Long time ago there was an issue about the $subject, filled by @fridex , but I can't find it now (maybe it even was in some previous repo).
I recently realized how difficult is to debug a failing tests for a task if we catch TaskError and turn it into {'status': 'error'}. We should let selinon handle those exceptions (i.e. (re-)raise them when caught or not catch them) instead of masking them into {'status': 'error'}.
A change similar to this should IMHO be done also in CVEchecker.py, dependency_snapshot.py, githuber.py, mercator.py (and maybe others)
Cannot find it, but I think it was on one of the internal issue trackers. IMHO it could simplify rescheduling failed tasks and possible task failure tracking/notifications.
Long time ago there was an issue about the $subject, filled by @fridex , but I can't find it now (maybe it even was in some previous repo). I recently realized how difficult is to debug a failing tests for a task if we catch
TaskError
and turn it into{'status': 'error'}
. We should let selinon handle those exceptions (i.e. (re-)raise them when caught or not catch them) instead of masking them into{'status': 'error'}
. A change similar to this should IMHO be done also in CVEchecker.py, dependency_snapshot.py, githuber.py, mercator.py (and maybe others)