Re: debugging the gtk fe
As i told you before i'm not very good at debugging code, but from your message
i've seen the problem is related to argc and argv pointers: they are hand-created
at line 1370 of gtk.c this way
/* FIXME: This can surely be done in a better way */
(char**) name = malloc(2 * sizeof(char*));
(char*) name[0] = malloc(8 * sizeof(char));
name[0] = "debconf";
name[1] = NULL;
and used this way
gtk_init (&args, &name);
those lines used to be cause of some (not many) strange crashes even
in an X environment.
may be this related to the crashes?
attilio
>-- Messaggio Originale --
>Date: Thu, 28 Jul 2005 16:27:00 +0200
>From: zinosat@tiscali.it
>Subject: Re: debugging the gtk fe
>To: debian-boot@lists.debian.org
>Cc: cjwatson@debian.org
>
>
>Here's what Mike has found out:
>
>Okay it looks like a bug outside of gtk Davide and I got a standalone dev
>env its attached. Bottom line is it looks like its outside the gtk code
at
>least now. I've got no idea what going on with the forked process etc etc.
>I've never attached gdb through forks but it looks like one of the child
>process crashed prob not the frontend.
>
>In short I don't see what I can do to help here except I do note that mutiple
>frontends are supported do the otheres work. Failing that a few printfs
to
>track the initialization of the gtk fronted so you could attach and break
>in the debugger before the crash is possible. I'm not in a position to build.
> If the bug is only in the gtk frontend then printfs and commenting out
code
>can isolate the problem area so you know where to attach the debugger thats
>what I have to do.
>
>Mike
>
>
>Here is and important email
>Hi Mike,
>I should mention that last night I did not find what the cause of the crash
>is, but I managed to grab an interesting gdb backtrace; atm I'm at work
and
>I don't have it. last night I posted it on the #debian-boot IRC channel,
>and here is a message that could be interesting
>
>http://lists.debian.org/debian-boot/2005/07/msg01014.html
>
>let me know if you managed to start debconf
>
>
>Here is the trace
> ---------------------- DirectFB v0.9.23 ---------------------
> (c) 2000-2002 convergence integrated media GmbH
> (c) 2002-2004 convergence GmbH
> -----------------------------------------------------------
>
>(*) DirectFB/Core: Single Application Core. (2005-07-21 22:02)
>(*) Direct/Memcpy: Using MMXEXT optimized memcpy()
>[New Thread 27106224 (LWP 14601)]
>(*) Direct/Thread: Running 'SDL Input' (INPUT, 14601)...
> (!!!) *** UNIMPLEMENTED [fusion_reactor_set_lock] *** [reactor.c:802]
>(*) DirectFB/Input: SDL Input 0.1 (Denis Oliver Kropp)
>(*) DirectFB/Genefx: MMX detected and enabled
>(*) DirectFB/Graphics: MMX Software Rasterizer 0.6 (convergence integrated
>media GmbH)
>(*) DirectFB/Core/WM: Default 0.2 (Convergence GmbH)
>[New Thread 66788272 (LWP 14602)]
>(*) Direct/Thread: Running 'EventBufferFeed' (MESSAGING, 14602)...
>
>(debconf:14598): Gdk-DirectFB-WARNING **: gdk_display_request_selection_notification
>Unimplemented function
>
>
>(debconf:14598): Gdk-DirectFB-WARNING **: gdk_window_set_keep_above() not
>implemented.
>
>
>(debconf:14598): Gdk-DirectFB-WARNING **: gdk_window_set_keep_below() not
>implemented.
>
>Detaching after fork from child process 14604.
>
>Program received signal SIGSEGV, Segmentation fault.
>[Switching to Thread -1208772928 (LWP 14598)]
>0x00000000 in ?? ()
>(gdb) bt
>#0 0x00000000 in ?? ()
>#1 0x006d2379 in frontend_delete (obj=0x9fed348) at frontend.c:208 #2
0x08048a30
>in cleanup () at debconf.c:41 #3 0x08048abf in sighandler (sig=17) at debconf.c:65
>#4 <signal handler called> #5 0x00fd8402 in __kernel_vsyscall () #6 0x0080bd3b
>in __read_nocancel () from /lib/libc.so.6 #7 0x006cf833 in confmodule_communicate
>(mod=0xa0f2a20) at confmodule.c:77 #8 0x0804918a in main (argc=4, argv=0xbf85f134)
>at debconf.c:153
>(gdb) quit
>
>
>
>
>On 7/27/05, Mike Emmel <mike.emmel@gmail.com> wrote:
>> On 7/27/05, Attilio Fiandrotti <fiandro@tiscalinet.it> wrote:
>> > since i'm very interested in the debugging of the frontend, could
>> > you add me in the cc list?
>>
>> Sure no problem your on for me. At the moment nothing happening I'm
missing
>>
>> linux-gate.so.1 => (0x00815000)
>> libdebconf.so => not found
>> libdl.so.2 => /lib/libdl.so.2 (0x00357000)
>> libc.so.6 => /lib/libc.so.6 (0x00205000)
>> /lib/ld-linux.so.2 (0x001e7000)
>> libdebian-installer.so.4 => not found
>>
>> And want the source for the gtk part.
>>
>> Mike
>>
>>
>> I'm currently missing these libs
>>
>>
>> >
>> > thanks
>> >
>> > attilio
>> > Mike Emmel wrote:
>> > > On 7/27/05, Davide Viti <zinosat@tiscali.it> wrote:
>> > >
>> > >>I recompiled DirectFB with --enable-debug=yes
>> > >>here are some more details:
>> > >>
>> > >>http://www.webalice.it/zinosat/dfb_2.7_dbgen2.png
>> > >>http://www.webalice.it/zinosat/dfb_2.7_dbgen.png
>> > >>
>> > >>hope this helps,
>> > >>Davide
>> > >>
>> > >
>> > > Yep you crashed :(
>> > >
>> > > I've got your app I'll try and run it.
>> > >
>> > > This is the last time I'll cc the list we can just continue with
>> > > private emails. Mine is mike.emmel@gmail.com if its not showing
>> > > for you.
>> > >
>> > > Mike
>> > >
>> > > _______________________________________________
>> > > directfb-dev mailing list
>> > > directfb-dev@directfb.org
>> > > http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
>> > >
>> >
>> >
>>
>
>
>__________________________________________________________________
>TISCALI ADSL 1.25 MEGA
>Solo con Tiscali Adsl navighi senza limiti e telefoni senza canone Telecom
>a partire da 19,95 Euro/mese.
>Attivala entro il 28 luglio, il primo MESE è GRATIS! CLICCA QUI.
>http://abbonati.tiscali.it/adsl/sa/1e25flat_tc/
>
>
>
__________________________________________________________________
TISCALI ADSL 1.25 MEGA
Solo con Tiscali Adsl navighi senza limiti e telefoni senza canone Telecom
a partire da 19,95 Euro/mese.
Attivala entro il 28 luglio, il primo MESE è GRATIS! CLICCA QUI.
http://abbonati.tiscali.it/adsl/sa/1e25flat_tc/
Reply to: