Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!sdd.hp.com!mips!sjsca4!konath
From: konath@sj.ate.slb.com (Kannan Konath)
Newsgroups: comp.unix.sysv386
Subject: Re: sh and csh scripts (SCO ODT 1.0 with SCO Unix 3.2.1)
Keywords: sh csh scripts
Message-ID: <1991Feb1.015045.227@sj.ate.slb.com>
Date: 1 Feb 91 01:50:45 GMT
References: <1991Jan23.225748.23327@sj.ate.slb.com> <8836@star.cs.vu.nl> <1991Jan25.195557.25955@kithrup.COM> <1991Jan29.230217.14825@informix.com>
Reply-To: konath@sj.ate.slb.com (Kannan Konath)
Organization: Schlumberger ATE, San Jose, CA
Lines: 29

In article <1991Jan29.230217.14825@informix.com> aland@informix.com (Colonel Panic) writes:
>Now, the bonus question: is there an analogous directive for setting 
>up a C shell script that will work in non-BSD systems ('#!/bin/csh' 
>doesn't cut it in SVR3)?

Actually, I had asked this question originally and it got lost in
all the responses.

If you use csh you can execute either sh or csh scripts.

If the first character of a shell script starts with a # then
then the script will be treated as a csh script.  If the first
character is anything but a # (I find it's usually a colon ":")
then the script will be treated as a bourne shell script.

You should however use the csh when doing such things.  There
is a short paragraph in the man pages of both sh and csh (SCO ODT 1.0)
that mention this behaviour.  It's an extremely clumsy way of
System V R3 doing things.  

But this is not as irritating as trying to extract tar files in which
file or directory names happen to be longer than 14 characters.  I
spend half my time counting the number of characters in filenames.
I wish they had kept the length of filenames down to 8 characters
since it's much easier to detect this at a  glance.  Fourteen
character gives you a headache and the ability to count to 14. 

kannan
konath@sj.ate.slb.com