Re: Exact types and non-8-bit-bytes Position Statement

["Clive D.W. Feather" <yyyyy@xxxxxxxxx>]

Eric Vought said:
>> All the problems in this area are to do with byte != octet. Restricting
>> systems to byte == cluster-of-octets won't make a significant difference.
> 
> It does make a difference. As discussed before on this group, if you have
> a system with 9 bit bytes, and an int has 4 bytes, then there are several
> different ways you can look at producing a 32-bit quantity in network
> order from it. If you have 16 bit bytes and a 32 bit int, it is much
> simpler to produce a 32-bit quantity in network order.

Oh no it isn't.

The problem is that your 32-bit quantity has got to be presented to the
network in 4 bytes, because otherwise the send() or write() won't do the
right thing. See the example in my other email for more details.

-- 
Clive D.W. Feather  | Work:  <yyyyy@xxxxxxxxx>   | Tel:  +44 20 8371 1138
Internet Expert     | Home:  <yyyyy@xxxxxxxxxx>  | Fax:  +44 20 8371 1037
Demon Internet      | WWW: http://www.davros.org | DFax: +44 20 8371 4037
Thus plc            |                            | Mobile: +44 7973 377646