Weirdness in CS 1.1 with sshd and tcsh


Subject: Weirdness in CS 1.1 with sshd and tcsh
schinder@pobox.com
Date: Fri Aug 06 1999 - 12:42:51 MDT


I finally made the jump to a glibc 2.1 distro by installing Champion
Server 1.1. on my Performa 6400 on Tuesday, upgrading from a
well-before-R4-and-heavily-upgraded LinuxPPC.

For a variety of reasons, I prefer tcsh to bash, so /bin/tcsh is my
login shell. I installed ssh-1.2.27 almost immediately from source.
Everything was working well at the console. The first indication of
trouble was yesterday when I first logged into my Mac from work using
ssh (but I've since learned the same trouble arises using telnet or
rsh). tcsh was acting very strangely. First, it was immediately
obvious that it didn't process my .tcshrc. Second, I had no shell
variables, and couldn't set any. Third, I had no job control. If I
started anything and then ^Z'd, it just hung. Both tcsh and the
stopped process were still around, the stopped proces in T state, but
I never got the prompt back. I could kill the stopped process from
another window, and then get the prompt back. (But I had to kill -9;
a simple kill seemed equivalent to "kill -CONT", and the stopped
process reappeared.) It was almost as if I was getting in some kind
of "secure shell" mode in tcsh.

I've checked the files in /etc/pam.d, tried mounting /dev/pts with
mode=0666, tried rebuilding tcsh from source with #define VFORK and
with --with-login, I've run install-ssh and used /usr/sbin/sshd
instead of my build, but nothing has seemed to help. I've found a
simple workaround, an immediate "exec /bin/bash", since bash doesn't
seem to have the same problems. But I'd really rather get back to
using my regular shell. Has anyone seen this before and know what's
going on and how to fix it? I'm going to keep digging, but I thought
I'd ask here first.

-- 
Paul Schinder
schinder@pobox.com



This archive was generated by hypermail 2a24 : Sun Sep 05 1999 - 13:46:32 MDT