Page 1 of 1

Fatal signal: Segmentation Fault (SDL Parachute Deployed)

Posted: 09 Mar 2005, 16:08
by oiper
Any word on this error. Others in game have reported it.

Code: Select all

Fatal signal: Segmentation Fault (SDL Parachute Deployed)
All players have crashed out at the same time on this error. Usualy it's just a singular case though.

Posted: 09 Mar 2005, 17:52
by maci
well i guess the server crashed ;)

Posted: 09 Mar 2005, 20:08
by oiper
Something like that. :roll:

Does anyone keep logs of crashes? So that we know when it's us or the server that dies?

Posted: 10 Mar 2005, 01:12
by Bjørn
No that's not a server crash, that's a client crash. That's all it says though, the client crashed. It's absolutely no use to us without further information about what you were doing, or a back trace of the stack (often most useful). For the latter you'd need to run tmw in a debugger, like gdb.

Posted: 10 Mar 2005, 02:38
by maci
All players have crashed out at the same time on this error. Usualy it's just a singular case though.

Posted: 10 Mar 2005, 02:45
by Bjørn
maci wrote:
All players have crashed out at the same time on this error. Usualy it's just a singular case though.
Yes but still, when the server crashes, the clients all crash as well? Would be nice if we could get a backtrace of that to pinpoint this problem...

Posted: 10 Mar 2005, 18:07
by nachdenki

Code: Select all

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1220643296 (LWP 13930)]
0x42074501 in _int_free () from /lib/tls/libc.so.6
(gdb) backtrace
#0  0x42074501 in _int_free () from /lib/tls/libc.so.6
#1  0x420734d6 in free () from /lib/tls/libc.so.6
#2  0xb7d41421 in operator delete(void*) (ptr=0x52)
    at ../../.././libstdc++-v3/libsupc++/del_op.cc:40
#3  0x0808547b in Being::clearPath() (this=0x42131300) at being.cpp:137
#4  0x0808bf16 in do_parse() () at game.cpp:557
#5  0x0808f6dc in game() () at game.cpp:118
#6  0x08092121 in main (argc=1, argv=0x52) at main.cpp:371
#7  0x42015574 in __libc_start_main () from /lib/tls/libc.so.6
that's what I got today :(
(It's not the first time the client crashes XD but the first time I was using the debugger to do a backtrace)

Posted: 10 Mar 2005, 23:58
by Bjørn
Well I just made one stability fix in being path handling, but it was in nextStep, not clearPath. Anyway, keep sending in your crashes, prerably against latest CVS, because of course we take every crash serious.

Posted: 11 Mar 2005, 14:52
by nachdenki

Code: Select all

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1220643296 (LWP 3577)]
Being::nextStep() (this=0x81d7078) at being.cpp:192
192             new_x = path->x;
Mhhh I think I have the latest cvs version (there are no newer updates). That happend after death.
Maybe there is still something wrong ? I took a look by myself but I wasn't able to find anything, the error makes no sense for me XD but that doesn't mean anything :o (there must be something wrong with path->next XD but you'll figure that out by yourself XD)

Posted: 11 Mar 2005, 16:06
by Bjørn
Well if you have the latest version the check above should now read:

if (path && path->next) {

instead of:

if (path->next) {

If you have the first version, then there would indeed still be some other problem there.

Posted: 11 Mar 2005, 17:19
by nachdenki
jep, it's still like that.
I got another crash as well

Code: Select all

#0  void std::list<Being*, std::allocator<Being*> >::merge<BeingCompare>(std::list<Being*, std::allocator<Being*> >&, BeingCompare) (this=0xbffff5c0, 
    __x=@0xbffff7e0) at being.cpp:130
#1  0x08085bfb in void std::list<Being*, std::allocator<Being*> >::sort<BeingCompare>(BeingCompare) (this=0xbffff7e0) at list.tcc:360
#2  0x08085967 in sort() () at being.cpp:108
#3  0x0808a951 in Engine::draw() (this=0x80f0ce8) at engine.cpp:284
#4  0x0808f6ca in game() () at game.cpp:116
#5  0x08092121 in main (argc=1, argv=0x10) at main.cpp:372
#6  0x42015574 in __libc_start_main () from /lib/tls/libc.so.6
erm, it's weird, both. mhh it may have something to do with keys... I think it crashs a lot more when I press repeatly ctrl (+g) when I come back from death (but I'm not sure :( ), moreover in my bubble is some weird stuff then (small boxex like [], red font color ... ) I dunno whether that's normal or not.

Posted: 11 Mar 2005, 23:49
by Bjørn
Not normal. I'm currently working on encapsulating player speech within the being object, so that there is more control over it, maybe there's a bug with that somewhere.

Posted: 13 Mar 2005, 17:00
by Bjørn
Well I finished that stuff and discovered another strange piece of code which I fixed. I haven't been able to crash it after these changes, so please try latest CVS once again to make sure.

Posted: 15 Mar 2005, 18:55
by maci
nice

well i was able to crash tmw after entering about 1KB to the chat box ;)