[Webkit-unassigned] [Bug 32437] [Qt] [LayoutTestController]Output of document.write() is vanished by waitUntilDone()/notifyDone()

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=32437





--- Comment #18 from Andras Becsi <abecsi at inf.u-szeged.hu>  2009-12-14 07:06:27 PST ---
(In reply to comment #17)
> (In reply to comment #14)
> > (In reply to comment #12)
> > > Created an attachment (id=44789)
 --> (https://bugs.webkit.org/attachment.cgi?id=44789) [details] [details] [details]
> > > Fix style error and renamed m_isLoading to m_hasDumped
> > > 
> > > I think m_isLoading is better off named as m_hasDumped and set to false when
> > > the test starts. With the introduction of m_loadFinished, it's purpose is to
> > > check whether DRT still needs to dump the results or not.
> > 
> > ok then. is m_loadFinished needed then? 
> > 
> 
> Yes, two variables are needed: one to tell DRT if the page has finished loading
> or not, and another to tell DRT if it's OK to dump. m_isLoading was trying to
> do both, which I think caused the problems.
> 
> notifyDone needs to know that the load has finished before it can dump, whereas
> maybeDump needs to know if notifyDone has already dumped. 
> 
> > btw, I don't think we want to add methods for resetting all variables:
> > 
> > +    void resetLoadFinished() { m_loadFinished = false; }
> 
> Layout tests can have multiple page loads so m_loadFinished needs to be set to
> false every time a new one is started, this prevents NotifyDone from dumping
> prematurely.

True, they can have multiple page loads. But they also can have a child
windows, which is created in QWebPage *DumpRenderTree::createWindow().

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