[Gambas-devel] Some wierd gdb behaviour (I think)

[Ron]

Ron schreef:
> Hi,
>
> Been trying to debug my program, because after adding some functionality 
> every now and then it crashes (gambas), before that it was rock stable.
> But upgraded from 2.8.1 to 2.8.2 in the meantime also...
>
> I compile and start my program like the website says, this is what I get:
>
> -----------------
> ron at ...218...:~/domotiga/DomotiGa$ gdb gbx2
> GNU gdb 6.4.90-debian
> Copyright (C) 2006 Free Software Foundation, Inc.
> GDB is free software, covered by the GNU General Public License, and you are
> welcome to change it and/or distribute copies of it under certain 
> conditions.
> Type "show copying" to see the conditions.
> There is absolutely no warranty for GDB.  Type "show warranty" for details.
> This GDB was configured as "i486-linux-gnu"...Using host libthread_db 
> library "/lib/tls/i686/cmov/libthread_db.so.1".
>
> (gdb) set args -p
> (gdb) run
> Starting program: /usr/local/bin/gbx2 -p
> Failed to read a valid object file image from memory.
> [Thread debugging using libthread_db enabled]
> [New Thread -1209992992 (LWP 8372)]
> [New Thread -1231660112 (LWP 8376)]
> [New Thread -1242149968 (LWP 8377)]
> [Thread -1242149968 (zombie) exited]
> [Thread -1231660112 (LWP 8376) exited]
> [New Thread -1231660112 (LWP 8378)]
> [New Thread -1242149968 (LWP 8379)]
> [New Thread -1250542672 (LWP 8380)]
> [Thread -1250542672 (zombie) exited]
> [New Thread -1250542672 (LWP 8381)]
> open /dev/sequencer: No such file or directory
> Qt: gdb: -nograb added to command-line options.
>          Use the -dograb option to enforce grabbing.
> [New Thread -1266635856 (LWP 8385)]
> [Thread -1266635856 (LWP 8385) exited]
> [New Thread -1266635856 (LWP 8386)]
> [Thread -1266635856 (LWP 8386) exited]
> [New Thread -1266635856 (LWP 8390)]
> [Thread -1266635856 (LWP 8390) exited]
> [New Thread -1266635856 (LWP 8391)]
> [Thread -1266635856 (LWP 8391) exited]
> Segmentation fault (core dumped)
> ron at ...218...:~/domotiga/DomotiGa$
> --------
>
> the first thing that worries me, are the 2 zombie threads in the 
> beginning, is this normal?
>
> Then when I click a certain menu, which lanches a part of my program 
> which calls rrdtool to generate graphs (SHELL sRRDCmd2 WAIT) were 
> sRRDCmd2 is a large command string which calls rrdtool graph...
>
> the gdb session gives Segmentation fault (core dumped)
> ron at ...218...:~/domotiga/DomotiGa$
> and the gdb session is closed, my program continues to run...
>
> Is this normal? Because I started another program in the shell? My 
> gambas program still runs and functions ok after this...
> What is wrong? This way I cannot debug my program without removing some 
> functionality first.
> Did I overlook something?
>
> Gambas 2.8.2 on Debian 4.0etch3
>
> Wanted to run it as root, but it's a headless server, and cannot get X 
> to run remote for root... will have to look tomorrow...
> Do you need to run it as root to get the debugging info (see the Failed 
> to read a valid object file image from memory. error)
>
> Regards,
> Ron_2nd
>
>   


OK,
so I compiled gdb 6.8 from scratch, and run my program with that, this 
is the first result:

ron at ...218...:~/domotiga/DomotiGa.112ok$ gdb gbx2
GNU gdb 6.8
Copyright (C) 2008 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later 
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "i686-pc-linux-gnu"...
(gdb) set args -p
(gdb) run
Starting program: /usr/local/bin/gbx2 -p
Failed to read a valid object file image from memory.
[Thread debugging using libthread_db enabled]
[New Thread 0xb7dca8e0 (LWP 4648)]
Qt: gdb: -nograb added to command-line options.
         Use the -dograb option to enforce grabbing.
[New Thread 0xb654cbb0 (LWP 4660)]
[Thread 0xb654cbb0 (LWP 4660) exited]
[New Thread 0xb5d4bbb0 (LWP 4661)]
[Thread 0xb5d4bbb0 (LWP 4661) exited]
[New Thread 0xb654cbb0 (LWP 4662)]
[Thread 0xb654cbb0 (LWP 4662) exited]
[New Thread 0xb654cbb0 (LWP 4666)]
[Thread 0xb654cbb0 (LWP 4666) exited]
...
[New Thread 0xb654cbb0 (LWP 5586)]
[Thread 0xb654cbb0 (LWP 5586) exited]
...

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb7dca8e0 (LWP 4648)]
0xb700e2bb in ?? () from /usr/lib/libcurl.so.3
(gdb) bt
#0  0xb700e2bb in ?? () from /usr/lib/libcurl.so.3
#1  0xb6feefa0 in ?? () from /usr/lib/libcurl.so.3
#2  0x080d6970 in ?? ()
#3  0xb7f65ff4 in ?? () from /lib/ld-linux.so.2
#4  0x080d66d8 in ?? ()
#5  0x00000001 in ?? ()
#6  0x00000000 in ?? ()
(gdb)


Ok, so I miss important debug information, dunno why yet, but at least 
gdb itself is not crashing like the 6.4 version did.
But am I correct to notice the sigsegv is due to problem in/with libcurl?

Any changes in gb.net.curl recently, say last 1 1/2 week?
Anyone using it in their program?

Will try to weed out the gb.net.curl dependant code from my program, to 
see if this is the only problem.

Any ideas as  how to get more debugging info inside my gdb would be fab.

Debian 4.0 etch 3, qt, Gambas 2.99 recent svn

Regards,
Ron_2nd