Please review 'text/dns' and 'application/dns'

[Simon Josefsson]

ned.freed at mrochek.com writes:

>> I missed the cut-off date, but meanwhile it is available from:
>
>> http://josefsson.org/draft-josefsson-mime-dns.txt
>
>> The forms where part of earlier versions of draft-josefsson-dns-url,
>> and there has been two versions of draft-josefsson-mime-dns published
>> at IETF before, so the types may have been discussed here before; I
>> don't recall.  Still, this version has been improved in many ways, so
>> I'd appreciate more review.
>
> Looks basically OK to me. The one thing I noted that is missing is some
> discussion of the charset used for text/dns; I assume it is US-ASCII.

Ah, thanks, there should be some discussion about that.  I think the
document should acknowledge existing practice that comments can be
encoded in some locally defined charset.  (This is another
interoperability issue, too.)

As for non-comments, non-ASCII are probably in use, and to my reading
they appear conforming to the specification.  The non-ASCII octet
values are not interpreted in any specific charset though, they are
just treated as opaque binary octets.  RFC 1035 (section 4) do allow
for escape sequences like \DDD to express arbitrary octets, so the
master files does not _have_ to be anything but ASCII, even if they
represent arbitrary octets.  I'm not sure how to phrase this in MIME
terms though.  Thoughts?  It appears as if any valid MIME charset
could be permitted, since the master file format only worries about
that actual on-disk encoding, not the character interpretation.

Thanks,
Simon