Re: Guidelines for instruction set design?

On 07 May 2009 16:35:36 +0100 (BST), Tony Finch <dot@dotat.at> wrote:


>gmt@cs.arizona.edu (Gregg Townsend) wrote:
>>
>> [The C language doesn't require] individually addressable characters
>
>You mean bytes. Characters don't fit in bytes.


C has always required individually addressable characters, but
characters are not required to be 8-bit. Prior to ANSI C (C89) bytes
were not even defined as a language item and it was coincidence that
there existed platforms on which characters could be considered
synonymous with bytes.


C89 defines "byte" in terms of character: a byte is an addressable
unit of data storage large enough to hold any member of the character
set of the environment.


C99 refines the C89 definition to specify that the "character set" be
the "_basic_ character set of the _execution_ environment".


C99 specifies multi-byte characters and makes a distinction between a
"basic character" whose bit pattern must fit into a single byte, and a
"wide character" which may require multiple bytes to express.


George
[Historical note: C was first implemented on the GE 635, a 36 bit word
addressed machine with a very complex instruction set that could also
handle 9 bit bytes. -John]

Return to the comp.compilers page.
Search the comp.compilers archives again.