From owner-freebsd-hackers  Fri Mar 17 09:59:13 1995
Return-Path: hackers-owner
Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id JAA01856 for hackers-outgoing; Fri, 17 Mar 1995 09:59:13 -0800
Received: from duality.gnu.ai.mit.edu (!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*!#$%^&*@duality.gnu.ai.mit.edu [18.43.0.236]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id JAA01815 for <hackers@FreeBSD.org>; Fri, 17 Mar 1995 09:59:12 -0800
Received: (from mycroft@localhost) by duality.gnu.ai.mit.edu (8.6.10/8.6.10) id MAA28775; Fri, 17 Mar 1995 12:58:49 -0500
Date: Fri, 17 Mar 1995 12:58:49 -0500
Message-Id: <199503171758.MAA28775@duality.gnu.ai.mit.edu>
From: "Charles M. Hannum" <mycroft@ai.mit.edu>
To: terry@cs.weber.edu
Cc: peter@bonkers.taronga.com, hackers@FreeBSD.org, tech-net@NetBSD.ORG
Subject: Re: Batch Telnet (Re: diskless and 3Com 509)
Sender: hackers-owner@FreeBSD.org
Precedence: bulk


   The message returned is "Connection closed by remote host.", not
   "Connection closed."

And you actually believed that message?  It's wrong.

    if (setjmp(peerdied) == 0)
        telnet(user);
    (void) NetClose(net);
    ExitString("Connection closed by foreign host.\n",1);

The code is structured rather poorly.  It ought to generate that
message where it actually notices that the socket is closed.

   The problem is clearly that the client is getting a close before it
   reads the data to be displayed.

Not at all.  If you actually look at the packets on the wire, you'll
see:

1) The client sends a FIN.

2) The server sends some data.

3) The client responds to the data with a RST.

tcpdump(8) leaves no doubt that the problem, if any, is in the client.

Actually, I misspoke about TCP.  The client can close the write side
by sending a FIN.  However, the telnet client is currently closing
both sides of the connection when it gets an EOF on stdin.  This is
really easy to fix.  In fact, I just did it.