Re: Derived widget in libglademm

[Philipp Klaus Krause <pkk spth de>]

Murray Cumming schrieb:
> On Sun, 2007-09-02 at 00:31 +0200, Philipp Klaus Krause wrote:
>> Murray Cumming schrieb:
>>
>>> I guess I'd need to see a complete compileable simple test case to
>>> figure this out.
> 
> Here is a self-contained test that uses libglademm and a test that does
> not, with the build command in a comment at the top of each file.
> 
> I see that on_realize() is indeed not called when using
> get_widget_derived(). From the g_warning() about using
> set_gl_capability() after realize, it seems that the realize signal has
> already been called before our C++ instance was constructed. That's
> unusual.
> 
> However, the default signals are being used in general. I added an
> on_expose_event() override to show that.
> 
> You are not the first person to discover this, it seems:
> http://lists.ximian.com/pipermail/glade-users/2005-March/002430.html
> And here's a message about the same problem with regular libglade:
> http://mail.gnome.org/archives/gtk-list/2006-June/msg00065.html
> which maybe has your solution, though it seems odd.
> "
> Solution: don't enable the visible attribute on widgets that you want 
> to catch the realize event and show this widgets manually with your 
> own code
> "

Thanks a lot for your help. I'm new to libglademm and it's been a few
years since I used gtkmm.

I solved the problem using a "creation function" (an extern "C" wrapper
around my constructor, looks somehow ugly in a C++ program). Now
on_realize() is called. I've posted a .tar.gz of a working program on
the list. on_realize(), on_configure_event(), on_expose_event() are
called correctly. The visible attribute is enabled in glade.

There's a small working sample program to the list in my "gtkglextmm
with libglade example - source attached" posting.

Philipp
P.S.: I currently have a problem with on_button_pres_event() not being
called, if I don't find a solution I'll post on the list soon.