[Webkit-unassigned] [Bug 176914] Async iframes in the scrolling tree should be in the z-index order

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Sep 28 08:29:02 PDT 2017


https://bugs.webkit.org/show_bug.cgi?id=176914

--- Comment #11 from Frédéric Wang (:fredw) <fred.wang at free.fr> ---
So I checked this a bit. The scrolling nodes happen to be attached at several places, for example some of them are: 

ScrollingStateTree::attachNode
AsyncScrollingCoordinator::attachToStateTree
|
\_
   RenderLayerCompositor::attachScrollingNode
   RenderLayerCompositor::updateScrollCoordinationForThisFrame
   RenderLayerCompositor::updateScrollCoordinatedLayer
   RenderLayerCompositor::updateScrollCoordinatedStatus
   |
   \_
      RenderLayerBacking::updateGeometry
      RenderLayerCompositor::rebuildCompositingLayerTree
      RenderLayerCompositor::updateCompositingLayers
   |
   \_
      RenderLayerCompositor::updateBacking
      RenderLayerCompositor::computeCompositingRequirements
      RenderLayerCompositor::updateCompositingLayers
|
\_
   AsyncScrollingCoordinator::ensureRootStateNodeForFrameView
   AsyncScrollingCoordinator::frameViewRootLayerDidChange
|
\_
   RenderLayerCompositor::reattachSubframeScrollLayers
   RenderLayerCompositor::updateCompositingLayers

ScrollingStateTree::attachNode basically appends the node to the parent and does not change anything if the node is already a child. Functions like rebuildCompositingLayerTree and computeCompositingRequirements browse the tree in the z-order (i.e. negZOrderList, normalFlowList, posZOrderList). However, if new nodes are inserted later, the scrolling nodes will never be reordered to match the z-order.

I am not sure what would be the best way to handle this. Maybe we should pass the ID of the previous sibling to attachNode or somewhat force a re-ordering somewhere.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170928/eaa6f1e1/attachment.html>


More information about the webkit-unassigned mailing list