[simspark-devel] problem with OpenGL setup (Mac OS X)

[Joschka B. <jos...@am...>]

Hi all,

I updated my local rcssserver3D source tree from CVS today, compiled,  
and did make install. The code compiles without problems (Mac OS X  
10.5.2), but when I try to start simspark, it crashes on setup of the  
OpenGL stuff:

(spark.rb) sparkSetupRendering
(spark.rb) using OpenGLSystem 'OpenGLSystemSDL'
(spark.rb) creating kerosin/OpenGLServer instance at /sys/server/opengl
Reading symbols for shared libraries . done
2008-04-01 11:18:32.194 simspark[82783:813] ***  
_NSAutoreleaseNoPool(): Object 0x85aac0 of class NSCFNumber  
autoreleased with no pool in place - just leaking
Stack: (0x9168712f 0x91593ec2 0x95c79de1 0x93f98b78 0x93f974f9  
0x93fa7736 0x95cd61aa 0x95cd43b0 0x95cd40f5 0x95cd3aa4 0x95cd35e3  
0x95cd15df 0xb2676 0xa930b....*a lot more of this...*

in gdb, the backtrace looks like this:

Program received signal SIGTRAP, Trace/breakpoint trap.
0x948740f4 in ___TERMINATING_DUE_TO_UNCAUGHT_EXCEPTION___ ()
(gdb) bt
#0  0x948740f4 in ___TERMINATING_DUE_TO_UNCAUGHT_EXCEPTION___ ()
#1  0x93f9c0fb in objc_exception_throw ()
#2  0x95d6fb81 in -[NSWindow _commonAwake] ()
#3  0x95c8b683 in -[NSWindow  
_commonInitFrame:styleMask:backing:defer:] ()
#4  0x95c8ab41 in -[NSWindow  
_initContent:styleMask:backing:defer:contentView:] ()
#5  0x95c8a4fe in -[NSWindow  
initWithContentRect:styleMask:backing:defer:] ()
#6  0x000b551f in -[SDL_QuartzWindow  
initWithContentRect:styleMask:backing:defer:] ()
#7  0x000b2c8c in QZ_SetVideoMode ()
#8  0x000a964e in SDL_SetVideoMode ()
#9  0x00a511e5 in OpenGLSystemSDL::Init (this=0x849fd0) at  
openglsystemsdl.cpp:144
#10 0x001314ac in kerosin::OpenGLServer::Init (this=0x849b90,  
openGLSysName=@0xbfffe0d8) at openglserver/openglserver.cpp:132
#11 0x001320ca in init (obj=0x849b90, in=@0xbfffe138) at openglserver/ 
openglserver_c.cpp:16
#12 0x00580483 in thisCall (objPointer=17381153, functionName=9484720,  
args=2691109104) at scriptserver/scriptserver.cpp:158
#13 0x005ce057 in rb_with_disable_interrupt () at stl_vector.h:123
#14 0x005d718c in rb_eval_string_wrap () at stl_vector.h:123
#15 0x005d7d6a in rb_eval_string_wrap () at stl_vector.h:123
#16 0x005d50da in rb_eval_string_wrap () at stl_vector.h:123
#17 0x005d795f in rb_eval_string_wrap () at stl_vector.h:123
#18 0x005d7d6a in rb_eval_string_wrap () at stl_vector.h:123
#19 0x005ded91 in rb_Array () at stl_vector.h:123
#20 0x005d7ca4 in rb_eval_string_wrap () at stl_vector.h:123
#21 0x005d50da in rb_eval_string_wrap () at stl_vector.h:123
#22 0x005d795f in rb_eval_string_wrap () at stl_vector.h:123
#23 0x005d7d6a in rb_eval_string_wrap () at stl_vector.h:123
#24 0x005d50da in rb_eval_string_wrap () at stl_vector.h:123
#25 0x005e2af2 in rb_apply () at stl_vector.h:123
#26 0x005e3371 in rb_eval_string () at stl_vector.h:123
#27 0x005cd3f5 in rb_protect () at stl_vector.h:123
#28 0x005cd457 in rb_eval_string_protect () at stl_vector.h:123
#29 0x0057f03b in zeitgeist::RbEvalStringWrap (str=@0xa06710f0,  
error=@0xbffff30c) at scriptserver/rubywrapper.cpp:32
#30 0x0057f098 in zeitgeist::ScriptServer::Eval (this=0x80b220,  
command=@0xa06710f0) at scriptserver/scriptserver.cpp:293
#31 0x005811e4 in zeitgeist::ScriptServer::Run (this=0x80b220,  
file=@0xbffff398) at scriptserver/scriptserver.cpp:277
#32 0x0058136b in ~shared_count [inlined] () at scriptserver/ 
scriptserver.cpp:286
#33 ~shared_ptr [inlined] () at detail/shared_count.hpp:209
#34 ~shared_ptr [inlined] () at detail/shared_count.hpp:131
#35 zeitgeist::ScriptServer::Run (this=0x80b220, fileName=@0xbffff408)  
at shared_ptr.hpp:286
#36 0x000025b8 in ~shared_ptr [inlined] () at main.cpp:101
#37 ~shared_count [inlined] () at detail/shared_count.hpp:131
#38 ~shared_ptr [inlined] () at detail/shared_count.hpp:209
#39 SimSpark::InitApp (this=0xbffff4bc, argc=1, argv=0xbffff564) at  
detail/shared_count.hpp:101
#40 0x0000f25c in spark::Spark::Init (this=0xbffff4bc, argc=1,  
argv=0xbffff564) at spark.cpp:118
#41 0x00002063 in main (argc=1, argv=0xbffff564) at main.cpp:118
(gdb)

It seems like the OpenGLSystem::Init method has problems setting the  
video mode or sth. like that. This is weird, because nothing changed  
on my system and the code was running before the update today (I'm not  
sure when I did make install on the whole source tree last time, but  
it was running for sure 2 weeks ago).

Is anybody aware of changes in the last 2 weeks that could cause this  
(I didn't find anything obvious in the ChangeLog and SF is giving me  
internal server error messages when I try to access the sserver- 
commits archive)? Is anybody having similar problems?

Any ideas are appreciated!

Thanks,
Joschka

P.S.: I tried a fresh checkout too, with the same result.