Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!snorkelwacker!spdcc!ima!haddock!karl
From: karl@haddock.ima.isc.com (Karl Heuer)
Newsgroups: comp.lang.c
Subject: Re: character strings, macros, ANSI c, multicharacter chars question
Keywords: string, multicharacter, ebcdic/ascii
Message-ID: <15583@haddock.ima.isc.com>
Date: 6 Jan 90 19:07:56 GMT
References: <1990Jan4.190431.19748@ultra.com>
Reply-To: karl@haddock.ima.isc.com (Karl Heuer)
Distribution: na
Organization: Interactive Systems, Cambridge, MA 02138-5302
Lines: 22

In article <1990Jan4.190431.19748@ultra.com> ted@ultra.com (Ted Schroeder) writes:
>I would like to create an integer that is full of '+' characters
>i.e. on an ASCII machine the four byte integer would be 0x2b2b2b2b.
>Unfortunately the code must run on both ASCII and EBCDIC machines.

First I'll mention the siamese character constant '++++', but only to
disrecommend it.  It'll probably give you the right answer, but you can't
count on the compiler accepting it without a warning.

Barry suggested ('+'<<24 | '+'<<16 | '+'<<8 | '+'); the briefer form
(0x01010101 * '+') is equivalent.  You can eliminate the byte- and word-size
assumptions with the following (note that the result is cast-free, and hence
usable at the preprocessor level, only if the compiler is ANSI):
	#if __STDC__
	#include <limits.h>
	#else
	#define UINT_MAX  (~(unsigned int)0)
	#define UCHAR_MAX ((unsigned char)UINT_MAX)
	#endif
	#define PLUS      (UINT_MAX / UCHAR_MAX * '+')

Karl W. Z. Heuer (karl@haddock.isc.com or ima!haddock!karl), The Walking Lint