Subject: Help! My BSDi MHSnet binaries don't work under 1.0
To: None <firstname.lastname@example.org>
From: Simon J Gerraty <email@example.com>
Date: 11/09/1994 00:07:43
if/when my system (zen) gets to talk to the world again, there are a
few messages queued. The 30 second summary is:
I've installed 1.0 (well the Oct24 binaries) on zen and its great.
Except for one (2 actually) problem.
I have a binary only comms package (MHSnet) which was compiled for a 0.9
vintage BSDi system. This is the transport for all my mail/news and
works fine on 0.9a (Feb12). Under 1.0 I get a semi useful error:
system error -- Could not lock "/var/spool/MHSnet/_state/lock": invalid argument
The file gets created ok, so I'm guessing that it is an flock() or fcntl()
problem. The kernel I built yesterday didn't have KTRACE enabled and
there is no symbol info in the binary so I'm a bit short of further info
just now. A new kernel with KTRACE is compiling now... (src from
Ok. I figured I'd revert back to uucp which was also working fine on the
old system. Running "cu -x chat" shows everything going smoothly until
the CONNECT string is matched and then we get a "caught hangup" message.
I'm dialed into my feed site now using kermit so the modems and ports work
fine. I have the "local" flag set in /etc/ttys for both modem ports and
I've checked that I can dial in/out ok.
I'm sure that with time I can sus the uucp problem, but suggestions would
The only other hassle is that nnmaster appears to being doing some set*uid()
stuff that no longer works.
Apart from that (and the Oct24 generic kernel panicing when lpd tried to
print to an off-line (I think) printer on lpt0) all is well.