Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site proper.UUCP
Path: utzoo!watmath!clyde!akgua!sdcsvax!sdcrdcf!hplabs!hpda!fortune!amd70!dual!proper!gam
From: gam@proper.UUCP (Gordon Moffett)
Newsgroups: net.unix
Subject: Why csh instead of sh? (includes free flame!)
Message-ID: <1206@proper.UUCP>
Date: Wed, 16-May-84 01:28:48 EDT
Article-I.D.: proper.1206
Posted: Wed May 16 01:28:48 1984
Date-Received: Sun, 13-May-84 08:22:37 EDT
References: <1192@proper.UUCP>, <1525@brl-vgr.ARPA>
Organization: Proper UNIX, Oakland, CA
Lines: 30
Keywords: C-shell,Bourne shell,scripts,translators,test(1),expr(1)

#
In response to my request for a sh -> csh translator, I have
already received two letters asking `WHY would you want to do that?'
and now I find that question posted to the net, so perhaps I can
forstall a deluge.

But first, a slight flame:  I guess you are all trying to be helpful
warning me against the evils of csh, but THAT WASN'T MY QUESTION!
I found this as annoying as asking about a good truffle shop and
being asked `Why would you eat chocolate?  It has caffiene in it!
...' and so on.  Annoying.  Questioning people's preferences is
pointless.

Now, why csh?  Because the particular script I want to translate is
heavily into using the test(1) and expr(1) commands, which are already
built-in to csh.  Yes, there will be overhead with a csh script start-up,
reading ~/.cshrc and all, but this will be offset by the numerous
calls to test and expr that I won't be using.

(As Kernighan & Pike point out, judicious use of sh's `case' construct
 will avoid having to call test(1) in some cases -- for pattern
 matching/string comparison -- but I've already done that as much
 as possible).

I generally prefer the speed of sh scripts and resort to csh for
instances like I have described above (though my login shell is
csh, nonetheless).

I believe we've already had a round of `csh vs sh' before, and I don't
wish to provoke another one.