Forward from Karri Niskala <karri.niskala@netmill.fi>:

 Hiya

 While making a little server installation for a new site ... I noticed that
the particular site managed to segfault Apache with php 3.0.15 on 
every hit.

--backtrace--
#0  0x400b1932 in free ()
#1  0x400b17f1 in free ()
#2  0x4018b842 in _efree (ptr=0x820d190) at alloc.c:168
#3  0x40187b7c in pval_destructor (pvalue=0x403ae670) at variables.c:84
#4  0x40189f21 in tc_destroy (tc=0x8325368) at token_cache.c:307
#5  0x40189f81 in tcm_destroy (tcm=0x40203878) at token_cache.c:321
#6  0x40182592 in php3_request_shutdown (dummy=0x0) at main.c:784
#7  0x8050d94 in ap_run_cleanup ()
#8  0x804fc4f in ap_clear_pool ()
#9  0x804fcb2 in ap_destroy_pool ()
#10 0x804fc3d in ap_clear_pool ()
#11 0x805c9da in ap_child_terminate ()
#12 0x805cf5b in ap_child_terminate ()
#13 0x805d068 in ap_child_terminate ()
#14 0x805d54b in ap_child_terminate ()
#15 0x805dc9e in main ()
--backtrace ends--

 I had to come up with something to at least stop the segfaults, 
so I was able keep Apache alive by commenting out the 'efree' call
in  pval_destruct (in variables.c) in the last block.
--
if (pvalue->value.func.arg_types) {
                        efree(pvalue->value.func.arg_types);
                }

--

 That surprisingly isn't a very long lasting solution, so do you
have any idea what might be causing the problem? Or could you
forward  the bug to someone who knows something about these things.