Bug 84807 – [chromium] Don't call calculateRenderPasses in CCLayerTreeHostImpl::scrollBegin()

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[Adrienne Walker]

calculateRenderPasses calls willDraw() on every layer and unless drawLayers is then called, there's a missing didDraw() pair.  This causes problems for video layers that want to lock and release resources during these two function calls.

[Adrienne Walker]

Created attachment 138696 [details]
Patch

[James Robinson]

Comment on attachment 138696 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=138696&action=review

Much better R=me

> Source/WebCore/platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:266
> +    ASSERT(passes.isEmpty());

can we make passes a local variable again and just have the renderSurfaceLayerList be a member variable, or is that teh crazy?

[Adrienne Walker]

(In reply to comment #2)
> (From update of attachment 138696 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=138696&action=review
> 
> Much better R=me
> 
> > Source/WebCore/platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:266
> > +    ASSERT(passes.isEmpty());
> 
> can we make passes a local variable again and just have the renderSurfaceLayerList be a member variable, or is that teh crazy?

Sadly, I think that'd be a larger refactoring, although it might be worthwhile to investigate it.

[WebKit Review Bot]

Comment on attachment 138696 [details]
Patch

Clearing flags on attachment: 138696

Committed r115162: <http://trac.webkit.org/changeset/115162>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.