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
{{ message }}
Repository navigation
write_unmatched clears once the pull request it names leaves the open set, so a machine-user write nobody matched closes its own alarm #787
The write was review 5883244267, the Maintainer's hand approval as the machine user, which record 0198 expects to alarm. The alarm cleared one pass after the merge, with nobody matching it. The "Factory alerts" issue that carried it (#786) closed itself at 23:15:27Z. Likely cause, not yet read in the code: reconciliation reads the machine user's writes only against open pull requests, so a write on a pull request that has since merged leaves the window.
This time the write was expected. But a write by the machine user's login that nobody made, followed by a merge, would raise the alarm and close it again within minutes, leaving only the opening notification.
The question to rule on: should write_unmatched stand until a person matches the write (or for the reconciliation window of decision 0112), whether or not its pull request is still open? Or is the opening notification enough? Either way, record the answer. If it should stand, it needs a boundary test and a mutant.
Not urgent: today the only use of the login is the Maintainer's own, by hand.
Activity