Draft: Test if empty branches in the output file exist
Add a macro which loops over all defined branches according to the BranchList folder and checks is empty branches exist. Empty branches indicate a problem. Add test to the CI which execute the macro for all created output files.
Merge request reports
Activity
changed milestone to %OCT23
added Build System CI/CD labels
requested review from @p.-a.loizeau
assigned to @p.-a.loizeau
This MR should only be merged after the creation of the MvdDigi is fixed. Currently the the CI pipelines should fail since new test were added which checks if empty branches in the output files exist. In case empty branches are found the test fails since this is an obvious bug. All existing branches should contain data.
added 20 commits
-
b80b35a3...5c7893f5 - 16 commits from branch
computing:master
- 84ce846a - Fix generation of MvdDigi in event based mode
- c47fff72 - Add back event mode in MVD cluster finder
- 8eb8f942 - Fix the reconstruction topologies at several places
- 96b4166e - Test if empty branches in the output file exist
Toggle commit list-
b80b35a3...5c7893f5 - 16 commits from branch
added 113 commits
-
96b4166e...15116b14 - 112 commits from branch
computing:master
- e508ae33 - Test if empty branches in the output file exist
-
96b4166e...15116b14 - 112 commits from branch
added 1 commit
- 03b5d7ca - Test if empty branches in the output file exist
added 54 commits
-
03b5d7ca...ea9125b8 - 53 commits from branch
computing:master
- dbac3f72 - Test if empty branches in the output file exist
-
03b5d7ca...ea9125b8 - 53 commits from branch
added 34 commits
-
dbac3f72...78e217d6 - 33 commits from branch
computing:master
- 02fa7efe - Test if empty branches in the output file exist
-
dbac3f72...78e217d6 - 33 commits from branch
added 5 commits
-
02fa7efe...b5b9540c - 4 commits from branch
computing:master
- 11a0246d - Test if empty branches in the output file exist
-
02fa7efe...b5b9540c - 4 commits from branch
removed milestone %OCT23
added 179 commits
-
11a0246d...39c00e70 - 178 commits from branch
computing:master
- a2db261a - Test if empty branches in the output file exist
-
11a0246d...39c00e70 - 178 commits from branch
I did not think about it before but the two remaining failed tests kind of make sense if the "no empty branch" convention was not made explicit before: in both case some Muon setup reco test fails with
[ERROR] Branch TrdTrack is empty
What is still strange is that this seems to happen only on
CentOs
while I would expect it to be general to all jobsSwitching to
DAILY
mode (by commenting out the cmakeWEEKLY
option), I can reproduce the error.
When comparing the logs, I get some completely different logs already at transport level, so I am wondering now if the problem is not simply that the random generator behaves differently undercentos
fromvae23
, leading to non-comparable results.If you want to have a look:
meld /lustre/cbm/users/ploizeau/apptainer/empties_*_nov22p1_daylies.txt