I'm encountering the following problem:

Objects/classobject.c:631: failed assertion `g->gc.gc_refs != 
_PyGC_REFS_UNTRACKED'
Abort

Program received signal SIGABRT, Aborted.
0x900429ac in kill ()
(gdb) bt
#0  0x900429ac in kill ()
#1  0x9009eb1c in abort ()
#2  0x95fec53c in __eprintf ()
#3  0x95f4e1d8 in PyInstance_New ()
#4  0x95fa9600 in PyEval_GetFuncDesc ()
#5  0x95fa6c64 in PyEval_EvalCode ()
#6  0x95fa7e30 in PyEval_EvalCodeEx ()
#7  0x95f5f354 in PyFunction_SetClosure ()
#8  0x95f4a8d0 in PyObject_Call ()
#9  0x95fa91ec in PyEval_CallObjectWithKeywords ()
#10 0x000aea30 in pypeak_stream_event_callback (s=0x0, 
type=406112, self=0x3) at peakmodule.c:499

(The stack seems corrupted as arguments make no sense in #10)

I use the current version of latest Mac OS X:
Python 2.3 (#1, Sep 13 2003, 00:49:11) 
[GCC 3.3 20030304 (Apple Computer, Inc. build 1495)] on darwin

Logged In: YES 
user_id=342163

Damn, I just found the problem. It isn't the core but simply my C module 
where I forgot to increase the refcount in an accessor method. Sorry for 
the trouble. I'm a python users since a few days and enjoying its great 
extensibility ;-)

Logged In: YES 
user_id=31435

Ya, the "peakmodule.c" in the traceback kinda revealed the 
conclusion at once <wink>.  It's almost certain that failures 
of this kind don't occur due to core code, since that's run by 
hundreds of thousands of people every day.

I'm delighted to see that you're running a debug-build 
Python!  That's the best way to minimize debugging time for 
your own C extensions (of course the Python developers use 
debug builds extensively for Python development too).