Bug 207387 – [ews] add commit-queue build step to close the bug

• Component: Tools / Tests
• Status: RESOLVED FIXED

[Aakash Jain]

[ews] add commit-queue build step to close the bug (after the patch has been committed to repository).

[Aakash Jain]

Created attachment 390088 [details]
Patch

[Aakash Jain]

Sample run: https://ews-build.webkit-uat.org/#/builders/26/builds/428

[Jonathan Bedard]

Comment on attachment 390088 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=390088&action=review

> Tools/BuildSlaveSupport/ews-build/steps.py:562
> +            self._addToLog('stdio', 'bug_id build property not found.\n')

Seems like we're duplicating this one a lot. Maybe add a 'has_bug_id' step to BugzillaMixin in a separate patch?

[Aakash Jain]

Comment on attachment 390088 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=390088&action=review

>> Tools/BuildSlaveSupport/ews-build/steps.py:562
>> +            self._addToLog('stdio', 'bug_id build property not found.\n')
> 
> Seems like we're duplicating this one a lot. Maybe add a 'has_bug_id' step to BugzillaMixin in a separate patch?

I am actually not sure if we should even have this additional check or not, theoretically this shouldn't ever be hit. ews django app does send both bug_id and patch_id with the patch to buildbot.

Also, if the patch_id is missing, ValidatePatch step will halt the build. For bug_id, it's theoretically possible to have missing bug_id though (e.g.: for security patch, if ews django app didn't send the bug_id for some reason).

However, it's possible for someone to define a new factory and add this build-step in some other order/manner, so it's probably fine to have this additional check.

[Aakash Jain]

Committed r256031: <https://trac.webkit.org/changeset/256031>

[Radar WebKit Bug Importer]

<rdar://problem/59263355>