From: Christoph Egger <Christoph_Egger@t-online.de>
  To  : ggi-develop <ggi-develop@eskimo.com>
  Date: Fri, 26 Nov 1999 15:24:45 +0100 (MET)

Re: libggi bug

On Thu, 25 Nov 1999, James Simmons wrote:

> 
> howdy!!!
> 
>    Well I was testing the CVS snap shot of libggi from today. Every
> program dumps core. I run gdb on flying_ggis. The problem is in
> ggiClose. Running threw I get this
> 
> 573             ggiClose(vis);
> (gdb) step
> ggiClose (visual=0x8050a00) at init.c:357
> 357             ggi_visual *vis,*pvis=NULL;
> (gdb) next
> 359             GGIDPRINT_CORE("ggiClose(\"%p\") called\n", visual);
> (gdb) 
> 361             if (!_ggiLibIsUp) return GGI_ENOTALLOC;
> (gdb) 
> 363             GGIDPRINT_CORE("ggiClose: closing\n");
> (gdb) 
> 365             for (vis = _ggiVisuals.visual; vis != NULL; pvis=vis,
> vis=vis->next) {
> (gdb) 
> 366                     if (vis == visual) break;
> (gdb) 
> 369             if (vis == NULL) return GGI_EARGINVAL;
> (gdb) 
> 371             ggLock(_ggiVisuals.mutex);
> (gdb) 
> 373             if (pvis == NULL) _ggiVisuals.visual = vis->next;
> (gdb) 
> 376             _ggiVisuals.visuals--;
> (gdb) 
> 378             ggUnlock(_ggiVisuals.mutex);
> (gdb) 
> 380             _ggiDestroyVisual(vis);
> (gdb) 
> 
> Program received signal SIGSEGV, Segmentation fault.
> _GGI_mansync_deinit (vis=0x8050a00) at child.c:265
> 
> This problem also locks the machine on fbcon. I will lock more into
> this. I wish I had two monitors :(
> 
>       -----==-                                                  
>       ----==-- _                                           
>       ---==---(_)__  __ ____  __       James Simmons
>       --==---/ / _ \/ // /\ \/ /       jsimmons@edgeglobal.com
>       -=====/_/_//_/\_,_/ /_/\_\       fbcon/gfx developer
>     The choice of a GNU generation
>


The bug you have reported is already known for some months.

I found it, when I tested a demo of my 3DtoolKit-Library
(http://home.t-online.de/home/christoph_egger/3dtk/).

All demos crashs on exiting at the same file child.c:265. Because this is
very annoying, I went back using the ggi-devel-990905.tar.gz -CVS-snapshot.
This is stable. _No_ crash! I can send it to you, if you
want (2812351 Byte size).

Andy was the first person, who I informed about that. When I remember
right, he has informed Marcus about that (Andy correct me, if I am
wrong!).


Cheers,

Christoph Egger
E-Mail: Christoph_Egger@t-online.de