Re: [glade--]unable to add new members to derived class

Bob Koutsky wrote:
> Hello,
Hi Bob,

> this is probably something very stupid, but my C++ is little rusty 
> (after years of Java) and I'm completely new to gtkmm & related stuff, 
> so please have a patience:
> 
> I create new gtk+ project using glade - single window containing a 
> label, nothing else. I set the language to C++, generate code, run 
> autogen.sh, run the program, it works. Then I add a single member to the 
> main window class header file:
> 
> #ifndef _WINDOW1_HH
> #  include "window1_glade.hh"
> #  define _WINDOW1_HH
> #include "glademm_support.hh"
> 
> class window1 : public window1_glade
> {  public:
>   int foo;
> };
> #endif
> 
> run make, run the program, and get segmentation fault. If I try to use 
> DDD, it apparently fails on the "Gtk::Main m(&argc, &atgv)" line in 
> function main, and it reports following:
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 1024 (LWP 1120)]
> 0x4049a82e in Glib::ObjectBase::is_derived_ () from 
> /usr/lib/libglibmm-2.0.so.1
> 
> My gcc is 2.95.3, glade-- is 1.1.2 and reports following when generating 
> code:
> Found pkg-config version 0.14.0
> Generating code for gtk 2.0.8 (pkg-config), gtkmm 2.0.1 (pkg-config), 
> [gnomemm 1.2.3 (pkg-config)]
> 
> 
> Can somebody please tell what is the problem and how to fix it? I would 
> like to convert to gtkmm, but I'm starting to miss Java and Swing 
> already 8-)

You encountered one of the most mysterious bugs I've ever seen. Can you do:
   rm *.o
   make
and try it again. I can't think of anything causing this but it's a try.

If this still fails, try to send the package to me (or a full backtrace) 
(please delete *.o, any executables and the like).

Clearly a "should never occur" bug.

    Christof

PS: If you don't see a personal benefit from C++/gtk/gnome, don't 
switch. I never switched to java since I don't see any benefits for me 
(only drawbacks). But this clearly depends on your perspective.