Bug 94961

Summary: WebGL should not flip textures on presentation if contents are unchanged
Product: WebKit Reporter: James Robinson <jamesr>
Component: New BugsAssignee: James Robinson <jamesr>
Status: RESOLVED FIXED    
Severity: Normal CC: abarth, cc-bugs, cmarrin, danakj, dino, dpranke, enne, jbauman, kbr, ojan, piman, webkit.review.bot, zlieber, zmo
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch kbr: review+

James Robinson
Reported 2012-08-24 11:27:08 PDT
WebGL should not flip textures on presentation if contents are unchanged
Attachments
Patch (9.21 KB, patch)
2012-08-24 11:33 PDT, James Robinson 		kbr: review+
DetailsFormatted DiffDiff
View All Add attachment proposed patch, testcase, etc.
James Robinson
Comment 1 2012-08-24 11:33:48 PDT
Created attachment 160461 [details] Patch
James Robinson
Comment 2 2012-08-24 11:34:48 PDT
The bug is chromium-specific (afaik) but the fix touches cross-platform code, so I haven't marked this with [chromium]. I think this could also be solved in a chromium-specific way that doesn't touch DrawingBuffer at all but this seems to me like something DrawingBuffer should be able to handle.
Kenneth Russell
Comment 3 2012-08-24 17:01:23 PDT
Comment on attachment 160461 [details] Patch The fix looks good as far as I can see. Some additional test cases to try would include printing of WebGL content, the other WebGL layout tests, and various demos on the web. r=me
James Robinson
Comment 4 2012-08-24 18:53:31 PDT
(In reply to comment #3) > (From update of attachment 160461 [details]) > The fix looks good as far as I can see. Some additional test cases to try would include printing of WebGL content, the other WebGL layout tests, and various demos on the web. r=me Thanks for the review. I'll check these locally before landing - our automated testing of things like webgl printing hasn't been too reliable in the past. Is there a way to force AA off for local testing other than code modification, by any chance?
Kenneth Russell
Comment 5 2012-08-27 14:55:07 PDT
(In reply to comment #4) > (In reply to comment #3) > > (From update of attachment 160461 [details] [details]) > > The fix looks good as far as I can see. Some additional test cases to try would include printing of WebGL content, the other WebGL layout tests, and various demos on the web. r=me > > Thanks for the review. I'll check these locally before landing - our automated testing of things like webgl printing hasn't been too reliable in the past. Is there a way to force AA off for local testing other than code modification, by any chance? zmo added the command line option --disable-gl-multisampling a while back; I think it should still work.
James Robinson
Comment 6 2012-08-27 15:35:00 PDT
Printing appears to be broken with or without this patch - after printing the WebGL content is cleared. I don't know how long this has been broken. Most pages / demos repaint the canvas on every frame so this problem is transient. Filed https://bugs.webkit.org/show_bug.cgi?id=95140 to cover that bug. I don't think this patch makes anything worse.
Kenneth Russell
Comment 7 2012-08-27 15:38:35 PDT
(In reply to comment #6) > Printing appears to be broken with or without this patch - after printing the WebGL content is cleared. I don't know how long this has been broken. Most pages / demos repaint the canvas on every frame so this problem is transient. Filed https://bugs.webkit.org/show_bug.cgi?id=95140 to cover that bug. I don't think this patch makes anything worse. Darn it. Thanks for testing anyway. Please go ahead with your patch.
James Robinson
Comment 8 2012-08-27 15:42:04 PDT
Committed r126810: <http://trac.webkit.org/changeset/126810>
Note You need to log in before you can comment on or make changes to this bug.