Bug 95891 - [CSS Shaders] Vertex attributes should be unbound after a custom filter is applied
Summary: [CSS Shaders] Vertex attributes should be unbound after a custom filter is ap...
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: Max Vujovic
URL:
Keywords:
Depends on:
Blocks: 71392
  Show dependency tree / graph
 
Reported: 2012-09-05 14:04 PDT by Max Vujovic
Modified: 2012-09-06 12:06 PDT (History)
4 users (show)

See Also:

Attachments
Patch (4.91 KB, patch)
2012-09-05 17:09 PDT, Max Vujovic 		no flags Details | Formatted Diff | Diff
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 Max Vujovic 2012-09-05 14:04:43 PDT 
Chromium sometimes prints the following error message:
"GL ERROR :GL_INVALID_OPERATION : glDrawElements: attempt to render with no buffer attached to enabled attribute 2"

This happens in the following situation: 
1) There are two FECustomFilters on the page.
2) One FECustomFilter has a detached mesh. One FECustomFilter has an attached mesh. The detached FECustomFilter has one more vertex attribute (a_triangleCoord) than the attached FECustomFilter.
3) The detached FECustomFilter is destroyed, but a_triangleCoord remains bound.
4) The attached FECustomFilter tries to render, but Chromium notices that there is no buffer attached to the a_triangleCoord attribute.

We should unbind the vertex attributes after a custom filter is applied.
Comment 1 Max Vujovic 2012-09-05 17:09:00 PDT 
Created attachment 162374 [details]
Patch
Comment 2 Dean Jackson 2012-09-05 18:18:47 PDT 
Did you mean to r? this?
Comment 3 Alexandru Chiculita 2012-09-05 18:28:01 PDT 
I think he did. Can you take a look please?
Comment 4 Max Vujovic 2012-09-05 21:04:56 PDT 
(In reply to comment #2)
> Did you mean to r? this?

Yup, thanks Dean!
Comment 5 Max Vujovic 2012-09-06 09:32:21 PDT 
Comment on attachment 162374 [details]
Patch

Bots look good. Setting cq?.
Comment 6 Dean Jackson 2012-09-06 10:15:17 PDT 
You can put cq+ yourself in this case (no changes needed from r+).

I'll do it now.
Comment 7 WebKit Review Bot 2012-09-06 10:44:35 PDT 
Comment on attachment 162374 [details]
Patch

Clearing flags on attachment: 162374

Committed r127758: <http://trac.webkit.org/changeset/127758>
Comment 8 WebKit Review Bot 2012-09-06 10:44:39 PDT 
All reviewed patches have been landed.  Closing bug.
Comment 9 Max Vujovic 2012-09-06 12:06:35 PDT 
(In reply to comment #6)
> You can put cq+ yourself in this case (no changes needed from r+).
> 
> I'll do it now.

Thanks Dean. I don't have committer status yet :)