Re: Wide_[Wide_]Character

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Sat, 12 Jul 2008 11:00:05 GMT, Dale Stanbrough wrote:

> Dmitry A. Kazakov wrote:
> 
>> RM 3.5.2 talks about "code positions" (=code points, I guess), represented
>> by Wide_Wide_Character. From this I conclude that it shall be UCS-4 with
>> some implementation-defined endianness.
> 
> Code points can be represented by any set of encodings. Wide_Character 
> seems to deliberately confine itself to the BMP, so UCS-2 would suffice 
> (and seems implied).
> 
> I can't see any implication that would cause me to think 
> Wide_Wide_Character is definitely UCS-4 (and not UTF-16).

How about this: Wide_Wide_Character may obviously use only the encodings
which would make  any Wide_Wide_String composed out of Wide_Wide_Characters
a properly encoded string in the same encoding. This automatically excludes
UTF-8 and UTF-16.

BTW, why do you care? (:-)) I wonder if there is any use of
Wide_[Wide_]Strings. IMO, anything one could wish from Unicode is provided
by UTF-8 and plain Strings...

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de