NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/lang/c-part1/5247

Internet Message Format | 1996-08-05 | 2.6 KB

Path: news.iag.net!news
From: [e-mail] (John R Buchan)
Newsgroups: comp.lang.c
Subject: Re: simple code, argc, argv, strcmp()
Date: 9 Feb 1996 05:49:48 GMT
Organization: Internet Access Group, Orlando, Florida
Message-ID: <4fen9s$[e-mail]>
References: <11f7cc$17261a.3b3@daprez> <4etj7c$[e-mail]> <[e-mail]> <4f0rjn$[e-mail]> <[e-mail]>
NNTP-Posting-Host: pm2-orl23.iag.net
X-Newsreader: WinVN 0.99.7

In article <[e-mail]>, [e-mail]
says...
>
>In article <4f0rjn$[e-mail]> [e-mail] "John R Buchan" writes:
>
>>You are correct, of course.  Stupidity on my part. I tend to get careless
>>about strcmp's reverse logic (ie returns False, when the strings match),
when
>>I am not paying attention (not an excuse, just an explanation).  I normally
>>compare the return value to 0, just to keep myself straight.
>
>
>That's the point though - strcmp() simply doesn't return a boolean value so
>saying 'ie returns False' is just plain wrong and is what gets you into
>trouble. strcmp() is in some sense more like a difference operation. e.g.
>(ignoring overflow) if you take 2 integers a and b and calculate a-b the
>result is >0 if a is larger, ==0 if they are equal and <0 if a is smaller,
>just as strcmp(sa,sb) is defined. IMHO the definition of strcmp is quite
>natural (but maybe the name doesn't quite reflect that definition).

<sigh> Yes, I am well aware that strcmp doesn't return a boolean value. When
it is actually used to determine the relative value of strings (ie in a sort
function), it makes perfect sense and its use is quite straightforward.

However, when you use it as the condition of an if (or any similar
conditional) the only thing that matters is whether the return is zero or
non-zero (False or True).  I guess it's a weakness of mine, but I tend to
read conditions (or any essentially boolean expressions) in terms of True and
False (an unfortunate side effect of this is that I can make the mistake
mentioned above).  When the function's name doesn't make its 'boolean' value
obvious, I find the code much easier to read, if I compare the return to
some 'obvious' value (eg, 0, NULL, etc).

As you indicated the name 'strcmp' doesn't really read well in a boolean.
Maybe it should changed to something like:

 if( theseStringsAreDifferent(str1, str2) ) /* Is this a bit wordy? :-} */

....but that name doesn't make sense in a relative value statement. Hmm..:)

--
John R Buchan           -:|:-     Looking for that elusive FAQ?  ftp to:
[e-mail]     -:|:-     rtfm.mit.edu /pub/usenet-by-group/....