Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site ccieng6.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!harpo!seismo!rochester!ritcv!ccieng5!ccieng6!dwr
From: dwr@ccieng6.UUCP ( Donald Wallace Rouse II)
Newsgroups: net.lang.c
Subject: Re: C programming style poll
Message-ID: <160@ccieng6.UUCP>
Date: Tue, 1-May-84 19:01:35 EDT
Article-I.D.: ccieng6.160
Posted: Tue May  1 19:01:35 1984
Date-Received: Wed, 2-May-84 06:31:20 EDT
References: <2735@azure.UUCP>
Organization: Computer Consoles Inc., Rochester, NY
Lines: 43

xxxxxxxxxxxxxxxxxxxxxxxx
>
>	(1)		y = (x == 5);
>
>
>	(2)		y = 0;
>			if (x == 5)
>				y = 1;
>
> By the way, I agree that the first form above may not be
> guaranteed to be entirely equivalent to the others, but if y is
> assigned to a Boolean expression, it should be tested with
>
>	if (y)
>	if (!y)
> or possibly
>	if (y == 0)
>	if (y != 0)
> but never with
>	if (y == 1)
>	if (y != 1)

According to the K&R bible, p. 41 at the top:

	Another useful form of type conversion is that
	relational expressions like i > j and logical
	expressions connected by && and || are defined to have
	value 1 if true, and 0 if false.  Thus the assignment

		isdigit = c >= '0' && c <= '9';

	sets isdigit to 1 if c is a digit and to 0 if not.

Therefore, it IS legal (though bad form, in my opinion)
to use the form "(y == 1)" or "(y != 1)".

I use this definition occasionally to access arrays, as in:

	char *	tf [] = {"false", "true"};
	...
	printf ("The supposition that x is 0 is %s\n", tf [x == 0]);

									D2