Xref: utzoo comp.sys.3b1:506 comp.sys.att:11903
Path: utzoo!mnetor!tmsoft!torsqnt!hybrid!scifi!bywater!uunet!lll-winken!elroy.jpl.nasa.gov!ames!vsi1!zorch!hico2!westmark!argon!ebh
From: ebh@argon.UUCP (Ed Horch)
Newsgroups: comp.sys.3b1,comp.sys.att
Subject: Re: ".GIF" files
Message-ID: <844@argon.UUCP>
Date: 25 Feb 91 16:37:03 GMT
References: <1991Feb6.215717.26551@cbnewsj.att.com> <1991Feb7.204039.5842@sonyd1.Broadcast.Sony.COM> <993@gnosys.svle.ma.us> <843@argon.UUCP> <996@gnosys.svle.ma.us>
Reply-To: ebh@argon.UUCP (Ed Horch)
Organization: Lightspace Designs, Highland Park, NJ
Lines: 19

In article <996@gnosys.svle.ma.us> gst@gnosys.svle.ma.us (Gary S. Trujillo) writes:
>In <843@argon.UUCP> ebh@argon.UUCP (Ed Horch) writes:
>> There may be a performance cost due to creating an additional
>> process for "cat a", but a smart shell will know how to short-
>> cut around that.
>
>What's this about a "smart shell"?  I've never heard of a shell that
>optimizes commands.  It seems inherantly very difficult to recognize
>that the two commands listed above are equivalent, since it requires
>a knowledge of what cat(1) does.  What shell are we talking about here?

The Korn Shell has subsumed several common commands as shell builtins.
I stand corrected on one point - cat isn't one of them, but echo and
test are, among others.  Yes, it does have knowledge about what those 
commands do.  The shell decides whether to perform any task either 
inline, in a subshell, or in a subprocess, depending upon what would 
be most optimal.

-Ed