Could you please compile PHP with debug symbols and try 
again. That would make the backtrace much more informative. 

I'm sorry, when compiling with debug-symbols (--enable-debug) it doesn't seem to happen...

Perhaps you can use this extra bit of information:
gcc = 3.3.2
glibc = 2.3.2
postgresql = 7.4.1
apache = 2.0.48

Even weirder, when I start apache2 -X, its opening two processes. The manager (I suppose) and one handler (I suppose). The manager process dies, the handler seems to survive (at least, the one with the higher pid and which is displayed in the ps xaufww output as the lower one).
When I attach gdb to that process, it doesn't really do anything, but if I attach gdb to the top process, it tells me there is a segfault and I can get the above stacktrace again.

If you have any clues in how to build a proper debuggable php, please let me know. My config.nice contains:
CFLAGS='-march=athlon-tbird -O3 -pipe' \
CXXFLAGS='-march=athlon-tbird -O3 -pipe' \
CC='gcc' \
CXX='g++' \
'./configure' \
'--with-apxs2=/usr/sbin/apxs2' \
'--with-zlib' \
'--with-mysql' \
'--with-pgsql'

I'll check whether it works correctly without the parameters, but they shouldn't make any difference.

Change your CFLAGS and CXXFLAGS to 
 
CFLAGS='g3 -pipe' \ 
CXXFLAGS='-g3 -pipe' \ 
 
Are you using apache2 with fork() or thread model? 

Ok, it's a bit clearer now:
Program received signal SIGSEGV, Segmentation fault.
0x4055e425 in php_sqlite_callback_invalidator (funcs=0x0) at /home/acm/source/php5-200401051230/ext/sqlite/sqlite.c:289
289             if (!funcs->is_valid) {
(gdb) bt
#0  0x4055e425 in php_sqlite_callback_invalidator (funcs=0x0) at /home/acm/source/php5-200401051230/ext/sqlite/sqlite.c:289
#1  0x40673df6 in zend_hash_apply (ht=0x81c9940, apply_func=0x4055e410 <php_sqlite_callback_invalidator>)
    at /home/acm/source/php5-200401051230/Zend/zend_hash.c:658
#2  0x4055e6be in php_sqlite_forget_persistent_id_numbers (rsrc=0x819f3c8) at /home/acm/source/php5-200401051230/ext/sqlite/sqlite.c:379
#3  0x40673df6 in zend_hash_apply (ht=0x406f5418, apply_func=0x4055e68c <php_sqlite_forget_persistent_id_numbers>)
    at /home/acm/source/php5-200401051230/Zend/zend_hash.c:658
#4  0x4055e74e in zm_deactivate_sqlite (type=1, module_number=4) at /home/acm/source/php5-200401051230/ext/sqlite/sqlite.c:395
#5  0x40670718 in module_registry_cleanup (module=0x818d9c8) at /home/acm/source/php5-200401051230/Zend/zend_API.c:1411
#6  0x40673df6 in zend_hash_apply (ht=0x406f5540, apply_func=0x406706d5 <module_registry_cleanup>) at /home/acm/source/php5-200401051230/Zend/zend_hash.c:658
#7  0x4066c2eb in zend_deactivate_modules () at /home/acm/source/php5-200401051230/Zend/zend.c:797
#8  0x4062a32b in php_request_shutdown (dummy=0x0) at /home/acm/source/php5-200401051230/main/main.c:1228
#9  0x406994ce in php_apache_request_dtor (r=0x81c4c78) at /home/acm/source/php5-200401051230/sapi/apache2handler/sapi_apache2.c:443
#10 0x406998b7 in php_handler (r=0x81c4c78) at /home/acm/source/php5-200401051230/sapi/apache2handler/sapi_apache2.c:547
#11 0x080686b2 in ap_invoke_handler ()
#12 0x08065a2b in ap_process_request ()
#13 0x08060c7b in _start ()
#14 0x08072f85 in ap_process_connection ()
#15 0x0806818b in ap_graceful_stop_signalled ()
#16 0x0806755e in ap_graceful_stop_signalled ()
#17 0x08066e74 in ap_mpm_run ()
#18 0x0806e0e2 in main ()
#19 0x402f494c in __libc_start_main () from /lib/libc.so.6


And as far as I can tell, I'm running the prefork model, although gdb does claim a thread gets forked off ?? (but from what I can tell from the configuration-line in gentoo's ebuild and the output of ps, it should be the prefork model)

See also bug #26852

This bug has been fixed in CVS.

Snapshots of the sources are packaged every three hours; this change
will be in the next snapshot. You can grab the snapshot at
http://snaps.php.net/.
 
Thank you for the report, and for helping us make PHP better.