Bug 255675 - [REGRESSION] Reverting a patch no longer posts anything in the original bug
Summary: [REGRESSION] Reverting a patch no longer posts anything in the original bug
Status: NEW
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Nobody
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2023-04-19 13:30 PDT by Myles C. Maxfield
Modified: 2023-04-26 08:10 PDT (History)
3 users (show)

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Myles C. Maxfield 2023-04-19 13:30:23 PDT 
https://bugs.webkit.org/show_bug.cgi?id=255473

This patch was reverted, but Fuji manually commented in it and reopened it themself. WebKitBot didn't comment in the bug _at all_.

When WebKitBot reverts a patch, it should _always_ comment in the bug of the patch it reverted.
Comment 1 Radar WebKit Bug Importer 2023-04-19 13:30:57 PDT 
<rdar://problem/108277861>
Comment 2 lingho@apple.com 2023-04-25 18:41:40 PDT 
I don't think webkitbot reopens a bug. I see cases where bug is reopened by Commit Bot.

webkitbot uses the "create-revert" option of webkit-patch consisting of the following steps:

        steps.DiscardLocalChanges,
        steps.Update,
        steps.RevertRevision,
        steps.CreateBug,
        steps.PrepareChangeLogForRevert,
        steps.PostDiffForRevert,

There is another webkit-patch option "revert" which consist of 
        steps.DiscardLocalChanges,
        steps.Update,
        steps.RevertRevision,
        steps.PrepareChangeLogForRevert,
        steps.EditChangeLog,
        steps.ConfirmDiff,
        steps.Build,
        steps.Commit,
        steps.ReopenBugAfterRevert,
Comment 3 Alexey Proskuryakov 2023-04-26 08:10:09 PDT 
This is generally happening in practice though.

https://bugs.webkit.org/show_bug.cgi?id=255610 was a revert bug created by webkitbot. It was filed at 10:02:18 PDT, and https://bugs.webkit.org/show_bug.cgi?id=255506 got reopened three seconds later. Commit queue only landed the revert at 10:15:04 PDT.

Looking at the original case that motivated this report, https://bugs.webkit.org/show_bug.cgi?id=255473, I see that Fujii reopened it manually BEFORE creating the revert bug. Webkitbot did add a dependency relationship.

So, seems like correct behavior - it would be wrong for webkitbot to say that it reopened the bug when it didn't. Remember that the comment says "Re-opened since this is blocked by <...>". Myles, is your expectation different?

> When WebKitBot reverts a patch, it should _always_ comment in the bug of the patch it reverted.

Note however that webkitbot does not comment after reverting, it only ever comments when creating a revert bug. I think that this is wrong, both commenting and reopening should happen when the revert lands. But maybe I'm missing why it's being done this way. In any case, there is no regression here, and we'd want a new bug report to investigate improvements.