[webkit-reviews] review granted: [Bug 201487] Generate event region for both the main graphics layer and the scrolled contents layer : [Attachment 378041] patch

[bugzilla-daemon at webkit.org]

Simon Fraser (smfr) <simon.fraser at apple.com> has granted Antti Koivisto
<koivisto at iki.fi>'s request for review:
Bug 201487: Generate event region for both the main graphics layer and the
scrolled contents layer
https://bugs.webkit.org/show_bug.cgi?id=201487

Attachment 378041: patch

https://bugs.webkit.org/attachment.cgi?id=378041&action=review




--- Comment #2 from Simon Fraser (smfr) <simon.fraser at apple.com> ---
Comment on attachment 378041
  --> https://bugs.webkit.org/attachment.cgi?id=378041
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=378041&action=review

Are there behavioral changes from using RenderLayerBacking::paintIntoLayer()
now that can be tested?

> Source/WebCore/rendering/RenderLayerBacking.cpp:1601
> +	   auto clipRect = enclosingIntRect(compositedBounds());

Isn't this wrong for the scrolled contents? You need to expand the clip there.

An issue with generating the event region for the entire scrolled contents is
that you're potentially traversing a massive render tree, most of which is not
visible. Ideally we'd generate event regions for areas with live tiles only.

> Source/WebCore/rendering/RenderLayerBacking.cpp:1604
> +	   auto layerOffset = toIntSize(graphicsLayer.scrollOffset()) -
roundedIntSize(graphicsLayer.offsetFromRenderer());

I think it would be a bit nicer to pass in the data to apply the scrolling
offset. I don't really like GraphicsLayers knowing about scrolling.