The calculation of non-fast-scrollable regions does not currently take the topContentOffset into account. Consequently, the logic that decides whether to stay on the scrolling thread, or drop down to an individual page element, can make the wrong choice. This is especially true for small scrollable regions (such as <select> elements), where the topContentInset may be quite close to the size of the scrollable element itself.
<rdar://problem/16706152>
Created attachment 230126 [details] Patch
Comment on attachment 230126 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=230126&action=review > LayoutTests/platform/mac/fast/scrolling/scroll-select-bottom-test.html:52 > + //debug("Page before: " + pageScrollPositionBefore + ", select before: " + selectScrollPositionBefore); > + //debug("Page after: " + pageScrollPositionAfter + ", select after: " + selectScrollPositionAfter); ???
Comment on attachment 230126 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=230126&action=review >> LayoutTests/platform/mac/fast/scrolling/scroll-select-bottom-test.html:52 >> + //debug("Page after: " + pageScrollPositionAfter + ", select after: " + selectScrollPositionAfter); > > ??? WK1 and WK2 sometimes produce different values here. Rather than baseline both, we comment these out. It's useful to uncommercial them while debugging.
(In reply to comment #4) > (From update of attachment 230126 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=230126&action=review > > >> LayoutTests/platform/mac/fast/scrolling/scroll-select-bottom-test.html:52 > >> + //debug("Page after: " + pageScrollPositionAfter + ", select after: " + selectScrollPositionAfter); > > > > ??? > > WK1 and WK2 sometimes produce different values here. Rather than baseline both, we comment these out. It's useful to uncommercial them while debugging. And by uncommercial, I of course mean uncomment them.
Comment on attachment 230126 [details] Patch Clearing flags on attachment: 230126 Committed r167807: <http://trac.webkit.org/changeset/167807>
All reviewed patches have been landed. Closing bug.