Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!ut-sally!pyramid!decwrl!sun!guy
From: guy@sun.uucp (Guy Harris)
Newsgroups: net.unix,net.unix-wizards
Subject: Re: question-- Bourne (and C) SHELL
Message-ID: <7365@sun.uucp>
Date: Wed, 17-Sep-86 15:07:38 EDT
Article-I.D.: sun.7365
Posted: Wed Sep 17 15:07:38 1986
Date-Received: Fri, 19-Sep-86 23:26:25 EDT
References: <1751@ittatc.ATC.ITT.UUCP> <7028@utzoo.UUCP> <150@humming.UUCP> <6228@sun.uucp> <574@ih1ap.UUCP>
Organization: Sun Microsystems, Inc.
Lines: 23
Xref: mnetor net.unix:5532 net.unix-wizards:7914

> Correct me if I'm missing something.  What's wrong with:
> 
> 	ls foo
> 	if [ $? != 0 ]; then
> 		echo foo does not exist
> 	fi

It doesn't indicate quite as clearly as the other version that the
failure/success of "ls" is being tested.  It merely indicates that the exit
status of "ls" is being compared against 0; you have to associate this with
failure/success.  (See past discussions in net.lang.c about why Boolean
types exist even if you language happens to implement them as integral
types.)

In addition, 1) "ls" does not always return an exit status indicating
whether it could find the files in question or not and 2) "ls" *does* always
print the name of the file, if it finds it, or an error, if it doesn't.  If
"test" had a "does this file exist" predicate, then as I mentioned that
would be what you should use.
-- 
	Guy Harris
	{ihnp4, decvax, seismo, decwrl, ...}!sun!guy
	guy@sun.com (or guy@sun.arpa)