Systems: SGI Origin 2000 IP27 running IRIX64 6.5.3f
SGI Indy IP22 running IRIX 6.5.3f
DEC Alpha running Digital UNIX V4.0D (Rev. 878)
Ssh version: 2.0.13
I was able to solve some of the problems I complained about previously:
the double mail check and the double printing of /etc/motd (the latter
on SGI only) can be avoided setting PrintMotd and CheckMail to "no" in
/etc/ssh2/sshd2_config, while the funny setting of the DISPLAY
environmental variable is indeed required for secure connections.
The speed problems seem CPU-related as they are worse on old slow
computers as the Indy. Thus the remaining beauty errors are:
1) Though it seems to be the default (see e.g. the lines
config->fall_back_to_rsh = TRUE;
config->use_rsh = TRUE;
in apps/ssh/sshconfig.c) the fallback mechanism to rsh when the remote
host does not have ssh does not work for me; I simply get
"FATAL: Connecting to remotehost failed: TCP/IP Failure".
I remember it working with version 1.2.21.
2) It seems that scp and sftp source .tcshrc without sourcing /etc/cshrc
before. As the latter file sets the path, some commands contained in my
.tcshrc don't work (e.g. hostname). I think non-interactive commands
like scp and sftp shouldn't source .tcshrc, but if they do, they should
also source /etc/cshrc, which is generally assumed as a requisite.
3) On SGI
a) /etc/issue is not printed
b) info about the remote system is not printed
c) info about the last login is printed incomplete (only date, not from)
d) the environment variables REMOTEHOST, REMOTEUSER and LANG are not set
e) the environment variables MAIL and TERM are set wrong; in particular,
if the latter is "iris-ansi" on the local host, it should be changed
to "iris-ansi-net" on the remote host (see "man rlogin")
f) when connecting from a DEC, the characters "[> F" are printed before
the first prompt
4) On DEC
a) the MAIL and PATH environment variables are set wrong
(the MAIL variable shouldn't be set at all)
b) when connecting from a SGI, I must press an enter to get a prompt.
Any ideas on how to solve the above things?
Regards
Guido Germano
Theoretical Physics, University of Bristol, England
tel. +44-117-928 8755, fax +44-117-9255624
http://www.phy.bris.ac.uk/staff/germano_g.html
