It's common for the last activity on a PR to be irrelevant:
- comment from triage asking for status
- author asking for status
- rustbot informing about a merge conflict
- author fixing a merge conflict
Ideally, we would track review activity:
- exclude comments by the author
- exclude rustbot comments
- exclude comments by others with certain keywords like "triage", "status", "bump", etc
- include all activity by the assigned reviewer
- include code-review activity
- exclude drafts
This obviously isn't possible with the basic Github filters available, but I think it would give us a much better idea of how long PRs are sitting without activity. Merge conflict notifications alone likely result in a de-facto upper limit to time-since-updated.
It's common for the last activity on a PR to be irrelevant:
Ideally, we would track review activity:
This obviously isn't possible with the basic Github filters available, but I think it would give us a much better idea of how long PRs are sitting without activity. Merge conflict notifications alone likely result in a de-facto upper limit to time-since-updated.