Skip to content

Exclude confirmed bugs from stale bot#30921

Merged
QuLogic merged 1 commit into
matplotlib:mainfrom
timhoffm:ci-stale-exclude
Jan 3, 2026
Merged

Exclude confirmed bugs from stale bot#30921
QuLogic merged 1 commit into
matplotlib:mainfrom
timhoffm:ci-stale-exclude

Conversation

@timhoffm

@timhoffm timhoffm commented Jan 2, 2026

Copy link
Copy Markdown
Member

We should not close confirmed bugs through a timeout mechanism. The bug exists and should be tracked as open bug as long as we don't fix it or explicitly decide to not handle this, in which case, we'd manually close
as "won't fix".

@rcomer

rcomer commented Jan 2, 2026

Copy link
Copy Markdown
Member

I appreciate the argument, but I believe this workflow has helped us to close more confirmed bugs "the right way" than automatically. So far, 41 confirmed bugs that were labelled by this workflow have subsequently been closed. Of these, only 13 were closed by the bot (have the "status: closed as inactive" label). The other 28 we either verified no longer exist, we decided they were "won't fix" or we actually fixed them via PR. I argue that we wouldn't have looked at and closed all 28 of these if the bot had not surfaced them, they would still be forgotten in the backlog.

Practically, although this workflow is designed to auto-close, I do not think it ever does in practice1 because we have it set up to look at least recently updated issues first and we also limit it so it doesn't generate too many notifications each run. So we created the stale-tidy workflow to actually do the closing (and remove the label if someone comments). So a compromise might be to have this workflow continue to label and comment on the confirmed bugs (so we get reminded about them) but change stale-tidy to ignore confirmed bugs. The only downside to this that I can see is that the label will not be automatically removed if someone comments after the bot.

Footnotes

  1. if we change days-before-issue-close to -1, it definitely never will

We should not close confirmed bugs through a timeout mechanism. The bug
exists and should be tracked as open bug as long as we don't fix it or
explicitly decide to not handle this, in which case, we'd manually close
 as "won't fix".
@timhoffm

timhoffm commented Jan 2, 2026

Copy link
Copy Markdown
Member Author

So a compromise might be to have this workflow continue to label and comment on the confirmed bugs (so we get reminded about them) but change stale-tidy to ignore confirmed bugs.

Done. This is is the best compromise.

@QuLogic QuLogic merged commit d99834c into matplotlib:main Jan 3, 2026
38 checks passed
@QuLogic QuLogic added this to the v3.11.0 milestone Jan 3, 2026
@timhoffm timhoffm deleted the ci-stale-exclude branch January 13, 2026 15:36
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants