[webkit-reviews] review requested: [Bug 72964] [chromium] Let LayerChromium subclasses override value of opaque() flag. Set it for image, video, and canvas layers. : [Attachment 117623] Patch

[bugzilla-daemon at webkit.org]

Dana Jansens <danakj at chromium.org> has asked  for review:
Bug 72964: [chromium] Let LayerChromium subclasses override value of opaque()
flag. Set it for image, video, and canvas layers.
https://bugs.webkit.org/show_bug.cgi?id=72964

Attachment 117623: Patch
https://bugs.webkit.org/attachment.cgi?id=117623&action=review

------- Additional Comments from Dana Jansens <danakj at chromium.org>
(In reply to comment #9)
> > We could do some complex logic to try to detect if every pixel of the
canvas was actually opaque but that's a bit involved. I think this should go on
WebGLLayerChromium since that actually has an 'has alpha' bit.
> 
> Sounds good. I will move the m_hasAlpha bit down to the WebGL layer.

The reason the m_hasAlpha bit is on the CanvasLayerChromium is because it is
also on the CCCanvasLayerChromium, even though the WebGL layer is the only
subtype to actually set the flag.

This is one way to move the alpha bits to the WebGL layer, but it leaves it out
of the pushPropertiesTo call entirely in the Canvas layer. Is that okay?  An
alternate means could be to have a virtual hasAlpha() function on the Canvas
layer and push that, and have the WebGL layer override and return its
m_hasAlpha.

Lots of good new feeback to go through also, thanks.