Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!rutgers!ames!husc6!seismo!mcnc!duke!ndd
From: ndd@duke.UUCP
Newsgroups: comp.unix.wizards
Subject: Re: Two process communication programs
Message-ID: <9746@duke.cs.duke.edu>
Date: Sat, 6-Jun-87 09:22:31 EDT
Article-I.D.: duke.9746
Posted: Sat Jun  6 09:22:31 1987
Date-Received: Sun, 7-Jun-87 03:24:40 EDT
References: <2071@emory.UUCP>
Reply-To: ndd@duke.UUCP (Ned D. Danieley)
Organization: Duke University, Durham NC
Lines: 31

In article <2071@emory.UUCP> km@emory.UUCP (Ken Mandelberg) writes:
>Most of the communications programs on Berkeley Unix (tip, rlogin, cu)
>are split into two two processes on the local side, a reader and
>writer. This makes sense to me in a System V environment with no
>"select" system call. But why the extra process in tip and rlogin,
>which are relatively recent and written just for BSD?
>-- 
>Ken Mandelberg      |  {akgua,sb1,gatech}!emory!km   USENET
>Emory University    |  km@emory                      CSNET,BITNET
>Dept of Math and CS |  km.emory@csnet-relay          ARPANET 
>Atlanta, Ga 30322   |  Phone: (404) 727-7963

Tip is not relatively recent: it was certainly available around 1980
on the V7 system I was using at the time. Perhaps no one has rewritten
it since.

And, depending on what you are using, tip may not involve just two processes.
I remember installing a security feature for tip that I got from the
net. The problem was that tip made a lock file in /usr/spool/uucp,
and then forked once to do the communicating. Unfortunately, you had to leave
tip running as uucp so it could remove the lock, which meant that a
user who did a shell escape would have uucp privileges. The hack was
to fork two processes: one for the read and one for the write; they would
reset to the user's id, and the parent would wait until they terminated,
and then remove the lock. In retrospect, it seems as if this could have been
handled without the extra process, but I thought it was a neat trick at
the time.

Ned Danieley
Basic Arrhythmia Laboratory
Duke University Medical Center