[imp] Does Imp 3.0 generate wrong header fields?

Otto Stolz Otto.Stolz@uni-konstanz.de
Tue, 29 Jan 2002 22:44:48 +0100


Hello,

quoted from the source of a message recently forwarded through
the Imp list:
 > From: =?Any?B?SmVz+nMtQW50b25pbyBBcnJveW8=?= <...>
 > MIME-version: 1.0
 > Content-type: text/plain; charset=Any
 > Content-transfer-encoding: 8BIT
 > User-Agent: Internet Messaging Program (IMP) 3.0

That particular encoding of a name in the From field is
governed by RFC 2047, section 3, cf.
<http://sunsite.dk/RFC/rfc/rfc2047.html>,
and the Content-Type field is governed by
RFC 2046, section 4.1.2, cf.
<http://sunsite.dk/RFC/rfc/rfc2046.html>,
both of which refer to the IANA registry for the
valid values of the charset parameter.

In the IANA registry,
<http://www.iana.org/assignments/character-sets>,
there is no character set named "Any". So, the
labels quoted above are clearly wrong. How is the
receiver supposed to make heads or tails of this mess?

The last line quoted above apparently indicates that these
erroneous labels were generated by IMP 3.0. Is that correct?

If so, there is a real problem. For interoperability with
other mailing software, it is absolutely essential that
IMP 3.0 produces labels with the correct MIME charset
parameter.

Without a correct MIME charset parameter (or when the receiving
side does not interpret this paramter correctly -- another
problem I have seen and reported with IMP 2.7), no characters
(beyond 7bit-ASCII) can sensibly be transmitted. In other words,
without this parameter, you can only transmit messages in Swahili,
Hawaiian, and English (disregarding particular punctuation marks
and loan words).

Best wishes,
   Otto Stolz