Issue metadata
Sign in to add a comment
|
Non-optimal rebaseline result from webkit-patch rebaseline-cl |
||||||||||||||||||||||||
Issue descriptionWhat steps will reproduce the problem? (1) git cl patch --force https://chromium-review.googlesource.com/c/chromium/src/+/929878/4 (2) webkit-patch rebaseline-cl --patchset=4 What is the expected result? The common baseline third_party/WebKit/LayoutTests/http/tests/devtools/layers/layer-canvas-log-expected.txt should be updated because all platforms produced the same new result. What happens instead? The script left the common baseline unchanged (which is stale), and added two new baselines (which are the same) for win and mac10.12. Also saw similar situation when I enabled SlimmingPaintV175 for experimental (https://chromium-review.googlesource.com/c/chromium/src/+/917226). Saw many mac10.12-specific new baselines which looked unnecessary.
,
Feb 23 2018
I just tried the reproduction steps, and the script still create non-optimal baselines. So this bug seems unrelated to the test results server issue.
,
Feb 23 2018
Is the non-existence mac10.13 platform the culprit? The script seems to think that mac10.13 still produce the same old result, so created mac10.12 baseline for tests that actually produced the same results on all platforms.
,
Mar 5 2018
,
Mar 5 2018
Re #3: yes. I think this is issue 812784 . Mac 10.13 trybot has been added to rebaseline-cl recently, which is supposed to fix the problem. Could you try rebasing your local branch (to get the up-to-date rebaseline-cl) and running rebaseline-cl again? If that solves the problem, please mark this issue as a duplicate of issue 812784 . Otherwise, I'll take another look. Thanks!
,
Mar 5 2018
|
|||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||
Comment 1 by martiniss@chromium.org
, Feb 23 2018