From ssh-owner@clinet.fi Tue Jul 18 19:23:30 1995 Return-Path: ssh-owner@clinet.fi Received: from clinet.fi (clinet.fi [193.64.6.1]) by ecstasy.olari.clinet.fi (8.6.11/8.6.9) with ESMTP id TAA14068 for ; Tue, 18 Jul 1995 19:23:28 +0300 Received: from hutcs.cs.hut.fi (root@hutcs.cs.hut.fi [130.233.192.2]) by clinet.fi (8.6.10/8.6.4) with SMTP id SAA06781 for ; Tue, 18 Jul 1995 18:40:12 +0300 Received: from shadows.cs.hut.fi by hutcs.cs.hut.fi with SMTP id AA19388 (5.65c8/HUTCS-S 1.4 for ); Tue, 18 Jul 1995 18:40:09 +0300 Received: (ylo@localhost) by shadows.cs.hut.fi (8.6.10/8.6.10) id SAA12787; Tue, 18 Jul 1995 18:40:12 +0300 Date: Tue, 18 Jul 1995 18:40:12 +0300 Message-Id: <199507181540.SAA12787@shadows.cs.hut.fi> From: Tatu Ylonen To: sdw@lig.net Cc: ssh@clinet.fi In-Reply-To: Subject: Re: Ssh installation comments Organization: Helsinki University of Technology, Finland > I finally have ssh working between my hpux 9.05 system and my Linux 1.2.8+ > server. I have compiled it on both (the hpux was 9.03) without problems or need to edit anything. More important than your Linux kernel version is the distribution that you are using. I am using the same kernel, Slackware 2.1.0 (gcc-2.7.0). I had no problems. I have been using it on linux for about four months, and haven't had any big proglems. Please let me know what version you have and what you had to do to get it to work. The HPUX code is not extensively tested; I don't use a HP machine regularly myself. > sshd doesn't seem to have a capability to record utmp information to > allow normal who, etc. information. On most machines, it should properly record utmp and wtmp information. However, almost every machine has the files in a different location, in a different format, complemented by {u,w}tmpx, has different values in ut_id (if it exists), has the files in a different directory, has lastlog.h or not, has _UTMP_FILE etc defined in the headers or not, etc. The number of combinations appears to be almost countless. Currently it tries to determine what you have, where and in which format. Sometimes it fails. If it cannot determine how to do something, it doesn't do it all. I don't now remember what was the situation with HPUX. (The relevant code is in login.c) I have almost come to the conclution that the best way would be to have a separate version of the file for each operating system, or at least for some systems. > Another problem I've noticed is that interrupt characters, while set with > properly according to stty, aren't being passed over the ssh connection > as rlogin/telnet normally would. (As Out Of Band data normally.) Yeah, they are passed as a normal data, and arrive in sequence with other characters. If this is a real problem, I can try to do something about it. > In otherwords, ^c, ^\, etc. have no effect that the characters disappear. This should not happen, and I have never seen this happen. However, the HPSUX pty code is very different from everything else I have seen - from example, EOF is not passed from the master side to the client side. There were substantial hacks to make it work at all on HPSUX (see the SIGCHLD stuff). Please let me know if you find out how to fix this. I'll try to check this when I come back (I am currently in IETF in Stockholm). Tatu