I upgraded an old P5/133 machine here from 1.4 to 1.4.1.  The 1.4.1
fsck_ffs coredumped on any and all of that machine's FFS partitions,
with a SIGSEGV. I verified the sum of fsck_ffs and the stock 1.4.1
kernel against another machine; both looked okay.

Grabbing and builing the -current fsck_ffs made the problem go away.
I gues it's due to the weird fdisk/bsd-label I have on this machine
(It used to run Linux, and its MBR/labels are still in some horrible
state after __endless__ problems with the 1.2 install script).

Has anyone else seen anything similar this?  Is it worth pulling up a
fix for 1.4.2, or has that already been done?