[Webkit-unassigned] [Bug 78404] [chromium] Replace RefPtr with OwnPtr for CCLayerImpl tree structure

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=78404





--- Comment #28 from Tien-Ren Chen <trchen at chromium.org>  2012-02-24 21:50:29 PST ---
(In reply to comment #26)
> (In reply to comment #25)
> > (From update of attachment 128839 [details] [details])
> > Clearing c-q for now, ASAN seems to have an issue with one of the unit tests..
> 
> Ah, ASAN caught an interesting bug in removeFromParent():
> 
> 82    void CCLayerImpl::removeFromParent()
> 83    {
> 84        if (!m_parent)
> 85            return;
> 86        for (size_t i = 0; i < m_parent->m_children.size(); ++i) {
> 87            if (m_parent->m_children[i].get() == this) {
> 88                m_parent->m_children.remove(i);
> 89                break;
> 90            }
> 91        }
> 92        m_parent = 0;
> 93      }
> 
> Line 88 was deleting 'this' and line 92 was attempting to set a member variable.  Shuffling the order around fixes this and results in a clean ASAN run of webkit_unit_tests.  I'll run the layout tests under ASAN, then land with this fix.

Now I think of it, why bother who is my mommy when I'm going to die anyway?
I think we can extend this function to return a PassOwnPtr to allow re-parenting a layer, so the PassOwnPtr can also extend the lifetime of 'this' until it returns.

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