I've started getting the SEGV again when running RDoc. The backtrace
looks like:


#0  0x400abbc1 in __kill () from /lib/libc.so.6
#1  0x400ab9c5 in raise () from /lib/libc.so.6
#2  0x400acfe1 in abort () from /lib/libc.so.6
#3  0x80bae98 in rb_bug () at error.c:178
#4  0x80a0a73 in sigsegv () at signal.c:393
#5  <signal handler called>
#6  0x80a3306 in st_foreach (table=0x7e, func=0x8068860 <mark_entry>, arg=0x0)
    at st.c:485
#7  0x806889d in rb_mark_tbl (tbl=0x7e) at gc.c:453
#8  0x8068d2d in rb_gc_mark (ptr=136551568) at gc.c:678
#9  0x8068d91 in rb_gc_mark (ptr=1076300836) at gc.c:707
#10 0x806880c in mark_locations_array (x=0xbffebe64, n=19510) at gc.c:417
#11 0x8068858 in rb_gc_mark_locations (start=0xbfffef38, end=0xbffebe64) at gc.c:436
#12 0x8069464 in rb_gc () at gc.c:1052
#13 0x80685c7 in rb_newobj () at gc.c:321
#14 0x80a3424 in rb_str_new (
    ptr=0x82997fd "considered to be a\n#   paragraph.\n#\n# * If a paragraph starts with a \"*\", \"-\", or with \"<digit>.\", then it is\n#   taken to be the start of a list. The margin in increased to be the\n#   first non-space"..., len=1) at string.c:39
#15 0x80a4120 in rb_str_substr (str=1076031216, beg=1213, len=1) at string.c:360
#16 0x80a5393 in rb_str_aref_m (argc=2, argv=0xbffec0fc, str=1076031216)
    at string.c:955
#17 0x8059f08 in call_cfunc (func=0x80a5340 <rb_str_aref_m>, recv=1076031216, 
    len=-1, argc=2, argv=0xbffec0fc) at eval.c:4276


This is on the latest CVS 1.6.7.  The same code seems to work find
with 1.7.  Anyone any ideas?


Cheers


Dave