[Webkit-unassigned] [Bug 30123] [GTK] Segfault while testing accessibility/iframe-bastardization.html

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=30123


chris fleizach <cfleizach at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |cfleizach at apple.com




--- Comment #5 from chris fleizach <cfleizach at apple.com>  2010-10-21 11:24:24 PST ---
(In reply to comment #4)
> (In reply to comment #3)
> > How might an AccessibilityUIElement be created with a m_element member that is null?
> 
> I can't give a categoric answer to this, but in the past I've observed that sometimes (probably depending on how the layout test was written) the platform dependant wrapper for the a11y object had not been created (normally because it was not rendered yet) at the moment of creating the AccessibilityUIElement. 
> 
> In those cases, the fix I found was normally to modify the layout test to actually perform all the  checks in a function called in the body.onLoad event (to make sure at least the document was loaded before doing anything).
> 
> In other situations the problem was that calling to a function like rootElement or focusedElement was returning NULL (see bug 38648), and perhaps there are other scenarios were you could get a null m_element.
> 
> Hence, because of this "experience" I'd say the extra check doesn't hurt and it's generally a good practice to do in AccessibilityUIElement.
> 
> But, as always, I might be wrong... :-)

A m_renderer can be detached and set to NULL, but the AX object might still be around (waiting to be cleaned up). it doesn't happen frequently, but it does happen

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