[Webkit-unassigned] [Bug 71181] Layout Test fast/forms/input-step-as-double.html is flaky

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=71181


Kent Tamura <tkent at chromium.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |benm at google.com,
                   |                            |steveblock at google.com




--- Comment #7 from Kent Tamura <tkent at chromium.org>  2011-11-03 23:35:24 PST ---
(In reply to comment #6)
> (In reply to comment #5)
> > When a mouse pointer moves on a node, WebKit usually calls the following functions:
> > 1. setHovered(true)
> > 2. defaultEventHandler() for 'mousemove'
> > 
> > However, if touch-target.html is invoked before input-step-as-double.html, this order is reversed.
> > 1. defaultEventhandler() for 'mousemove'
> > 2. setHovered(true)
> > 
> > I don't know why this happens.  Anyway, SpinButtonElement::m_upDownState is reset to Indeterminate in setHovered(true).  So SpinButtonElement::defaultEventHandler() calls input-stepUpFromRenderer(-1).
> 
> Can you see the difference in EventHandler::m_touchPressed ? Maybe we're not clearing some touch-related states properly.

m_touchPressed is false in the former case, and m_touchPressed is true in the latter case.

It seems EventHandler::m_touchPressed is never cleared if the last call of EventHandler::handleTouchEvent() had one or more touches.

* Should we clear m_touchPressed?
* Shouldn't we change the setHovered()-mousemove order in the case of m_touchPressed?
* Should we support the reversed order in SpinButtonElement?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.