Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83 (MC840302); site mcvax.UUCP
Path: utzoo!watmath!clyde!floyd!harpo!decvax!mcvax!guido
From: guido@mcvax.UUCP (Guido van Rossum)
Newsgroups: net.emacs
Subject: Re: emacs and 4.2 (why eot-process doesn't work)
Message-ID: <5692@mcvax.UUCP>
Date: Fri, 17-Feb-84 08:57:35 EST
Article-I.D.: mcvax.5692
Posted: Fri Feb 17 08:57:35 1984
Date-Received: Sat, 18-Feb-84 04:17:40 EST
References: <16542@sri-arpa.UUCP>
Organization: CWI, Amsterdam
Lines: 16

Eot-process doesn't work because the kernel doesn't implement it.  The
4.2 manual page pty(4) states under the description of TIOCREMOTE how
an eot should be sent: "a write of 0 bytes is like typing an end-of-file
character".  However, the BUGS section tells us: "It is not possible to
send an EOT".  (This was not present in the 4.1c version of the manual.)
Inspection of the kernel code shows that some attempts are made to
handle writes of 0 bytes special, but somehow this doesn't make it to
the other end of the line.
  Someone at our site hacked the 4.1c version to use the master tty in
normal mode rather than TIOCREMOTE mode, thus with input editing
enabled, and to send a control-D character.  This worked, but had the
disadvantage that processes in windows could influence the tty status
with stty system calls.  Later we reverted to an "official" version.
--
Guido van Rossum, {philabs,decvax}!mcvax!guido
Centre for Mathematics and Computer Science, (CWI, formerly MC), Amsterdam