[Glade-users] Window communication.

[damon ximian com (Damon Chaplin)]

On 16 Aug 2001 13:16:28 -0700, James Campbell wrote:
> Hello,
>
> I'm running into an interesting problem that I thought was covered
> in the FAQ.
>
> I've got an ATM networking application that spawns off multiple
> window threads to actually display sending and receiving ATM
> cells.
>
> The spawned windows are normal GTK top level windows, with
> gnome_less widgets in them displaying from a buffer.
>
> The problem I'm getting into is having the main callback functions
> be able to hit the non-global widgets in the spawned windows.
>
> I've tried the solution in the glade faq, but I can't seem to get it
> to work properly.
>
> Example (simplified):
>
> 1. main_window = lookup_widget(GTK_WIDGET(button), "main_window");
> 2. window = create_window();
> 3. gtk_widget_show(window);
> 4. gtk_object_set_data(GTK_OBJECT(main_window), "window" window);
> 5. windows = gtk_object_get_data(GTK_OBJECT(main_window), "window");
> 6. window_less = lookup_widget(windows, "windowless");
>
> Here is what I think I'm doing:
>
> Creating a pointer from the main window (and subsequent callback functions)
> to one of the spawned windows (4).
>
> Then calling the pointer to get data about the spawned window and set it to
> a variable (5).
>
> Finally using lookup_widget on the pointer I get from (5) to find the
> gnome_less
> widget "windowless" that is in one of the toplevel windows.
>
> I know I'm getting something back from all that mess, as lookup_widget doesn't
> barf, but when I try to display something on the widget, it crashes without
> much explaination.

The idea looks OK. Note that you have to be careful that you don't try
to use these pointers after the windows are destroyed.

I'd suggest using debugging messages like

  g_print ("Window: %p\n", window);

to see whereabouts it is going wrong.

Didn't the stack trace give any clue?

Damon