Bug 58323 – compositing/objects/composited-object-alignment.html causing failures in other tests since its introduction in r83518

• Component: Layout and Rendering
• Status: ASSIGNED

[Jessie Berlin]

http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r83519%20(11485)/compositing/overflow/ancestor-overflow-pretty-diff.html

The error there is coming from the code in LayoutTests/compositing/objects/resources/composited-subframe.html, which is used by composited-object-alignment, and is a bit buggy:
1)It uses setTimeout without first calling layoutTestController.waitUntilDone.
2)It tries to call some nonexistent parent.testDone() function.
3)The platform-specific results are sitting in the cross-platform directory.

This test should be disabled until it is fixed.

[Jessie Berlin]

Disabled the test in http://trac.webkit.org/changeset/83588

[Adam Roben (:aroben)]

<rdar://problem/9271337>

[Chris Marrin]

(In reply to comment #0)
> http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r83519%20(11485)/compositing/overflow/ancestor-overflow-pretty-diff.html
> 
> The error there is coming from the code in LayoutTests/compositing/objects/resources/composited-subframe.html, which is used by composited-object-alignment, and is a bit buggy:
> 1)It uses setTimeout without first calling layoutTestController.waitUntilDone.
> 2)It tries to call some nonexistent parent.testDone() function.

This is cruft I failed to get rid of when changing the test to do a pixel test rather than printing the layer tree. I'll get rid of it.

> 3)The platform-specific results are sitting in the cross-platform directory.

The intent is for all the results to be cross-platform. I avoided text and borders and anything else that would cause differences in the pixel results.

[Chris Marrin]

Looks like pixel results don't match up, at least not between WebKit and WebKit2. So I will just generate platform specific results for all cases

[Chris Marrin]

Created attachment 89224 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 89224 [details]
Patch

Looks like LayoutTests/platform/mac/compositing/objects/composited-object-alignment-expected.png has the wrong mime type.

[WebKit Review Bot]

Comment on attachment 89224 [details]
Patch

Rejecting attachment 89224 [details] from commit-queue.

Failed to run "['./Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '--bot-id=ec2-cq-01', '--port..." exit_code: 2

Last 500 characters of output:
annot stat `LayoutTests/compositing/objects/composited-object-alignment-expected.checksum': No such file or directory
Failed to copy LayoutTests/compositing/objects/composited-object-alignment-expected.checksum LayoutTests/platform/mac/compositing/objects/composited-object-alignment-expected.checksum. at /mnt/git/webkit-commit-queue/Tools/Scripts/svn-apply line 428.

Failed to run "[u'/mnt/git/webkit-commit-queue/Tools/Scripts/svn-apply', u'--reviewer', u'Simon Fraser', u'--for..." exit_code: 2

Full output: http://queues.webkit.org/results/8881917

[Ahmad Saleem]

It is still skipped - https://searchfox.org/wubkat/rev/1dfbd6f7279ea1201b5afb9719b4163c8fab1566/LayoutTests/TestExpectations#3153