GitHub is sending us PR synchronize event for a PR that does not exist.

[Mariatta]

Same issue as described in python/the-knights-who-say-ni#148.
Bedevere is attempting to apply a label to such PR, causing gidgethub.BadRequest: Not Found error.

bedevere/bedevere/bpo.py

Lines 35 to 42 in 7cd16ef

	@router.register("pull_request", action="opened")
	@router.register("pull_request", action="synchronize")
	async def set_status(event, gh, *args, **kwargs):
	"""Set the issue number status on the pull request."""
	issue_number_found = ISSUE_RE.search(event.data["pull_request"]["title"])
	if not issue_number_found:
	issue = await util.issue_for_PR(gh, event.data["pull_request"])
	status = SKIP_ISSUE_STATUS if util.skip("issue", issue) else FAILURE_STATUS

Perhaps we can catch the error here, so I won't keep getting error alerts.
Thanks.

[brettcannon]

Yep, if GitHub gives us bad data then we should log that fact and then move on.

[Mariatta]

🍝 (copy pasta 😄)

Update from GitHub:

PR 5581 exists, but it just has been marked as "suspicious/spammy", so we and the bots can't access it. The PR was still open and we were receiving the "synchronize" events, breaking the bots.

I have asked GitHub to close the PR, and it has been closed now.

We can still improve the code though if this ever happens again :)

[lysnikolaou]

@brettcannon @Mariatta How do you think the right approach for this would be? Should we wrap every call to post with a try/except, where we log the event (which would be a bigger change) or maybe just do that for the code in set_status?

[brettcannon]

@lysnikolaou this definitely isn't a priority to fix, but yes, a well-documents try/except with a tight except clause would be the way to handle it.