Re: People with From: <mail@domain.com>

[Florian Weimer <fw@deneb.cygnus.argh.org>]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Florian Weimer <fw@deneb.cygnus.argh.org> writes:
> 
> > Perhaps I should add something to the manual which explains that?
> > Which section would be appropriate?
> 
> Yes, an explanation in the manual would be good.  I've done some
> cursory browsing of message.info and emacs-mime.info, and I couldn't
> find anything about the auto-choose mechanism.  Hm.  I think there
> should be a section somewhere which says that Gnus (or message, or
> mml) looks at the message to automatically find a way to encode it
> before sending.  Sometimes, the message is split into several parts,
> and the right charset is chosen, stuff like that.

I'm going add it to emacs-mime.texi, somewhere around the description
of MML.  After all, charset translation from MULE to MIME takes place
during MML-to-MIME translation.

BTW: I've browsed some source code and discovered a code/comment
discrepancy in "mm-encode-body", non-MULE case:

  (if (not (featurep 'mule))
      ;; In the non-Mule case, we search for non-ASCII chars and
      ;; return the value of `mm-default-charset' if any are found.
      (save-excursion
	(goto-char (point-min))
	(if (re-search-forward "[^\x0-\x7f]" nil t)
	    (or mail-parse-charset
		(mm-read-charset "Charset used in the article: "))
	  ;; The logic in `mml-generate-mime-1' confirms that it's OK
	  ;; to return nil here.
	  nil))

The comment says "mm-default-charset", but actually it's
"mail-parse-charset".  The value that's actually used by Gnus is
"message-default-charset", as it seems (in fact, this is a FAQ ;).
I'll change the comment accordingly.