Bug 48395 - Accelerated animation with missing values in keyframes is broken
Summary: Accelerated animation with missing values in keyframes is broken
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Layout and Rendering (show other bugs)
Version: 528+ (Nightly build)
Hardware: Mac OS X 10.5
: P2 Normal
Assignee: Simon Fraser (smfr)
URL: http://dev.sencha.com/animator/demos/...
Keywords:
Depends on:
Blocks:
 
Reported: 2010-10-26 20:44 PDT by Simon Fraser (smfr)
Modified: 2010-10-27 14:32 PDT (History)
8 users (show)

See Also:


Attachments
Testcase (748 bytes, text/html)
2010-10-26 20:44 PDT, Simon Fraser (smfr)
no flags Details
Patch (10.15 KB, patch)
2010-10-26 22:13 PDT, Simon Fraser (smfr)
no flags Details | Formatted Diff | Diff
Patch (16.78 KB, patch)
2010-10-27 09:35 PDT, Simon Fraser (smfr)
mitz: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Simon Fraser (smfr) 2010-10-26 20:44:36 PDT
Created attachment 71976 [details]
Testcase

http://dev.sencha.com/animator/demos/healthy-choice/ shows problems when running with accelerated compositing; the second set of divs do not become visible at all.
Comment 1 Simon Fraser (smfr) 2010-10-26 20:46:26 PDT
The workaround is to explicitly specify opacity: 1 in keyframes that need it.
Comment 2 Simon Fraser (smfr) 2010-10-26 20:49:22 PDT
This probably regressed in r66339.
Comment 3 Simon Fraser (smfr) 2010-10-26 22:13:42 PDT
Created attachment 71983 [details]
Patch
Comment 4 Jarred Nicholls 2010-10-26 23:18:25 PDT
Looks good to me.
Comment 5 Simon Fraser (smfr) 2010-10-27 08:26:12 PDT
Comment on attachment 71983 [details]
Patch

Removing ? until I figure out if I have to stuff the values into the first keyframe also.
Comment 6 Simon Fraser (smfr) 2010-10-27 09:35:46 PDT
Created attachment 72045 [details]
Patch
Comment 7 WebKit Review Bot 2010-10-27 09:37:15 PDT
Attachment 72045 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style']" exit_code: 1
WebCore/rendering/RenderLayerBacking.cpp:1121:  Tests for true/false, null/non-null, and zero/non-zero should all be done without equality comparisons.  [readability/comparison_to_zero] [5]
Total errors found: 1 in 9 files


If any of these errors are false positives, please file a bug against check-webkit-style.
Comment 8 Jarred Nicholls 2010-10-27 09:46:15 PDT
bool isFirstOrLastKeyframe = !key || key == 1; // per style check... ?
Comment 9 Simon Fraser (smfr) 2010-10-27 09:49:01 PDT
http://trac.webkit.org/changeset/70657

(In reply to comment #8)
> bool isFirstOrLastKeyframe = !key || key == 1; // per style check... ?

I think what I have is more readable. I am purposefully ignoring the style check.
Comment 10 Jarred Nicholls 2010-10-27 09:54:58 PDT
Indeed.  Good deal.
Comment 11 WebKit Review Bot 2010-10-27 14:03:24 PDT
http://trac.webkit.org/changeset/70657 might have broken GTK Linux 32-bit Debug and Qt Linux Release
Comment 12 Csaba Osztrogonác 2010-10-27 14:32:00 PDT
(In reply to comment #11)
> http://trac.webkit.org/changeset/70657 might have broken GTK Linux 32-bit Debug and Qt Linux Release

FYI: Qt specifix expected results landed in: http://trac.webkit.org/changeset
/70710

Please don't ignore commnts of sheriffbot! He usually tell the truth.