arlad crashed (i386-netbsd1.3H)

[Ken Raeburn]

Using the binaries I built from the 12-28 snapshot (haven't applied
the patches yet), I got an abort in arlad:

(gdb) bt
#0  0x400d872b in kill ()
#1  0x400d84c5 in abort ()
#2  0x400afcce in __assert ()
#3  0x4031c in remove_this (h=0xdea00, n=0) at ../../util/heap.c:216
#4  0x403d6 in heap_remove_head (h=0xdea00) at ../../util/heap.c:235
#5  0x9faf in invalidator (arg=0x0) at ../../arlad/fcache.c:641
#6  0x3a2fd in Create_Process_Part2 () at ../../lwp/lwp.c:629
#7  0xfffefdfc in ?? ()
#8  0x1 in ?? ()
#9  0x3a09a in LWP_MwaitProcess (wcount=1, evlist=0xefbfce74)
    at ../../lwp/lwp.c:567
#10 0x3a140 in LWP_WaitProcess (event=0x4c60) at ../../lwp/lwp.c:585
#11 0x5345 in main (argc=0, argv=0xefbfd8c0) at ../../arlad/arla.c:910
(gdb) 

(gdb) fr 3
#3  0x4031c in remove_this (h=0xdea00, n=0) at ../../util/heap.c:216
216         assert (n < h->sz);
(gdb) p n
$1 = 0
(gdb) p h->sz
$2 = 0

(gdb) p *h
$5 = {cmp = 0x8ca8 <expiration_time_cmp>, max_sz = 4000, sz = 0, 
  data = 0xfc000}
(gdb) 

Anything else useful I can look up for you?  I'm firing it up again
with debugging logging turned on, maybe that'll turn up something more
informative if it crashes again.

Ken