On ZHF:
A similar point was raised here as part of the previous retrospective, but as a newcomer willing to contribute a few hours to ZHF I was turned away after fixing a single package due to UX shortcomings:
-
It’s hard to find something I’m able to fix, having filtered the jobset by architecture:
- Some failures are intermittent (ie. due to timeouts), with no way to easily discern these from jobs which have been failing for months. Ideally one should be able to request a rebuild or at least be able to mark them them as an transient failure.
- There’s no way to find jobs with failures I’m more familiar with or feel like looking into (e.g. C build failures).
- Some failures appear related but clutter the search results (ie. a huge list of kernel modules).
-
Hydra being fairly slow doesn’t help either.
I wonder whether it would be useful to be able to somehow tag build failures (at least during ZHF) - most likely would require a triage step, but being able to search for labels might help.