Lars Magne Ingebrigtsen writes in ding@gnus.org: > Content-Type: text/plain; charset=iso-8859-1 This should be `iso-2022-jp' as the message contains iso-2022-jp encoded characters. Automagically figuring out the charset is one area that tm is very strong at, so there's probably some code there to be reused. The way I composed this message was to add `text/plain' tags around the line containing Japanese and tm autodetected what charsets to put in the MIME headers. > SL Baur writes: >> Subject lines don't appear to be getting the funky MIME quoting, and >> all I'm seeing the body are tildes. I'm running XEmacs/Mule in a >> Japanese locale.