value.Additionally
ec['parallel']
is also not set yielding ampirun -n None
which of course fails.This PR adds the prepare_step and set_parallel call to initialize those. It also uses a temporary directory for the test binary because the builddir also doesn't exist.