Bug 102310

Summary: [chromium] WebGL texImage2D fails with deferred image decoding
Product: WebKit Reporter: Hin-Chung Lam <hclam>
Component: WebGLAssignee: Hin-Chung Lam <hclam>
Status: RESOLVED FIXED    
Severity: Normal CC: dino, jamesr, kbr, nick, senorblanco, webkit.review.bot
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch
none
Patch for landing none

Description Hin-Chung Lam 2012-11-14 17:59:25 PST
When deferred image decoding is enabled WebGL's texImage2D using an Image object will fail.

This is due to the fact that it generates an ImageFrame marked as incomplete which WebGL rejects.
Comment 1 Hin-Chung Lam 2012-11-14 18:09:13 PST
Created attachment 174307 [details]
Patch
Comment 2 James Robinson 2012-11-14 18:16:52 PST
Comment on attachment 174307 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=174307&action=review

> LayoutTests/fast/images/webgl-teximage2d.html:1
> +<script src="../js/resources/js-test-pre.js"></script>

add a <!DOCTYPE html> before this

> LayoutTests/fast/images/webgl-teximage2d.html:2
> +<script src="../canvas/webgl/resources/webgl-test.js"></script>

why isn't this test in fast/canvas/webgl ?
Comment 3 James Robinson 2012-11-14 18:18:20 PST
Seems fairly reasonable to me, but I think Ken or Stephen are more familiar with this bit.
Comment 4 Hin-Chung Lam 2012-11-14 18:26:07 PST
I talked to Dirk about this. Since I want to run this test with --enable-deferred-image-decoding virtual suite there's a couple options:

1. Add a virtual suite that runs *one* test in fast/canvas/webgl
2. Copy the test from fast/canvas/webgl
3. Write a simple webgl test in fast/images that does exactly what I need

I did just 3.
Comment 5 Kenneth Russell 2012-11-14 19:05:56 PST
Comment on attachment 174307 [details]
Patch

Looks good. Please make sure the WebGL conformance suite and the fast/canvas/webgl/ layout tests pass with --enable-deferred-image-decoding, as there are many other tests which already exercise this path, some potentially more strenuously. r=me
Comment 6 Stephen White 2012-11-15 08:22:42 PST
Comment on attachment 174307 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=174307&action=review

I have no comments; will leave for Ken.

> Source/WebCore/ChangeLog:10
> +        generates an ImageFrame marked as incompelte, which WebGL rejects. This results

incompelte -> incomplete
Comment 7 Kenneth Russell 2012-11-15 10:44:19 PST
Note: please mark the patch cq? (Details link) if you would like it submitted to the commit queue.
Comment 8 Hin-Chung Lam 2012-11-15 10:45:43 PST
(In reply to comment #7)
> Note: please mark the patch cq? (Details link) if you would like it submitted to the commit queue.

Thanks. I'll land this manually.
Comment 9 Hin-Chung Lam 2012-11-15 10:47:52 PST
Created attachment 174486 [details]
Patch for landing
Comment 10 WebKit Review Bot 2012-11-15 12:36:14 PST
Comment on attachment 174486 [details]
Patch for landing

Clearing flags on attachment: 174486

Committed r134816: <http://trac.webkit.org/changeset/134816>
Comment 11 WebKit Review Bot 2012-11-15 12:36:18 PST
All reviewed patches have been landed.  Closing bug.