Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!necntc!ima!think!barmar
From: barmar@think.COM (Barry Margolin)
Newsgroups: comp.lang.c
Subject: Re: C Style
Message-ID: <20126@think.UUCP>
Date: 22 Apr 88 22:29:04 GMT
References: <11216@brl-adm.ARPA> <2111@chinet.UUCP> <4403@garfield.UUCP> <226@hotlr.ATT> <130@obie.UUCP> <5981@utcsri.UUCP> <1982@ubc-cs.UUCP> <126@atpal.UUCP> <2823@mmintl.UUCP> <255@oink.UUCP> <3583@haddock.ISC.COM>
Sender: usenet@think.UUCP
Reply-To: barmar@fafnir.think.com.UUCP (Barry Margolin)
Organization: Thinking Machines Corporation, Cambridge, MA
Lines: 27

In article <3583@haddock.ISC.COM> karl@haddock.ima.isc.com (Karl Heuer) writes:
>>There are times when a var can very intentionally have a non-zero (true)
>>value other than TRUE (1).  The classic kind of case of this is [isalpha].
>
>Then isalpha() is not a boolean, in my book.  I'm not convinced there's any
>good reason for this; it would be trivial to rewrite the function to return a
>true boolean.  In most instances this would have zero run-time cost.

Well, if you're coding in C, you'll have to live with C's quirks.
Nowhere does the C language specify that any particular non-zero value
should be returned by the isXXX functions to indicate truth.  The
person who wrote isalpha() might have done #define TRUE 2, and you
might have done #define TRUE 1.  Who is to say that you are right and
he is wrong?  The only thing that C specifies is that you must both do
#define FALSE 0.  Remember, C doesn't have real booleans, and #define
TRUE 1 isn't going to change that fact.

I can think of one reason, however, why 1 should be used as a standard
truth value.  Single-bit fields in structures are generally used as
flags, and it would be nice to be able to say flag = isXXX(...) rather
than flag = (isXXX(...) != FALSE).  But this is just a wish; portable
code must currently use the more verbose version.

Barry Margolin
Thinking Machines Corp.

barmar@think.com
uunet!think!barmar