[Webkit-unassigned] [Bug 18064] assert-fail in WebCore::FrameLoader::saveDocumentState (document is null)

[bugzilla-daemon at webkit.org]

http://bugs.webkit.org/show_bug.cgi?id=18064


julien.chaffraix at gmail.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |julien.chaffraix at gmail.com




------- Comment #2 from julien.chaffraix at gmail.com  2008-03-27 08:07 PDT -------
(In reply to comment #1)
> Created an attachment (id=20080)
 --> (http://bugs.webkit.org/attachment.cgi?id=20080&action=view) [edit]
> remove bogus assertion
> 
> I'm pretty sure the ASSERT is bogus, because a frame can contain no document
> under certain circumstances, and because there's an if() test immediately after
> that protects against the case that the ASSERT is checking for.

I would be more cautious about removing the ASSERT because first it is never
triggered on the Mac port only on Qt/Gtk (as mentionned in
http://bugs.webkit.org/show_bug.cgi?id=17984).
You should also be carefull as ASSERTs are not enforced on release build so the
null check is a necessity for those cases.

Finally in bug17984, I showed another way of resolving the bug that does not
involve removing the ASSERT (but may lead to memory leaks).

> 
> Also because although the ASSERT trips on the debug build (for example when
> clicking the reference rendering link on http://acid3.acidtests.org/), the
> release build works fine in the same circumstances.
> 
> Thus this patch removes the ASSERT.
> 

Your patch lacks a test case (it is pretty straightforward as the assertion can
be triggered by a page with an iframe and just reloading it).


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