Listener changes should update the region.
Created attachment 399362 [details] patch
Comment on attachment 399362 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=399362&action=review > LayoutTests/fast/scrolling/mac/wheel-event-listener-region-element-invalidation.html:20 > + extra line > LayoutTests/fast/scrolling/mac/wheel-event-listener-region-element-invalidation.html:29 > + function listener() { > + results.textContent += 'wheel\n'; > + } This is never called, right? It could be an empty function. > LayoutTests/fast/scrolling/mac/wheel-event-listener-region-element-invalidation.html:33 > + if (window.internals) > + results.textContent += internals.layerTreeAsText(document, internals.LAYER_TREE_INCLUDES_EVENT_REGION | internals.LAYER_TREE_INCLUDES_ROOT_LAYER_PROPERTIES); Beware of setting element contents here then fetching layers again later; layout will change making the two trees hard to compare. I prefer to store the first layer tree in a variable then set textContent once at the end. > LayoutTests/fast/scrolling/mac/wheel-event-listener-region-element-invalidation.html:36 > + passive.addEventListener('wheel', () => { results.textContent += 'passive wheel\n' }, { passive: true }); Ditto > Source/WebCore/rendering/RenderLayer.cpp:-6973 > - // FIXME: This should not be conditioned on PLATFORM(IOS_FAMILY). See <https://webkit.org/b/210216>. Can https://webkit.org/b/210216 be closed after this? > Source/WebCore/rendering/RenderLayer.cpp:6973 > +#if ENABLE(ASYNC_SCROLLING) This doesn't feel like quite the right #ifdef; isn't it also if we have touch events or editable regions?
> This doesn't feel like quite the right #ifdef; isn't it also if we have > touch events or editable regions? It doesn't, but it is the #ifdef currently used for RenderLayerBacking::updateEventRegion and this should match.
Created attachment 399474 [details] patch
Committed r261742: <https://trac.webkit.org/changeset/261742> All reviewed patches have been landed. Closing bug and clearing flags on attachment 399474 [details].
<rdar://problem/63270443>