Re: [Nmh-workers] Non-ASCII Characters in bodies and subjects

[Ken Hornstein]

>Is there an unenforced rule that all headers must be all ASCII, when a draft
>is stored in UTF-8 format? Or perhaps the defacto standards allow such
>headers?

Well, it depends what you mean by "enforced".  No one's going to serve you
with an arrest warrant if that happens.  But the RFCs are quite explicit:
in the general case, headers can only contain ASCII (I'm ignoring the
message/global MIME type; I have not yet seen those in the wild).  It
happens that a lot fo software will happily transmit such messages, but
the big problem is on the receiving end; without knowing what the
character set is, you can't properly display it.

Now there's a scheme for encoding non-ASCII characters in headers; that
should happen automatically by mhbuild, at the same time the draft has
the other MIME stuff added to it.  It should be pretty obvious when this
happens if you look at the raw file.  If that's not happening then I
REALLY screwed up.

--Ken