[Webkit-unassigned] [Bug 29803] [Qt] QWebHistory and QWebPage autotests crash

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=29803





--- Comment #10 from Simon Hausmann <hausmann at webkit.org>  2009-09-30 05:31:41 PDT ---
(In reply to comment #9)
> (In reply to comment #8)
> >  QWebPage::~QWebPage()
> >  {
> > +    d->createMainFrame();
> > 
> > Why does it make sense to create a main frame in the destructor? Shouldn't
> > there instead be a test to see if it was not created in the following code?
> > 
> > At least I think it deserves a comment.
> 
> Actually d->createMainFrame() is a test, it test if mainFrame was created.
> MainFrame should be created as soon as possible, but not in constructor because
> of signals. WebCore::Page really depends on mainFrame and use it in different
> places (in destructor too). 
> 
> Maybe we should move all checks to WebCore::Page, but:
> - the change will affect more platforms, not only Qt
> - it will be much more difficult to control mainFrame creation time. If Page
> automagically will create frame how to emit signal?
> In general it is possible but what we can gain with it?

I think we should remove the creation from the QWebPage destructor and add just
that one single if in the Page destructor.

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