Path: utzoo!utgpu!watmath!att!dptg!rutgers!gatech!ncar!ico!ism780c!haddock!karl
From: karl@haddock.ima.isc.com (Karl Heuer)
Newsgroups: comp.lang.c
Subject: Re: swap(x,y)
Message-ID: <14479@haddock.ima.isc.com>
Date: 31 Aug 89 00:18:06 GMT
References: <MYwDEFy00WB842JEUx@andrew.cmu.edu> <8350@boring.cwi.nl> <190@titania.warwick.ac.uk> <230@crdos1.crd.ge.COM>
Reply-To: karl@haddock.ima.isc.com (Karl Heuer)
Organization: Interactive Systems, Cambridge, MA 02138-5302
Lines: 21

In article <230@crdos1.crd.ge.COM> davidsen@crdos1.UUCP (bill davidsen) writes:
>However, there are reasons for doing tricks such as the XOR thing.  They
>enable me to write a macro which does the swap for any type.

Only for integral types.  It won't work on float, pointer, struct, etc.

If you insist on using a single macro, I think that
	#define swap(x,y,TYPE) {TYPE _swap_tmp_=(x); (x)=(y); (y)=_swap_tmp;}
is a better bet than the XOR hack.

I think if I were proposing a *language change* to solve this problem, I
wouldn't ask for a builtin swap operator; it's too limited.  Might as well
invent something that has other applications as well, such as the displacing
assignment operator (where  x := y  stores value y in lvalue x, but the value
of the entire expression is the previous value of x rather than the new one)
or even my reverse sequential evaluation operator (where  x,,y  evaluates
first x and then y, but (unlike the comma operator) returns y as its value;
cf. `prog1' in lisp).  Then swap would be  x=y:=x  or  x=(y,,y=x)

Karl W. Z. Heuer (ima!haddock!karl or karl@haddock.isc.com), The Walking Lint
(Followups on this last paragraph should go to comp.lang.misc, please.)