non-ASCII newsgroup names

[Katsumi Yamaoka <yamaoka@jpl.org>]

(I wish it were not a disturbance to syncing Gnus and Emacs.)

Hi,

I've improved the non-ASCII newsgroup names handling of Gnus.
This is kind of a major change, so I did a job only in the
Gnus CVS trunk, a.k.a., No Gnus.

(The changes I also installed to the v5-10 branch are not
necessarily related to non-ASCII group names.  They fix
miscellaneous bugs that I found while I was working.)

People who subscribe to or want to subscribe to newsgroups of
which the names contain non-ASCII characters had better use the
trunk for some time.  Note that the back ends other than nntp or
nnrss have not been completed yet for non-ASCII group names.

To read and post those newsgroups, there are three variables
that you have to set properly:

gnus-group-name-charset-method-alist
====================================
  If the newsgroup names are all encoded with the gb2312 charset
  in the nntp server news.cn.example.com, and the newsgroup names
  in news.jp.example.com are all encoded with the euc-jp charset,
  set it as:

  (setq gnus-group-name-charset-method-alist
	'(((nntp "news.cn.example.com") . gb2312)
	  ((nntp "news.jp.example.com") . euc-jp)))

  The car of each element is a select-method object.  If it is
  very long, like

  (nntp "quimby" (nntp-address "news.gnus.org") (nntp-end-of-line ...

  , you can truncate it into (nntp "quimby").

  For Gnus v5.11, it has to be exactly the same as the actual
  method.

gnus-group-name-charset-group-alist
===================================
  This variable specifies regexps matching group names and
  corresponding charsets.  For instance:

  (setq gnus-group-name-charset-group-alist
	'(("\\`gnus\\.alt\\.chinese\\." . gb2312)
	  ("\\`gnus\\.alt\\.japanese\\." . euc-jp)
	  (".*" . utf-8)))

  Where the last element is the default charset to decode group
  names.  Note that this variable is ignored if the match is made
  with `gnus-group-name-charset-method-alist'.

nnmail-pathname-coding-system
=============================
  *XEmacs users must set this*.  Emacs users necessarily need not
  do.  This overrides the value of `file-name-coding-system' which
  specifies the coding system used to encode and decode file names.
  Since the Agent, the nntp marks feature, and the nnrss back end
  use non-ASCII group names in file names, the coding system has
  to be set properly anyway.  The default value is nil.

  In XEmacs (w/ Mule), only `file-name-coding-system' is the means
  to specify the coding system used to encode and decode file names.
  On the other hand, Emacs uses `default-file-name-coding-system'
  if `file-name-coding-system' is nil.  Normally the value of
  `default-file-name-coding-system' is initialized according to
  the locale, so you will need to do nothing if the value is
  suitable to encode and decode non-ASCII group names.

There are two main points in this improvement.  One is to make
encoded non-ASCII group names all unibyte internally.  Since
multibyte ones and unibyte ones coexisted formerly, Gnus sometimes
failed in identifying a group.

The other is to make `file-name-coding-system' be bound to
`nnmail-pathname-coding-system' when accessing files of which
the names contain non-ASCII group names.  In addition to this,
I've modified `nnmail-group-pathname' so as not to encode file
names by `nnmail-pathname-coding-system'.  It produced the
encoding of file names doubly (the first is done by itself, the
second is done by `file-name-coding-system').

Regards,