Experiment with UTF-8 in message-IDs

Julien ÉLIE <[email protected]> Sun, 09 Oct 2011 21:27:16 +0200
Newsgroups gmane.ietf.usenet.format
Organization TrigoFACILE -- http://www.trigofacile.com/
Message-ID <[email protected]>
Hi all,

In the IETF working group for IMA (Internationalized eMail Address),
there is a current thread about UTF-8 in message-IDs:
    http://www.ietf.org/mail-archive/web/ima/current/threads.html#04330

Quick references in the thread:

http://www.ietf.org/mail-archive/web/ima/current/msg04430.html
http://www.ietf.org/mail-archive/web/ima/current/msg04344.html
http://www.ietf.org/mail-archive/web/ima/current/msg04345.html
http://www.ietf.org/mail-archive/web/ima/current/msg04420.html
http://www.ietf.org/mail-archive/web/ima/current/msg04422.html



RFC 5536 (USEFOR) currently allows only ASCII characters in message-IDs.

INN 2.4 and INN 2.5 have always rejected message-IDs containing
non-ASCII chars.  (I have not looked at INN 2.3 and before.)  When
a message-ID is not valid per RFC 850/1036/... and now 5536, the
article is rejected.

200 news.trigofacile.com InterNetNews server INN 2.6.0 (20110908 prerelea=
se) ready (transit mode)
IHAVE <=C2=A9@fr>
435 Syntax error in message-ID
MODE READER
200 news.trigofacile.com InterNetNews NNRP server INN 2.6.0 (20111003 pre=
release) ready (posting ok)
ARTICLE <=C2=A9@test>
501 Syntax error in message-ID
QUIT
205 Bye!


(Note that 435 is answered to IHAVE for legacy reasons; 501 should be
the real response code per RFC 3977.)




My question is:  should we try right now to relax the check so as to allo=
w
UTF-8 in message-IDs?
If yes, is there something else to enforce?  (NFC normalization?)

Of course, other requirements from RFC 5536 will remain (that is to say
no comments in the Message-ID: header field, and no ">" or WSP).
U+00A0 (&nbsp; in HTML) and other spaces encoded in UTF-8 are allowed,
aren't they?



We plan on releasing INN 2.5.3 soon, so perhaps we can relax the check
starting from INN 2.5.3.  I will ask in the INN workers mailing-list,
if naturally there is no complaints in this USEFOR mailing-list against
going this way.

--=20
Julien =C9LIE

=AB I don't know if it's what you want, but it's what you get. =BB
  (Larry Wall)