Charles Lindsey writes:
I am worried about the receivers that don't understand UTF-8, of
which sendmail seems the chief offender.
Hmmm. I was using ``receiver'' to refer to the final message reader. If
all the message readers understand =?utf-8?= then you can just use that
without running into any trouble with sendmail.
Sorry, it seems I was wrong, though actually only the first on your list
seems to be legal UTF-8 according to RFC 2044.
It takes extra work for a UTF-8 reader to exclude the other encodings.
Does anyone have the relevant excerpt from the real UTF-8 definition?