[webkit-reviews] review denied: [Bug 23360] Make repainting compositing-aware for accelerated compositing : [Attachment 27268] Final repaint patch: hook up to composited layers.

[bugzilla-daemon at webkit.org]

Dave Hyatt <hyatt at apple.com> has denied Simon Fraser (smfr)
<simon.fraser at apple.com>'s request for review:
Bug 23360: Make repainting compositing-aware for accelerated compositing
https://bugs.webkit.org/show_bug.cgi?id=23360

Attachment 27268: Final repaint patch: hook up to composited layers.
https://bugs.webkit.org/attachment.cgi?id=27268&action=review

------- Additional Comments from Dave Hyatt <hyatt at apple.com>
Put a FIXME in RenderInline, since you aren't really dealing with the problem
yet.

You really shouldn't get a null containingBlock() ever from within
containingBlockWith()/Height().  I do approve of nuking a useless FIXME though.
:)

I think it's ok to have an enclosingCompositingLayer on RenderObject itself. 
That will make the code cleaner in RenderObject::containerForRepaint.

OwnPtr please for RenderView::m_compositor.

updateRootLayerPosition should not be in updateWidgetPositions.  I would
suggest updateLayerPositions or updateLayerPosition for the root layer, or
perhaps you meant to do it post-layout (either where updateLayerPositions is or
in performPostLayoutTasks).

I think you could just fix selection by adding enough info to get back to the
repaint container to SelectionInfo objects, and then you wouldn't need the
special view method.

Caret painting could also cache the repaint container that it needs to use.