Expected: "Normal flow" geometry is always used from a scroll-snap-align set element, even if it has sticky position scheme. Observed: Scenario where snapping scroller doesn't register children's normal flow snap alignments, if position:sticky "sticks" them into the scrollport towards the scrolling start edge, meaning their normal flow is passed the scrollport's "fold"? 1. Navigate to testcase -> https://jsfiddle.net/bzantvmk/ 2. Scroll horizontally, notice how the scroller only snaps to the beginning and end. 3. Alter the 'position' of '.panel' blocks to 'static' and see snapping work correctly for each panel in the scroller. In most my mixing of scroll-snap and sticky, alignment is calculated correctly, but this is a case where it's completely broken. cc wenson_hsieh@apple.com
wenson_hsieh@apple.com Don't want to bug, but just wondering if this was ever triaged and reproduced?
<rdar://problem/43015593>
(In reply to jonjohnjohnson from comment #1) > wenson_hsieh@apple.com Don't want to bug, but just wondering if this was > ever triaged and reproduced? Sorry for the delay — I'm still able to reproduce this on the latest macOS Mojave beta.
Here is a video of the snapport not calculating any of the snap positions correctly and only snapping to the beginning and ends of the scroll. http://cl.ly/1h280g262V1D Again, if you cause a re-layout by altering the position, snap positions get calculated correctly. But not on initial layout.
fred.wang@free.fr When looking into how "scroll snap offsets" are created in https://bugs.webkit.org/show_bug.cgi?id=191816 I wonder if you any insight into this quite similar issue?