Bug 145140 – REGRESSION(142590): Scroll-snap points are improperly snapping to earlier index values

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Brent Fulgham]

My fix in Bug 142590 for resizing events introduced a snapping logic bug that causes some scroll-snap gestures to trigger the "nearestActiveSnapPoint" logic, throwing the scroll position back to the previous snap point.

This must be fixed!

[Brent Fulgham]

The new "nearestActiveSnapPoint" logic is firing while scroll snap animations are running. We need to add an "isScrollSnapInProgress" predicate, much like the existing "isRubberBandInProgress" to avoid certain "fix-up" logic that we don't want running while we are in the process of moving to a new position.

[Brent Fulgham]

<rdar://problem/21006738>

[Brent Fulgham]

Created attachment 253345 [details]
Patch

[Beth Dakin]

Comment on attachment 253345 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=253345&action=review

> Source/WebCore/platform/mac/ScrollAnimatorMac.h:135
> +    bool isRubberBandInProgress() const override;

Did you mean to remove the "virtual" from isRubberBandInProgress()?

[Brent Fulgham]

(In reply to comment #4)
> Comment on attachment 253345 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=253345&action=review
> 
> > Source/WebCore/platform/mac/ScrollAnimatorMac.h:135
> > +    bool isRubberBandInProgress() const override;
> 
> Did you mean to remove the "virtual" from isRubberBandInProgress()?

I just switched to the new 'override' style guideline, where we don't also leave the 'virtual' declaration (it's implied by the 'override' keyword).

[Brent Fulgham]

Committed r184513: <http://trac.webkit.org/changeset/184513>