Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!cbosgd!ihnp4!homxb!whuts!mtune!rutgers!ukma!uunet!mcvax!ukc!its63b!aiva!richard
From: richard@aiva.UUCP
Newsgroups: comp.arch,comp.unix.wizards
Subject: pseudo terminals
Message-ID: <186@aiva.ed.ac.uk>
Date: Fri, 30-Oct-87 08:14:03 EST
Article-I.D.: aiva.186
Posted: Fri Oct 30 08:14:03 1987
Date-Received: Tue, 3-Nov-87 01:53:35 EST
References: <573@elxsi.UUCP> <682@its63b.ed.ac.uk> <796@quacky.UUCP> <421@nuchat.UUCP> <3619@sol.ARPA> <705@its63b.ed.ac.uk>
Reply-To: richard@uk.ac.ed.aiva (Richard Tobin)
Followup-To: comp.unix.wizards
Organization: AI Applications Institute, Edinburgh University
Lines: 37
Keywords: output pagination
Xref: utgpu comp.arch:2596 comp.unix.wizards:4871

[In a discussion of output pagination in comp.arch...]
In article <705@its63b.ed.ac.uk> simon@its63b.ed.ac.uk (Simon Brown) writes:
>Unfortunately, it isn't (so far as I know) possible to detect tty-mode changes
>over a pty, which means that the paging could *not* be automatically disabled
>if the command puts the terminal (ie, the slave end of the pty) into raw or
>cbreak mode, which wouldn't be the best at all.

Yes, this is something I've found awkward.  I wanted to write a program that
would provide command line editing in the style of tcsh (or indeed Simon
Brown's ssh) for all programs.  This would be done using a pseudo-terminal
and a filter program.  However, the inability to detect cooked-raw-cbreak
mode changes means this won't work if you want to be able to use screen
editors and the like.

In fact, it is possible to detect mode changes sometimes - a pseudo terminal
in packet mode can detect when the slave end flushes input, which usually
happens when changing modes (NB this is all BSD-specific).

What would be really useful would be a device that merely passed everything
from its slave end to its master end, including ioctls.  Ioctls calls at the
slave end would (optionally? - certainly in the case of GET ioctls) suspend
until the master end had processed it.  Has any unix system had such a device?
As well as the paging/editing applications, it could be used for prototyping
other device drivers; much of the debugging could be done with the driver
in user code, rather than the kernel.

I did, in fact, write such a device driver and used it on a sun to run
suntools in a window and across the network.  Response was, of course, quite
abysmal.

As this has little to do with computer architecture, I'm redirecting
followups to comp.unix.wizards.

-- 
Richard Tobin,                         JANET: R.Tobin@uk.ac.ed             
AI Applications Institute,             ARPA:  R.Tobin%uk.ac.ed@nss.cs.ucl.ac.uk
Edinburgh University.                  UUCP:  ...!ukc!ed.ac.uk!R.Tobin