[webkit-reviews] review granted: [Bug 30098] commit-queue can wrongly reject patches if the buildbots are behind : [Attachment 45834] Patch

[bugzilla-daemon at webkit.org]

Eric Seidel <eric at webkit.org> has granted Adam Barth <abarth at webkit.org>'s
request for review:
Bug 30098: commit-queue can wrongly reject patches if the buildbots are behind
https://bugs.webkit.org/show_bug.cgi?id=30098

Attachment 45834: Patch
https://bugs.webkit.org/attachment.cgi?id=45834&action=review

------- Additional Comments from Eric Seidel <eric at webkit.org>
Seems:
 162		 self._update_status("Unabled to successfully build and test",
None)

should include the revision number that it tried, and should include the
results spew, no?

Also, shouldn't it prefix with ERROR: so that it shows up purple?

Probably should have a comment to explain the --no-update here:
 170		 self.run_bugzilla_tool(["land-attachment", "--force-clean",
"--non-interactive", "--no-update", "--parent-command=commit-queue",
"--build-style=both", "--quiet", patch["id"]])

I think this is definitely better than what we have.  Sad that we'll end up
running the layout tests twice which will slow us down a lot.  Hopefully that
will be moot once run_webkit_tests.py is with us.

I'll say r+, but I think this needs modifications to report the status of the
failed test build better.  That will help us more easily debug flakey tests
(and will make the dashboard correctly show purple).