Bug 92469 - [chromium] CCLayerTreeHost placeholder scissor should be deviceViewportSize instead of viewportSize
Summary: [chromium] CCLayerTreeHost placeholder scissor should be deviceViewportSize i...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Layout and Rendering (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Shawn Singh
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-07-27 00:42 PDT by Shawn Singh
Modified: 2012-07-27 15:04 PDT (History)
6 users (show)

See Also:


Attachments
Patch (1.87 KB, patch)
2012-07-27 00:53 PDT, Shawn Singh
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Shawn Singh 2012-07-27 00:42:43 PDT
http://code.google.com/p/chromium/issues/detail?id=139005

After tracing the problem from missing tiles --> skipped paints --> bad occlusion --> bad scissor rect, it finally showed up that we are accidentally setting the placeholder scissorRect to the viewportSize.  As I understand, the space of the scissorRect is expected to be in the pixel accurate content space of the targetSurface; and for this main-thread placeholder default scissor, that means we should have been using deviceViewportSize instead of viewportSize.  It does indeed fix the flickering issue that showed up suddenly on canary.  

Dear Reviewers:  if you wish to land it right away, please go ahead, and feel free to create another bug on me to add a unit test.   If you prefer to wait and land this after unit test is created, I'll work on it right away to get this landed soon.
Comment 1 Shawn Singh 2012-07-27 00:53:30 PDT
Created attachment 154868 [details]
Patch
Comment 2 Dana Jansens 2012-07-27 08:32:33 PDT
Comment on attachment 154868 [details]
Patch

good catch. lgtm
Comment 3 Adrienne Walker 2012-07-27 08:33:57 PDT
The code change looks right to me, but if you can get a test in today, let's land them together.
Comment 4 Adrienne Walker 2012-07-27 14:05:54 PDT
Comment on attachment 154868 [details]
Patch

R=me.  After some conversation with Shawn, I'm not sure that there's a really good way to test this without some serious refactoring of how CCLayerTreeHost works.  This seems clearly like the right fix, so let's just land it.
Comment 5 WebKit Review Bot 2012-07-27 15:04:32 PDT
Comment on attachment 154868 [details]
Patch

Clearing flags on attachment: 154868

Committed r123921: <http://trac.webkit.org/changeset/123921>
Comment 6 WebKit Review Bot 2012-07-27 15:04:36 PDT
All reviewed patches have been landed.  Closing bug.