Bug 111514 – Layout Test plugins/plugin-clip-subframe.html is failing

• Component: Tools / Tests
• Status: RESOLVED FIXED

[Hayato Ito]

The following layout test is failing on WebKit Mac and Win.

plugins/plugin-clip-subframe.html

Dashboard:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=plugins%2Fplugin-clip-subframe.html

Probable cause:
I cannot tell a exact revision from the dashboard.

The test was introduced 6 days ago in:
   http://trac.webkit.org/changeset/144236  Plugin in iframe may not display

[Hayato Ito]

Updated a chromium TestExpectations:

http://trac.webkit.org/changeset/144869
webkit.org/b/111514 [ Mac ] plugins/plugin-clip-subframe.html [ Failure ]

[John Bauman]

Created attachment 192170 [details]
Patch

[Eric Seidel (no email)]

We should not need to force layouts from JavaScript.  If that does de-flake this, then we're missing a layout somewhere!

[John Bauman]

Ok, I'm not certain this deflakes it on Win7 (dbg)(2) (as I can't reproduce the flakiness myself), but it seemed likely that this would fix the issue. I figured that it's probably doing layout at paint time, and since I'm using setTimeout in the test that means it's not doing a second paint/layout by the time the test finishes. I don't know if there's a better explanation.

[Ryosuke Niwa]

*** Bug 112324 has been marked as a duplicate of this bug. ***

[Ryosuke Niwa]

I've confirmed that this test change fixes the bug 112324.

If this signifies a real bug in WebCore, please file a separate bug and fix it there. It's not acceptable to have a flaky test like this one.

[Ryosuke Niwa]

In fact, the bug 112324 might have been the culprit of Mac WK2 EWS being so behind.

[John Bauman]

Actually, I'm not certain that this patch fixes the type of flakiness found in bug 1123234. Are you sure you weren't just lucky when testing it? I'll make a patch that's certain to fix that flakiness, just in case this one doesn't help.

[Ryosuke Niwa]

(In reply to comment #8)
> Actually, I'm not certain that this patch fixes the type of flakiness found in bug 1123234. Are you sure you weren't just lucky when testing it? I'll make a patch that's certain to fix that flakiness, just in case this one doesn't help.

Yeah. I was seeing the 100% sample failure rate on my MBP with Mac WK2. With this change, its sample failure rate is 0%.

[WebKit Review Bot]

Comment on attachment 192170 [details]
Patch

Clearing flags on attachment: 192170

Committed r145794: <http://trac.webkit.org/changeset/145794>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Ryosuke Niwa]

You're right. plugins/plugin-clip-subframe.html is still flaky on some bot :(