Re: [XaraXtreme-dev] XCode - built Mac Version will not display wxXtra SplashScreen

["Ben Fowler" <ben.the.mole@xxxxxxxxx>]

On 01/07/06, Alex Bligh <alex@xxxxxxxxxxx> wrote:
> > [ snip ]
> >
> > On (a) isn't it sufficient that the 'this' pointer is the same -- have
> > I shown that it is the same? Am I missing something about C++ and
> > constructors?
>
> I don't think we've checked that the (consistent) this pointer ties
> up with event.EventObject (i.e. what is being passed to IsKindOf() ).

I'll take a look at that.

> > b)
> >   i) I think that wxFrame is a base class of wxAdvSplashScreen, and
> > this is why its constructor is called before wxAdvSplashScreen's runs.
>
> Ah. I was misreading the stack dump. I thought it was wxAdvSplashScreen's
> constructor that was calling ::Create in wxFrame. Actually the "forked"
> construction mechanism (i.e. calling ::Create is done at the wxFrame level),
> so at the point it gets into the "Create" loop, wxFrame's constructor has
> not completed, so the VFT pointer in the object is still pointing to a
> wxFrame VFT and this is confusing the IsKindOf() stuff.

Sounds reasonable. Given that TLWs need to murder the
wxAdvSplashScreen, it is tight that wxAdvSplashScreen is itself a TLW.
Whilst one possibilitity would be to check ahead  of
CamResource::DoneInit() that we are not trtying kill the window we are
constructing.

> So I think I know how to fix this cleanly - do a separate call into
> Frame's Create function rather than rely on wxFrame to do it, because
> the VFT pointer by then will be set to wxAdvSplashScreen's VFT as the
> underlying constructors will have exited even though Create() won't yet
> have been called. I'll try to fix that up this PM.

Thanks.

> >   ii) I need to check that I am using the mac/carbon/frame.h
>
> OK.

This is so.

Ben