From:	XXguille AT XXiies DOT XXes (Guillermo Rodriguez Garcia)
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Portability and size_t type related question
Date:	Thu, 13 May 1999 22:28:09 GMT
Organization:	Telefonica Transmision de Datos
Lines:	43
Message-ID:	<373d50d7.7978822@noticias.iies.es>
References:	<Pine DOT SUN DOT 3 DOT 91 DOT 990513161505 DOT 12629T-100000 AT is>
NNTP-Posting-Host:	iies249.iies.es
Mime-Version:	1.0
X-Newsreader:	Forte Agent 1.5/32.451
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

El día Thu, 13 May 1999 16:18:56 +0300 (IDT), Eli Zaretskii
<eliz AT is DOT elta DOT co DOT il> escribió:

>> >> Like 'char' is always 1 byte.
>> >
>> >That's not true, either.  There are compilers (mostly for embedded 
>> >systems) where `char' is 32-bit wide.
>> 
>> Then a byte is defined as 32 bits on those systems.
>
>AFAIK, there's no such thing as a `byte' in the C language description.  

Yes, there is. Just read what I quoted here from the ANSI C standard
rationale.

>So defining a byte as 32 bits doesn't help for the issue at hand which 
>had to do with portability of C programs.

According to the ANSI standard, a char is always 1 byte wide, whatever
the size of one byte is.

>Most people think that byte is a synonym for 8 bits.

Then, they're wrong :-)

>> The
>> exact definition of byte, for every system, is the minimum addressable
>> memory unit.
>
>I'm not against this definition, but I'm not sure it's true.  A compiler 
>for embedded system could disallow 8-bit bytes because that would produce 
>inefficient code, not because individual bytes aren't addressable.

Then that compiler is not conforming to the ANSI standard.

You know, I'm not making this up myself :-) It is all covered in the
standard and in the rationale...

Regards,
GUILLE
----
Guillermo Rodriguez Garcia
XXguille AT XXiies DOT XXes (ya sabes :-)

- Raw text -