[Webkit-unassigned] [Bug 102215] WebGL: Avoid unnecessary format conversion for texImage2D by using GL extension

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=102215





--- Comment #6 from Jun Jiang <jun.a.jiang at intel.com>  2013-01-23 05:02:19 PST ---
Hi, Kenneth and Gregg. Thanks for your comments.  
The initiative of this patch is to try to support BGRA888 extension in WebGL which could be used for tex{Sub}Image2D. That I post this patch was intended for discussion. And I had noticed that the format and internal format of texImage2D() should be the same as that of texSubImage2D().That's why I mess the code for texSubImage2D() by checking texture format and doing RGBA8 to BGRA8 conversion if needed. I added FIXME comments there and hope this limitation in GLES could be relaxed in some way if possible and the code following FIXME part could be removed in the end.The failure for the WebGL conformance test is a direct result when BGRA format support is added and it is in my expectation. The vital issue to me is that the limitation exposed by GLES seems difficult to be relaxed(part of spec and not mean to be relaxed even in an extension manner) and it still has other tough things remaining to be addressed , such as what Gregg showed in which different formats are generated for different levels. I don't want to mess the code and I admit it is inappropriate and not ready to add support for BGRA8 extension for tex{Sub}Image2D in WebGL this way due to its complexity and cost.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.