Re: XEmacs, Gnus and mm-coding-system priorities.

[Katsumi Yamaoka <yamaoka@jpl.org>]

>>>>> In <b9yy8ga7wc7.fsf@jpl.org> Katsumi Yamaoka wrote:

> Anyway, thank you for the new patch.  I wish we could merge it
> by the time Gnus v5.10.7 is released.

I'm ready to install your patch in the Gnus trunk and the v5-10
branch.  My patch is available at:

http://www.jpl.org/mm-patch-20041207.txt

I hope it is tested by many XEmacs users in the Latin worlds.
There I made slightly modifications:

1. Added some dummy definitions to suppress compile warnings.
2. Provided a small macro used to reduce a funcall overhead in
   Emacs.
3. Corrected the order of arguments specified to the
   latin-unity-massage-name function.
4. Removed the chars-region variable which is unused.

If there's no comment, I will commit it when you drop a signed
paper into a mailbox.

(For the change to v5-10 branch, I think it can be considered to
be a bug fix since Gnus didn't have an ability to unify Latin-9
into Latin-1 in XEmacs.)

P.S.
I found another problem, although it is not related to this
subject.  The following form causes an error in XEmacs, but
Emacs works.

(rfc2047-encode-string (string
			(make-char 'latin-iso8859-1 95)
			(make-char 'japanese-jisx0208 38 66)))

The occasion using mixture text of Latin and Japanese is few,
though.