readme.mime

Emacs 插件代码,增加unicode支持和中文支持

                                        -*- text -*-

  This file roughly describe the mime coding problems about GBK in GNU
Emacs21, which covers two famous mail agents of Emacs:
    
 - Mew: A great MIME mail reader for Emacs/XEmacs (http://www.mew.org/)
 - Gnus: Gnus Network User Services (http://www.gnus.org/)


1. Multipurpose Internet Mail Extensions(MIME)
   
  The MIME standard is covered mainly by RFC 1521 and RFC 1522, etc.
To special the textual encoding with GBK, the mail should use a line
in the header which looks like
   
   Content-Type: Text/Plain; charset=gbk

  Here, `gbk' is the name of the `mime charset'.
  According to RFC 1251, the charset name w/o the publication of
a formal specification and its registration with Internet Assigned
Numbers Authority (IANA), or by private agreement, must begin with
an additional "X-". Therefore, we can find that there are large amount
Internet mails with header like

   Content-Type: Text/Plain; charset=x-gbk

  But, now a days GBK is a registered charset of IANA:
---------------8<---------------
Name: GBK                                                 
MIBenum: 113
Source: Chinese IT Standardization Technical Committee  
        Please see: <http://www.iana.org/assignments/charset-reg/GBK>
Alias: CP936
Alias: MS936
Alias: windows-936 
---------------8<---------------

  See

    http://www.iana.org/assignments/character-sets

for the informations of charset list, and

    http://www.iana.org/assignments/charset-reg/GBK

for the registration text compiled by Anthony Fok <[e-mail]>.

  Note that, there are 3 aliases of GBK in the publication of IANA, it
may be better to count them in, although it seems that no one uses
them.


2. Mew setup

  To use GBK encoding in Mew, added the following Elisp codes into
your init file of Emacs, and make sure that these will be evaluated
before Mew loads.

    ;; For GBK
    (eval-after-load "mew"
      '(progn
         (setq mew-cs-database-for-decoding
               (append
                '(("gbk"         . gbk)
                  ("x-gbk"       . gbk)
                  ("cp936"       . gbk)
                  ("ms936"       . gbk)
                  ("windows-936" . gbk))
                mew-cs-database-for-decoding))
         (add-to-list 'mew-cs-database-for-encoding
                      '((ascii chinese-cns11643-5
                               chinese-cns11643-6
                               chinese-cns11643-7)
                        gbk
                        "base64"
                        "B"))))

  This setup has been simply tested on Mew 3.3 w/ GNU Emacs 21.3.

3. Gnus setup

  Be patient to read up this section!!!

  In Gnus(running in GNU Emacs21), the mime charset to encode the
text is determined by a property `mime-charset' of the current coding
system of the text buffer, which is not specified in coding system
`Chinese-gbk' of the old versions of mule-gbk, so that Gnus encoded
the GBK text with wrong coding system. It has been fixed since
mule-gbk 0.1.2004062701.

  And the encoding system to decode the text of the mail is determine
by the `charset=...' information in the mail header.

  For GBK support, we need to evaluate the following Elisp codes
before Gnus loads up.

    ;; For GBK
    (eval-after-load "mm-util"
      '(setq mm-charset-synonym-alist
             (append
              '((x-gbk       . gbk)
                (cp936       . gbk)
                (ms936       . gbk)
                (windows-936 . gbk))
              mm-charset-synonym-alist)))

  Because the file `mm-util.el' in Gnus, which defines
`mm-charset-synonym-of-gbk-alist' is also a part of GNU Emacs, we
added the above setup for Gnus into `mule-gbk.el'. Thus, the only
task left to you is to make sure that mule-gbk is loaded before the
library `mm-util'(or Gnus) :)

  This setup has been simply tested on Gnus 5.10.6 w/ GNU Emacs 21.3.

4. Bug report

  Bug reports or suggestions (mailto:[e-mail]) are greatly
welcome!

--

SU Yong <[e-mail]>
Sun, 27 Jun 2004 04:33:44 +0800