Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!crdgw1!uunet!taumet!mike
From: mike@taumet.com (Michael S. Ball)
Newsgroups: comp.lang.c++
Subject: Re: 64 bit architectures and C/C++
Message-ID: <712@taumet.com>
Date: 7 May 91 14:43:59 GMT
References: <168@shasta.Stanford.EDU> <12563@dog.ee.lbl.gov> <312@nazgul.UUCP> <184@shasta.Stanford.EDU>
Reply-To: mike@taumet.UUCP (Michael S. Ball)
Organization: Taumetric Corporation, San Diego
Lines: 25

In article <184@shasta.Stanford.EDU> shap@shasta.Stanford.EDU (shap) writes:
>In article <312@nazgul.UUCP> bright@nazgul.UUCP (Walter Bright) writes:
>>
>>The most aggravating problem we have is it seems we (Zortech) are the only
>>compiler for which:
>>	char
>>	signed char
>>	unsigned char
>>are all distinct types! For example,

>Have I misunderstood something in the standard?

You have.  The standard states that char is a distinct type which has
the same characteristics as either signed char or unsigned char.
This appears to be a fairly late change, and earlier versions of the
standard read as you assume.

It's a bit strange, but makes very little difference in C.  Carrying
the same definition over to C++ can make things very strange indeed.
The only solution given this definition is to make all characters
which are really characters "char", and use signed char and unsigned
char only for short arithmetic values.
-- 
Michael S. Ball			mike@taumet.com
TauMetric Corporation		(619)697-7607