1242911 - Bugzilla updated based on an old revision; publish didn't finish

Status: RESOLVED INVALID

(MozReview Graveyard :: General, defect)

Description

[Andreas Pehrson [:pehrsons]]

Attachment: Mozreview error; bugzilla error about deadlock

Not sure how to give you guys solid steps to reproduce here, but something has gone very wrong with https://reviewboard.mozilla.org/r/22565/ for bug 1208371.

Here's what I did.

* I pushed to mozreview, and got bug 1202309.
* After waiting a bit more, things seemed fine, mozreview was showing the expected state.
* I published in the web UI (and this has worked before after encountering bug 1202309).
* I got an error message per the attachment. This has also happened before and been fine, it just takes a little longer for bugzilla to get fully updated.

Expected:
Mozreview to be in published state.
Bugzilla to correctly reflect who has been requested to review what.

Actual result:
Mozreview is in a draft state.
Bugzilla was updated erronously:
* See bug 1208371 comment 574, mozreview requested review from roc and jib for bug 1208371 attachment 8676177 [details], though it shouldn't have requested review at all. An old revision of that attachment had roc and jib as reviewers. mt is listed as reviewer but had r+ed.
* See bug 1208371 comment 579, mozreview requested review from roc and mt for bug 1208371 attachment 8706309 [details], though it should have requested it from mt. mt and bwc are reviewers, bwc had r+ed. roc was reviewer in an older revision.
* See bug 1208371 comment 582 for another similar case.

It also seems it didn't request review for any patches it should have requested it for, or it hadn't gotten there yet, since it's stuck in a draft state still.

How can I get out of the draft state?

And how can I get the reviewers properly set up?

Comment 1

[Andreas Pehrson [:pehrsons]]

Based on mdoglio's advice on IRC I discarded the draft and pushed and published again. It gave an HTTP 500 on publish but eventually finished (took a couple of minutes, much less than I waited the previous time). The stack trace for that 500 error is at bug 1202309 comment 6 aiui.

Comment 2

[BMO Automation]

MozReview is now obsolete. Please use Phabricator instead. Closing this bug.