Bug 53154

Summary: Fix multisampling support in DrawingBuffer
Product: WebKit Reporter: Kenneth Russell <kbr>
Component: CanvasAssignee: Kenneth Russell <kbr>
Status: RESOLVED FIXED    
Severity: Normal CC: abarth, cmarrin, eric, gman, jamesr, mdelaney7, senorblanco, webkit.review.bot
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: All   
OS: All   
Attachments:
Description Flags
Patch jamesr: review+, kbr: commit-queue-

Description Kenneth Russell 2011-01-25 18:58:28 PST
Recent changes to the DrawingBuffer class to add antialiasing support to the accelerated 2D canvas code path had some bugs and were illegally using certain enums. These issues need to be fixed and the Chromium port implemented.
Comment 1 Kenneth Russell 2011-01-25 19:11:15 PST
Created attachment 80159 [details]
Patch
Comment 2 Stephen White 2011-01-25 21:25:06 PST
Comment on attachment 80159 [details]
Patch

Thanks!
Comment 3 Stephen White 2011-01-25 21:26:57 PST
Comment on attachment 80159 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=80159&action=review

> Source/WebKit/chromium/ChangeLog:10
> +        GL_CHROMIUM_framebuffer_multisample) through WebGraphicsContext3D.

Not a blocker, but I'm not sure I understand why we don't plumb through the ANGLE extensions verbatim.  Feel free to de-confuse me at your convenience.  :)
Comment 4 James Robinson 2011-01-25 22:52:01 PST
Comment on attachment 80159 [details]
Patch

R=me.  I'm also curious about the extension name mapping.
Comment 5 Kenneth Russell 2011-01-26 14:07:04 PST
Committed r76717: <http://trac.webkit.org/changeset/76717>
Comment 6 Kenneth Russell 2011-01-26 14:12:09 PST
(In reply to comment #3)
> (From update of attachment 80159 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=80159&action=review
> 
> > Source/WebKit/chromium/ChangeLog:10
> > +        GL_CHROMIUM_framebuffer_multisample) through WebGraphicsContext3D.
> 
> Not a blocker, but I'm not sure I understand why we don't plumb through the ANGLE extensions verbatim.  Feel free to de-confuse me at your convenience.  :)

I don't remember why we decided to expose a different extension than GL_ANGLE_framebuffer_multisample and GL_ANGLE_framebuffer_blit from the command buffer code. The ANGLE extension is written against OpenGL ES 2.0, and maybe we wanted slightly different or relaxed semantics so that we didn't have to have absolutely identical behavior between desktop and OpenGL ES 2.0 implementations.

CC'ing Gregg in case he remembers.
Comment 7 WebKit Review Bot 2011-01-26 14:52:22 PST
http://trac.webkit.org/changeset/76717 might have broken Qt Linux Release