Red Hat Bugzilla – Bug 16823
ssh vs rsh/rexec and keyboard mapping of <del> and <bs>
Last modified: 2008-05-01 11:37:58 EDT
Aug 23 14:41:26 russianblue xinetd: Bad line received from identity
server at 18.104.22.168: 1016
Aug 23 14:41:26 russianblue pam_rhosts_auth: allowed to
email@example.com as root
Aug 23 14:41:26 russianblue PAM_unix: (system-auth) session opened
for user root by (uid=0)
Aug 23 14:41:26 russianblue in.rshd: firstname.lastname@example.org as root:
cmd='exec /bin/csh -cf "setenv PATH
xauth remove mudpuddle.cs.wustl.edu:0.1 ; xauth merge -"'
Aug 23 14:41:26 russianblue automount: attempting to mount entry
the common shell script "xrsh"
# Some System V systems don't understand the #! construct.
# If your system does understand it, put ": " at the beginning of the
XRSH_VER='xrsh version 5.8-amc3' # AMC
XRSH_RCS='$Header: /usr/u/jjd/xrsh/RCS/xrsh.sh,v 1.12 1994/06/13 15:53:11
jjd Exp $'
XRSH_TIME='Time-stamp: <94/06/13 11:43:59 jjd>'
# Copyright 1991 by James J. Dempsey <email@example.com>
when used to access a RH 6.1 system correctly maps <del> and <bs> keys to
delete. The key would normally send ^[[3c or some such key codes.
under RH 6.2 and pinstripe, these codes are ignored and "~" and <bell> are
I can "xterm -ls -n foo" once in and the keyboard mapping is still bad.
xmodmap and the other x utilities see a correct string for the keypress,
the system still ignores them. a ^H sent or a ^? sent is correct. it is
the multi-char escape sequence that gets munged.
using rsh/rexec/ssh2 does not show the problem.
Changing the $PATH does not fix it either.
using xmodmap to remap the keys does not fix it.
quite annoying (whatever breaks it is in xfree86*.rpm, since
I can apply theses rpm's to a 6.1 system and break it there).
not a really big problem, but as a system-admin with 100's of machines to
at, it gets a bit tiring to remember ^H not <bs> :-)
I cannot replicate this. Have you tried a newer release at all? Is the bug
still present for you in red Hat Linux 6.2/7.0/7.1?
I know the pain you're having all too well from a long time ago though. Not
with the same software, but with other broken setup... That was in the "other"
OS though. ;o) It seems different people have had different but similar
experiences to yours, so I'd like to see if it is still present and fix it,
but I need a case that I can replicate here to get started first.
This happens on 6.1, 6.2, 7.0, and 7.1. It happens comming from solaris, tru64,
and from secure-crt
off a WinDoze PC. It started with 3.3.6 of XFree86.
This does not happen to a version of Xfree86 4.0.x when compiled on Netbsd.
I have not tried a recompile and install on Linux.. I'll have to do that when I
want some more abuse :-)
I can supply the functional version of xrsh we use on our slowlaris boxes, via
or anon-ftp let me know.
xrsh source sent to firstname.lastname@example.org
I cannot reproduce this bug.