[webkit-reviews] review denied: [Bug 111126] Enable GPU-GPU texture copy in texImage2D() for HTMLVideoElement if hardware accelerated video decode is in use : [Attachment 192373] Patch

[bugzilla-daemon at webkit.org]

Kenneth Russell <kbr at google.com> has denied Jun Jiang <jun.a.jiang at intel.com>'s
request for review:
Bug 111126: Enable GPU-GPU texture copy in texImage2D() for HTMLVideoElement if
hardware accelerated video decode is in use
https://bugs.webkit.org/show_bug.cgi?id=111126

Attachment 192373: Patch
https://bugs.webkit.org/attachment.cgi?id=192373&action=review

------- Additional Comments from Kenneth Russell <kbr at google.com>
View in context: https://bugs.webkit.org/attachment.cgi?id=192373&action=review


Thanks for your continued work on this. It looks good overall, and the Chromium
API change looks fine, but there's a new layering violation that requires one
more change.

> Source/WebCore/platform/graphics/MediaPlayer.h:350
> +    bool copyVideoTextureToPlatformTexture(GraphicsContext3D*,
WebGLTexture*, GC3Dint level, GC3Denum type, GC3Denum internalFormat, bool
premultiplyAlpha, bool flipY);

Code under platform/graphics is not supposed to reference WebCore types outside
of platform/graphics like WebGLTexture*. While there are other violations of
this rule in this class and in the WebGL implementation, it isn't good to add
more. Please change the WebGLTexture* argument to Platform3DObject (defined in
GraphicsTypes3D.h).

> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:781
> +	   return false;

Since according to the layering rules the WebGLTexture object won't be
available here, please move these checks back up to WebGLRenderingContext.cpp
and have them apply to all ports -- they aren't unreasonable restrictions.
Also, please document the current restrictions in MediaPlayer.h.