Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!lll-crg!caip!princeton!allegra!alice!ark
From: ark@alice.UucP (Andrew Koenig)
Newsgroups: net.lang.c
Subject: Re: swap() macro
Message-ID: <5688@alice.uUCp>
Date: Sun, 22-Jun-86 10:27:11 EDT
Article-I.D.: alice.5688
Posted: Sun Jun 22 10:27:11 1986
Date-Received: Tue, 24-Jun-86 03:30:57 EDT
References: <370@anasazi.UUCP>
Organization: Bell Labs, Murray Hill
Lines: 44

> #define swap(x, y) \
>   if (1) \
>   { \
>     register int *xp, *yp, t; \
>     xp = &(x); \
>     yp = &(y); \
>     t = *xp; \
>     *xp = *yp; \
>     *yp = t; \
>   } \
>   else
>
> --------------------------------
>
> The "if (1) ... else" allows the user to put a semicolon after the macro
> invocation, and makes code such as
> 
>   if (today == tuesday)
>     swap(*a++, *b++);
>   else
>     swap(*c++, *d++);
> 
> work correctly.

On the other hand, if you forget a semicolon:

	if (x < y)
		swap (x, y)
	printf ("%d %d\n", x, y);

you will spend a long time wondering why it doesn't print anything.

Of course, if you write in C++, you can say:

	inline void swap (int& x, int& y)
	{
		register int t = x;
		x = y;
		y = t;
	}

and be done with it.  It's a true function, the compiler checks
types for you, you can define several swap functions for different
types, and so on.