[Webkit-unassigned] [Bug 56393] Without checking existence of the renderer of the element, tries to access the enclosing layer.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=56393





--- Comment #23 from Joe Wild <joseph.wild at nokia.com>  2011-06-14 12:18:01 PST ---
(In reply to comment #22)
Now I understand your question.

I added this ASSERT so that this situation would fail on all platforms
so I could write the test.  If we had Symbian layout tests, that would
be a much better alternative.

Orignally, I put the ASSERT at the site of
the call as you suggest, but that did not make any sense once I applied
the patch for checking the null renderer.

IntRect FrameView::windowClipRect()
  ...
  ASSERT(elt && elt->renderer());

  // Patch for null renderer
  RenderLayer* layer = 0; 
  if (elt && elt->renderer()) 
    layer = elt->renderer()->enclosingLayer(); 

So I thought the best place to include the ASSERT was in 
enclosingLayer().

I agree that the ASSERT is a bit odd.  However, I could not think
of a better way to expose this problem in the Layout tests.  I am 
open to other suggestions.  While this currently only fails on Symbian,
I assume we don't want WebKit depending on undefined C++ behavior.

I hope this makes sense.  If not, please ask more questions.

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