http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=css3%2Ffilters%2Fcustom%2Feffect-custom-parameters.html I'm not sure who the best person to look into this is, or if there's anything much we can do about it (or if they're just rounding errors).
We can add a Linux 32 baseline. It looks like a rounding difference. I guess we could just mark LINUX DEBUG as IMAGE.
I think this is related to https://bugs.webkit.org/show_bug.cgi?id=84813. The color difference is expected from one GPU to the other.
*** Bug 84813 has been marked as a duplicate of this bug. ***
Created attachment 153540 [details] Patch
Stephen, do these look OK to you?
Comment on attachment 153540 [details] Patch Hah, that diff is not so useful. Let me try again.
Created attachment 153542 [details] Patch
Oh, I bet this is related to the bugzilla server upgrade. I'll file a new bug for it.
(In reply to comment #5) > Stephen, do these look OK to you? Alex would know better if the pixel results are correct (once we can see them). The TestExpectations changes look fine.
(In reply to comment #9) > (In reply to comment #5) > > Stephen, do these look OK to you? > > Alex would know better if the pixel results are correct (once we can see them). The TestExpectations changes look fine. Yes, I could check them once I get to see them. I will try to apply the patch on my machine instead.
(In reply to comment #10) > (In reply to comment #9) > > (In reply to comment #5) > > > Stephen, do these look OK to you? > > > > Alex would know better if the pixel results are correct (once we can see them). The TestExpectations changes look fine. > > Yes, I could check them once I get to see them. I will try to apply the patch on my machine instead. It should be working now. You may have to press reload on the Review Patch link.
(In reply to comment #11) > It should be working now. You may have to press reload on the Review Patch link. It doesn't work yet for me, but I've applied it on my machine. The images look fine. It seems like color correction is just a little bit different, but the functionality that those tests are asserting is working correctly.
Created attachment 153552 [details] Patch for landing
Comment on attachment 153552 [details] Patch for landing Rejecting attachment 153552 [details] from commit-queue. Failed to run "['/mnt/git/webkit-commit-queue/Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '-..." exit_code: 1 ERROR: /mnt/git/webkit-commit-queue/LayoutTests/ChangeLog neither lists a valid reviewer nor contains the string "Unreviewed" or "Rubber stamp" (case insensitive). Full output: http://queues.webkit.org/results/13307330
Created attachment 153555 [details] Patch for landing
Comment on attachment 153555 [details] Patch for landing Clearing flags on attachment: 153555 Committed r123252: <http://trac.webkit.org/changeset/123252>
All reviewed patches have been landed. Closing bug.