gobby + sobby segfault immediately upon starting a session

[alex]

I'm using the gentoo ebuilds for gobby-0.3.0 etc. If I run sobby, I get this output: Generating RSA key pair... Opening session... Segmentation fault

Gobby opens up fine, but if I try to create a session or join a session, it segfaults immediately. I'm sorry I don't know how to get any more detailed debug information. Please let me know if I can be of more help in debugging this.

[anonymous]

You may use the GNU debugger (GDB) to supply us with a backtrace: Type 'gdb {s|g}obby' in a console and then 'run' (at the gdb prompt, as soon as it appears). Make gobby/sobby crash and enter 'bt' in gdb. Post its output here.

[armin]

Oops, I wasn't logged in when I wrote that. Sorry.

[alex]

sobby through gdb: enerating RSA key pair... Opening session...

Program received signal SIGSEGV, Segmentation fault. 0xb7bc205c in memcpy () from /lib/libc.so.6

That's using stripped binaries, I'll re-emerge with FEATURES="nostrip". Here's a sobby backtrace: (gdb) bt #0 0xb7bc205c in memcpy () from /lib/libc.so.6 #1 0xb7d1d0c8 in std::string::_Rep::_M_clone () from /usr/lib/gcc/i686-pc-linux-gnu/3.4.4/libstdc++.so.6 #2 0xb7d1d1c4 in std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string ()

from /usr/lib/gcc/i686-pc-linux-gnu/3.4.4/libstdc++.so.6

#3 0x080892f4 in obby::serialise::attribute::set_value<unsigned int> () #4 0x0808b81e in obby::serialise::attribute::set_value<unsigned int> () #5 0xb7b6f56d in libc_start_main () from /lib/libc.so.6 #6 0x08050761 in ?? ()

As I said, I'll give non-stripped binaries a go.

[alex]

Formatted properly

(gdb) bt
#0  0xb7bc205c in memcpy () from /lib/libc.so.6
#1  0xb7d1d0c8 in std::string::_Rep::_M_clone () from /usr/lib/gcc/i686-pc-linux-gnu/3.4.4/libstdc++.so.6
#2  0xb7d1d1c4 in std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string ()
   from /usr/lib/gcc/i686-pc-linux-gnu/3.4.4/libstdc++.so.6
#3  0x080892f4 in obby::serialise::attribute::set_value<unsigned int> ()
#4  0x0808b81e in obby::serialise::attribute::set_value<unsigned int> ()
#5  0xb7b6f56d in __libc_start_main () from /lib/libc.so.6
#6  0x08050761 in ?? ()

[alex]

Compiled with -ggdb, unstripped:

(gdb) run
Starting program: /usr/bin/sobby
Generating RSA key pair...
Opening session...

Program received signal SIGSEGV, Segmentation fault.
0xb7c0205c in memcpy () from /lib/libc.so.6
(gdb) bt
#0  0xb7c0205c in memcpy () from /lib/libc.so.6
#1  0xb7d5d0c8 in std::string::_Rep::_M_clone () from /usr/lib/gcc/i686-pc-linux-gnu/3.4.4/libstdc++.so.6
#2  0xb7d5d1c4 in std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string ()
   from /usr/lib/gcc/i686-pc-linux-gnu/3.4.4/libstdc++.so.6
#3  0x08087f61 in Server (this=0xbfe86110, argc=1, argv=0xbfe86194) at ustring.h:862
#4  0x0808a50e in main (argc=-1348243436, argv=0xafa37014) at main.cpp:27

[phil]

Could you please use "up" until you reach the "Server" part and then issue a "print this"? This should give us a dump of the server object with all member variables.

[phil]

Could you by chance step through the program and collect some variables on the way? It is quite difficult for us to track it down remotely.

[alex]

I've never used gdb before today, so right now I'm not sure how to go about stepping through the program. I can do some reading tomorrow though. I issued a 'print this' after going 'up' to the server object. It doesn't look very useful at all to me, but what do I know.

(gdb) up
#3  0x08087f61 in Server (this=0xbfdb53d0, argc=1, argv=0xbfdb5454) at ustring.h:862
862     {
(gdb) print this
$1 = (class Sobby::Server * const) 0xbfdb53d0

I I go up again it shows this:

#4  0x0808a50e in main (argc=-1348042732, argv=0xafa68014) at main.cpp:27
27      main.cpp: No such file or directory.
        in main.cpp

[phil]

Sorry, "print *this" would be more right.

Could you please compile sobby from source? (It's as easy as ./configure && make -- No need to install it.) Then try if this copy does also segfault. If so invoke gdb on it, it should now present a more meaningful source line. If not you could just use next several times instead of run, until it crashes.

[phil]

What's the output of the following command?

 gcc -v 2>&1|tail -n 1

[alex]

gcc -v 2>&1|tail -n 1 gcc version 3.4.4 (Gentoo 3.4.4-r1, ssp-3.4.4-1.0, pie-8.7.8)

I'm currently doing some major library rebuilding, I may have a botched gcc 3.3-3.4 transition, in which case I'm sorry for the false bug report.

#3  0x080887c9 in Server (this=0xbfe49940, argc=1, argv=0xbfe499c4) at ustring.h:862
862     {
(gdb) print *this
$1 = {<> = {<No data fields>}, <sigc::trackable> = {callback_list_ = 0x0}, m_port = 6522, m_interactive = false,
  m_main_loop = {pCppObject_ = 0x809bef0}, m_server = {_M_ptr = 0x809d25c}, m_autosaver = {_M_ptr = 0x0},
  static m_cmd_map = @0x809aa88}

[alex]

gobby+sobby work like a charm now, my system was borked. Apologies for the false bug.