00:11
<ljharb>
i had to revert 3006; the biblio task failed on it because it was trying to publish a sha that had already been published previously. that should probably be a silent success condition?
00:17
<bakkot>
uhhh the task uses the commit number, not the sha
00:17
<bakkot>
and the only way for the commit number to be reused is if you force-push to main, I think? reverts count as commits for the purposes of rev-list --count
00:18
<bakkot>
if you didn't force-push, I don't know what's going on there
00:33
<ljharb>
ah ok then that's what did it, i did force push, to remove my first attempt to fix it
00:34
<ljharb>
won't be doing that again :-)
02:19
<jmdyck>
pulling in the new merges, finding some bugs.
03:57
<jmdyck>
actually only one bug.
04:12
<jmdyck>
and one new inconsistency
04:31
<jmdyck>
https://github.com/tc39/ecma262/pull/3017
05:41
<ljharb>
jmdyck: if there's any chance of making a github action to detect these things on PRs - even if it remained optional - i think it'd be very helpful
14:12
<jmdyck>
If it's optional, is there anything to prod the PR-owner to check the results?
14:20
<jmdyck>
Is there a way to set things up so that the results go to me, and then I can decide if a reported error is due to a problem in my code or a problem in the PR?
16:11
<ljharb>
they’ll still see the failures - and certainly part of the action could be to notify you
18:47
<jmdyck>
So they'd still get a red-X failure, but instead of "Required" in an oval, it'd say "Optional"?
20:46
<ljharb>
it's just say nothing, since only things with "required" are required, but yep