Summary: | REGRESSION: Scroll snapping triggers body scroll | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Bruno Stasse <augus.dupin> | ||||||
Component: | Scrolling | Assignee: | Simon Fraser (smfr) <simon.fraser> | ||||||
Status: | RESOLVED FIXED | ||||||||
Severity: | Normal | CC: | augus.dupin, sam, simon.fraser, smoley, webkit-bug-importer, wenson_hsieh | ||||||
Priority: | P2 | Keywords: | InRadar | ||||||
Version: | Safari Technology Preview | ||||||||
Hardware: | Mac (Intel) | ||||||||
OS: | macOS 10.15 | ||||||||
Attachments: |
|
Description
Bruno Stasse
2020-12-21 07:49:16 PST
Thanks for filing. I'm not sure about the regression status as I can reproduce this on Safari 13.1.3 as well as STP 117 (14.1). Yes, that's right I can too. However, I'm pretty sure it was not the case in STP 115 (I don't know about 116, I didn't test it). Latching isn't working correctly; with scroll snap we fail to tell ScrollingTreeLatchingController that we've received events, so the latch times out. This might be: if (processWheelEventForScrollSnap(wheelEvent)) return false; // FIXME: Why don't we report that we handled it? Created attachment 416694 [details]
Patch
Comment on attachment 416694 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=416694&action=review > Source/WebCore/platform/cocoa/ScrollController.h:172 > bool processWheelEventForScrollSnap(const PlatformWheelEvent&); Not for this change obviously, but maybe it's time to replace these boolean return values with something a bit more expressive. Yeah, I spent 5 minutes trying to think of an enum name and failed. We need this handled/unhandled enum more extensively. |