[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 #29 from Alexis Menard <alexis.menard at openbossa.org>  2011-06-14 13:40:30 PST ---
(In reply to comment #28)
> (In reply to comment #26)
> > (In reply to comment #24)
> > > > enclosingLayer().
> > > The first idea was better than the current one.
> > > One thing bother me. If the test is good enough you don't need the ASSERT, it should just crash properly therefore it will be catch at a LayoutTest failure. A crashing test is good enough to be treated as a regression/failure.
> > 
> > That's the problem.  The test without the ASSERT is not good enough
> > to crash on the Unix platforms.  The Unix platforms allow a null
> > "this" pointer and Symbian does not.  I can't think of a way to 
> > make a better test.
> 
> Either ASSERT earlier or make a specific layout test only executed on Symbian (look in LayoutTests/platform/qt, we have different sets of Skipped tests, e.g. one with Qt 4.8).

And it doesn't matter if you don't cover other platforms in your tests because anyway they didn't crash earlier. What you want to make sure is that Symbian is not crashing. So just let the test like it is, run it on all platforms (UNIX will never crash) but Symbian will (without the patch) and is covered if a regression could happen.

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