Bug 141537

Summary:

REGRESSION: Should not send wheel events with a delta of 0

Product:

WebKit

Reporter:

Beth Dakin <bdakin>

Component:

Layout and Rendering

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

bdakin, bfulgham, cmarcelo, commit-queue, esprehn+autocc, kangil.han, simon.fraser, thorton

Priority:

Keywords:

InRadar, Regression

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Bug Depends on:

Bug Blocks:

141821

Attachments:

Description	Flags
Patch	simon.fraser: review+

Description Beth Dakin 2015-02-12 14:36:58 PST

This is a regression from http://trac.webkit.org/changeset/172832/trunk/Source/WebCore/dom/Element.cpp We should not be sending wheel events to the DOM with a delta of 0. This affects webmail.apple.com

rdar://problem/18903118

Comment 1 Beth Dakin 2015-02-12 14:58:01 PST

Created attachment 246478 [details]
Patch

Comment 2 Simon Fraser (smfr) 2015-02-12 15:05:38 PST

Comment on attachment 246478 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=246478&action=review

> Source/WebCore/dom/Element.cpp:282
> +    if (!(event.deltaX() || event.deltaY()))

This would be clearer as if (!deltaX && !deltaY()) I think.

Comment 3 Beth Dakin 2015-02-12 15:09:38 PST

http://trac.webkit.org/changeset/180018

Comment 4 Beth Dakin 2015-02-12 16:14:08 PST

Re-baselined tests in: http://trac.webkit.org/changeset/180021

Comment 5 Brent Fulgham 2015-02-17 14:38:46 PST

This change breaks snap scroll points, but you couldn't have known that since I have not gotten around to making suitable tests!

Comment 6 Beth Dakin 2015-02-17 14:42:50 PST

(In reply to comment #5)
> This change breaks snap scroll points, but you couldn't have known that
> since I have not gotten around to making suitable tests!

I wonder why snap scroll points rely on getting a delta of 0?

Comment 7 Brent Fulgham 2015-02-17 14:47:40 PST

(In reply to comment #6)
> (In reply to comment #5)
> > This change breaks snap scroll points, but you couldn't have known that
> > since I have not gotten around to making suitable tests!
> 
> I wonder why snap scroll points rely on getting a delta of 0?

I'm not sure. I'm debugging the code now to see why it matters. I think we have some kind of edge case where the user has ended the wheel motion and the snap animation is about to start. If we return 'false' in this case, the event is treated as unhandled and the animation never starts.

I'll know more once I finish debugging.