> I know at least 4 other sites who want to get their mail from my mailserver
> via UUCP....
> They will kill me, if that will not be fixed soon.

A workaround, which I'm using myself, is to configure dial-up PPP, and
use UUCP over TCP.  You lose throughput, but at least it works!  :-)

UUCP over serial lines hasn't worked under NetBSD since Charles Hannum
put "the new com driver" for the Intel port into the tree.  (Not the
recent speedup fix, but back when the current driver replaced the
previous one.)  Mind you, I'm not saying that Charles broke anything,
just that it stopped working at pretty much exactly that time -- and not
just on the Intel port.

Incidentally, this was about the same time when another problem was
introduced that's still in the code.  Ever since that time, I'm plagued
with crashes on my older hardware, both Intel and VAX.  On my old i386
boxes, I get "page fault in supervisor mode" panics, and on VAX, I get
"SEGV in kernel mode" panics, which I suspect are the same thing.  The
frequency of the problems has changed over time (I track current pretty
closely), so while the i386 has been more or less unuseable at times in
the past, it's actually stable enough to use (with care) these days,
while my VAX is practically unbootable with 1.3_ALPHA.

Other operating systems have proved reliable on the same hardware: I've
tested Linux on the Intel boxes, and Ultrix and 4.3BSD-Reno on the VAX,
with no problems whatsoever.  I assume it's a VM bug that hits on older
(slower) hardware, and that it won't be fixed at least until the VM
subsystem is replaced, which I know is planned.

On modern hardware, I have no problems at all -- we run 1.2G in
production at work, and are very happy with it.  Looking forward to 1.3!
At home, though, I'll be moving to OpenBSD on Intel, and 4.3BSD-Reno on
the VAX.  Reno may be old, and may have its share of bugs, but at least
it's runnable on my old MicroVAX!  :-)

-tih