On Friday 21 December 2018 22:37:42 CET, Bruno Wolff III wrote:
I also got the same problem when I attached an ebcdic file as an attachment, so there is probably more to this than I thought.
The linefeed isn't UTF-8, it's email. A MIME part delimiter starts with the four bytes 13, 10, 45, 45 (all decimal). The part that's subject to the "charset=" specification ends just before those four bytes, and the part that's subject to the next "charset=" starts a little later.
Many clients assume that the CR LF are the end of the part that's ending, but syntactically it's the first bytes AFTER the end of the part.
It's less clear than one might wish in the RFCs, but if you combine the ABNF in RFCs 822 and 2046 it's clear enough. Or look at the example on https://tools.ietf.org/html/rfc2046#page-21 and note how it says "DOES end" near the bottom of the page: if you look closely there "e n d CR LF CR LF - -" etc. The first CR LF is part of the bodypart and is what the comment describes, the second is part of the delimiter.
Mutt is right, just not very bug-compatible. All praise mutt. Arnt
