ietf-nntp NNTP and 16-bit charsets

[Russ Allbery]

Andrew Gierth <andrew at erlenstar.demon.co.uk> writes:

> 16-bit charsets in MIME still have to conform to the requirements of the
> content-transfer-encoding used, which means that if the charset does
> strange things with NUL or CR/LF octets, it can't be sent in 7bit or
> 8bit but only in binary, QP or base64.

Right.  Good point.

> but the "binary" content-transfer-encoding isn't supported by any
> existing implementations, and as far as I can see there is no way to fix
> that, because in the CRLF.CRLF terminating sequence, the first CRLF is
> part of the article body.

Confirmed.  *sigh*

Okay, that means I was definitely right and we need to go the same
direction that RFC 3030 went for SMTP.  I've added this to the things to
look at doing after we get this RFC out.  Basically, NNTP will need to
introduce new commands for transferring binary articles, and while we're
at it we can just specify the data length up front.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>