[Webkit-unassigned] [Bug 106903] [GTK] Possible bogus unref of accessible table cell

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=106903





--- Comment #12 from Mario Sanchez Prada <mario at webkit.org>  2013-02-04 05:43:51 PST ---
(In reply to comment #11)
> > So, attaching a one-liner patch :)
> 
> Suhweet!! :) That fixes the problem I was seeing with Orca (but could not come 
> up with a reliably-reproducible test case for). Thank YOU!!!

Glad to hear that. Thanks!

> Any suggestions regarding a layout test to go with this patch? (Or could we go 
> layout-testless on the grounds that it's a one-liner patch that fixes a 
> hard-to-repro, but pain in the arse bug?)

Yes, I think I could write a layout test that calls repeatedly to cellForColumnAndRow() for the accessibilityUIElement representing the table, which should crash without the patch and not with it.

Or we could just write an unit tests to call atk_table_ref_at() directly, too.

I'll work on that now.

> I am now going to build webkit with this patch in my GNOME jhbuild 
> environment to see if it makes the Orca + Evolution crasher go away. 
> /me crosses her fingers.

/me crosses fingers too!

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