| Summary: | Optimize commit-queue performance for green trees | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Adam Barth <abarth> | ||||||
| Component: | New Bugs | Assignee: | Adam Barth <abarth> | ||||||
| Status: | RESOLVED FIXED | ||||||||
| Severity: | Normal | CC: | eric | ||||||
| Priority: | P2 | ||||||||
| Version: | 528+ (Nightly build) | ||||||||
| Hardware: | Other | ||||||||
| OS: | OS X 10.5 | ||||||||
| Attachments: |
|
||||||||
|
Description
Adam Barth
2010-09-22 02:21:15 PDT
Created attachment 68352 [details]
work in progress
I need to update existing tests and add new tests. Factoring out the CommitQueueTask object should allow for much better testing of this logic. Comment on attachment 68352 [details] work in progress View in context: https://bugs.webkit.org/attachment.cgi?id=68352&action=review This patch is absolutely fantastic. > WebKitTools/Scripts/webkitpy/tool/bot/commitqueuetask.py:136 > + if not self._patch.is_rollout(): > + if not self._test(): > + if not self._test(): > + if not self._build_and_test_without_patch(): > + return False > + raise self._script_error This block is kinda confusing. I think it may need some comments to annotate what its doing. We could also have a _validate_and_land() command which it called instead, and reverse the logic. But it's kinda nice having the ifs be always for the failure cases. Created attachment 68477 [details]
Patch
Comment on attachment 68477 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=68477&action=review As far as I can tell this is "right" :) > WebKitTools/Scripts/webkitpy/tool/bot/commitqueuetask.py:41 > + # commit-queue is processing. Do one last minute check before landing. this is no longer last minute. we can remove that part Committed r68102: <http://trac.webkit.org/changeset/68102> |