Bug 55918 - Make LayoutTests robust to latency btwn playing event and frame rendering
Summary: Make LayoutTests robust to latency btwn playing event and frame rendering
Status: UNCONFIRMED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: Other OS X 10.5
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-03-07 19:39 PST by Victoria Kirst
Modified: 2011-10-19 15:18 PDT (History)
4 users (show)

See Also:

Attachments
Patch (4.40 KB, patch)
2011-03-07 19:55 PST, Victoria Kirst 		no flags Details | Formatted Diff | Diff
Patch (5.08 KB, patch)
2011-03-09 12:57 PST, Victoria Kirst 		no flags Details | Formatted Diff | Diff
Patch (5.14 KB, patch)
2011-03-09 13:14 PST, Victoria Kirst 		commit-queue: commit-queue-
Details | Formatted Diff | Diff
Show Obsolete (2) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Victoria Kirst 2011-03-07 19:39:23 PST 
Make LayoutTests robust to latency btwn playing event and frame rendering
Comment 1 Victoria Kirst 2011-03-07 19:55:08 PST 
Created attachment 85011 [details]
Patch
Comment 2 Victoria Kirst 2011-03-07 19:59:12 PST 
These layout tests depended on the play event firing immediately after a call to play(). I added logic to pause and seek so that the pixel results will always be consistent. Rebaselines will come after this patch is committed.

See crbug.com/74982
Comment 3 Mihai Parparita 2011-03-07 20:18:52 PST 
Comment on attachment 85011 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=85011&action=review

> LayoutTests/media/controls-after-reload.html:14
> +            video.pause();

Do the pause/currentTime lines have to be inside the "if (window.layoutTestController)" blocks? Generally we want layout tests to behave the same way when run inside the DRT as when they're run inside a normal browser.

> LayoutTests/platform/chromium/test_expectations.txt:3059
> +// Temporarily make these tests flaky.

Why is this necessary? If you leave them as having just IMAGE expectations, we'll get notified when they start to unexpectedly pass.
Comment 4 Victoria Kirst 2011-03-09 12:57:52 PST 
Created attachment 85219 [details]
Patch
Comment 5 Victoria Kirst 2011-03-09 13:00:17 PST 
Gah, sorry about that patch, Mihai! I must have been sleepy when I uploaded it. I have fixed the comments you mentioned (and a few other silly things) and uploaded a new one.

I got rid of the GPU modifier on test_expectations.txt because this should affect both software and GPU rendered tests.
Comment 6 Victoria Kirst 2011-03-09 13:14:56 PST 
Created attachment 85221 [details]
Patch
Comment 7 WebKit Commit Bot 2011-03-10 17:57:48 PST 
Comment on attachment 85221 [details]
Patch

Rejecting attachment 85221 [details] from commit-queue.

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

Last 500 characters of output:
c3/JavaArray ....
java/lc3/JavaClass ....
java/lc3/JavaObject .....................................
java/lc3/StringMethods .
java/lc3/forin .
java/lc3/instanceof .
jquery ...........
loader ..
mathml .......
mathml/presentation ..................
media ................
media/controls-after-reload.html -> failed

Exiting early after 1 failures. 13129 tests run.
383.26s total testing time

13128 test cases (99%) succeeded
1 test case (<1%) had incorrect layout
9 test cases (<1%) had stderr output

Full output: http://queues.webkit.org/results/8148088
Comment 8 Darin Adler 2011-06-18 12:31:10 PDT 
Comment on attachment 85221 [details]
Patch

Clearing review flag since this patch caused a test failure on the commit bot.
Comment 9 imasaki 2011-10-19 15:18:29 PDT 
vrk@, do you have update on this?