Path: utzoo!attcan!uunet!know!cs.utexas.edu!rutgers!att!dptg!ulysses!andante!alice!ark
From: ark@alice.att.com (Andrew Koenig)
Newsgroups: comp.lang.c
Subject: Re: Comparing strings... (long but maybe worth it)
Message-ID: <11488@alice.att.com>
Date: 14 Oct 90 14:26:17 GMT
References: <2205.271700c2@cc.nu.oz.au> <11485@alice.att.com> <1990Oct14.095336.2819@zoo.toronto.edu>
Organization: AT&T Bell Laboratories, Liberty Corner NJ
Lines: 24

In article <1990Oct14.095336.2819@zoo.toronto.edu>, henry@zoo.toronto.edu (Henry Spencer) writes:

> Some (for example, me) would say that this was aided by your reliance on
> implicit comparisons to zero, which hurt readability and do nothing for
> your stated goal of greater efficiency.  Why did you bother doing that?

Because for me it increases readability.

> >I imagine that further optimizations are possible...

> Definitely. ...
> [optimization description deleted]

Neat!  And you haven't even unrolled any loops!

> For extra credit, discuss how current implementations by
> both Berkeley and AT&T attempt this optimization and thereby introduce
> a bug (hint:  consider the fact that char is signed on most machines).

I'm afraid I may have given it away in another posting, which I made
before I saw this one.
-- 
				--Andrew Koenig
				  ark@europa.att.com