You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Sometimes Robosignatory is really backed up (e.g. if texlive) and it could take a while for the builds to be signed. Instead of just waiting 10 mins, since we're a persistent service we should just asynchronously wait forever until the builds are signed so that we can issue the follow-up distRepo.
The text was updated successfully, but these errors were encountered:
I think I was originally against waiting forever, especially since the next run would pick it up anyway. If it failed (rather than wait forever) and we actually had some sort of notification system set up then we could investigate.
I could see the argument for making it wait forever (or even just wait much longer), but am interested in specifically what problem we're trying to solve?
I often see pipeline builds fail because the tagger (signing/distrepo) took too long. Would be nice if we could make those wait if there is an existing distrepo running or something. I guess that won't really solve the signing problem either. Maybe:
Sometimes Robosignatory is really backed up (e.g. if texlive) and it could take a while for the builds to be signed. Instead of just waiting 10 mins, since we're a persistent service we should just asynchronously wait forever until the builds are signed so that we can issue the follow-up distRepo.
The text was updated successfully, but these errors were encountered: