Re: [gtkmm] Glade-2, libgtkmm, Xml::create, callbacks

[Daniel Serpell <daniel atmlab utfsm cl>]

El Thu, Sep 04, 2003 at 05:48:36PM +0200, Murray Cumming escribio:
> 
> I think that would be rather bizarre, and I'd rather that you thought
> about the alternative suggestions in my other email.
> 

I suppose that you are referig to this:
> As for specifying signal handlers in the RAD:
> In theory the RAD could do some of this. I think you'd need to:
> - give each widget a name in Glade (Already possible)
>   This will allow you to find the instance that will emit the signal, at
> run time.
>   You might need to consider the parent-child relationship of the
> widgets too. For instance, you might have 2 instances of your mywindow
> window, but you don't want the 2nd one to handle signals from the button
> in the first one.
> - specify the name of the widget whose class will handle the signal, as
> well as the name of the signal handler method.
> - deal with C++ name-mangling to find a pointer to the method that will
> handle the signal.
>

I don't quite undesrstand you, because I don't see the relation between
the gtk+ object instance created by libglade and the c++ object instance
that you want to receive the method:, I see that the name given in glade
is the name of a gtk+ object instance, created at run-time by libglade,
not associated with any c++ class in user's code.

> I don't think the RAD could do everything you want, because some of your
> intent can not be expressed without code.

Ah, perhaps you are saying that you can connect gtk+ signals with gtk+
pre-defined handlers, but this is already possible (if you set the
handler in a button to "gqt_main_quit" in glade, and then call the
autoconnect function, it works).

> As I said in the other email, and as I've been saying for ages,
> libglademm 1 did this, and I'd like someone to investigate that.

Thanks, I'll look at it tonight.

    Daniel.