improving nnrss.el

improving nnrss.el

[Katsumi Yamaoka]

Hi,

I and Tsuyoshi AKIHO are now planning to make nnrss.el support
multilingual text.  There was such a demand and a workaround
using the charset group parameter was suggested by the author.

http://article.gmane.org/gmane.emacs.gnus.general/53944

It works only for subjects and text/plain parts by slight
modification of nnrss.el (inserting raw octets into multibyte
buffer breaks them).  Furthermore, to modify mm-view.el so that
it may recognize the charset group parameter will make it work
also for text/html parts.  The troublesome of this way is to set
the group parameter manually.  We are scheming for automating
it, keeping the backward compatibility.  The plan includes
enabling it to use non-ascii group names (i.e., utf-8 encoded).

Regards,

Re: improving nnrss.el

[Katsumi Yamaoka]

[-- Attachment #1: Type: text/plain, Size: 683 bytes --]

Hi,

>>>>> In <b9y4qivqcks.fsf@jpl.org> Katsumi Yamaoka wrote:

> I and Tsuyoshi AKIHO are now planning to make nnrss.el support
> multilingual text.

[...]

> The plan includes enabling it to use non-ascii group names (i.e.,
> utf-8 encoded).

If you are interested in it, you can try the latest Gnus CVS and
the following patch:

ftp://ftp.jpl.org/pub/tmp/nnrss-patch-20041214.gz
or http://www.jpl.org/ftp/pub/tmp/nnrss-patch-20041214.gz

It is still under development and we need to verify it further.
Non-ASCII group names aren't supported yet with the `j' group
command etc.  I made the following function for the replacement
of `completing-read'.  However, it is rather slow.


[-- Attachment #2: Type: application/emacs-lisp, Size: 659 bytes --]

[-- Attachment #3: Type: text/plain, Size: 10 bytes --]


Regards,

Re: improving nnrss.el

[Katsumi Yamaoka]

[-- Attachment #1: Type: text/plain, Size: 3267 bytes --]

>>>>>> In <b9y4qivqcks.fsf@jpl.org> Katsumi Yamaoka wrote:

>> I and Tsuyoshi AKIHO are now planning to make nnrss.el support
>> multilingual text.

> ftp://ftp.jpl.org/pub/tmp/nnrss-patch-20041214.gz
> or http://www.jpl.org/ftp/pub/tmp/nnrss-patch-20041214.gz

The new patch is here:

ftp://ftp.jpl.org/pub/tmp/nnrss-patch-20041221.gz
or http://www.jpl.org/ftp/pub/tmp/nnrss-patch-20041221.gz

To use this patch, you need to use the latest Gnus CVS.

I and Tsuyoshi AKIHO will install it into the Gnus trunk in the
near feature after he finished the paper work to FSF.  The new
nnrss.el supports multilingual text and non-ASCII group names.
You who are already subscribing to rss groups will need to do
some tedious tasks before using the new nnrss.el.  I'm sorry for
that but it will give you a real advantage.

Please note, you won't get a good result if your Emacs doesn't
support the `utf-8' coding system.

If you are an Emacs user and don't mind the past articles may
break a little, you have nothing to do except for removing of
the `charset' group parameter.

1. Remove the `charset' group parameter for the rss groups if
   you've set them.  It was a possible workaround in order to
   decode non-ASCII articles, but it is no longer necessary.

2. You may be able to continue to read past articles using the
   new nnrss.el, although articles which have broken from the
   first cannot be helped (it likely occurred because of the old
   nnrss.el).  To attempt to do that, read each rss file as
   binary, decode the contents using the proper coding system
   (which was used as the `charset' group parameter, maybe), and
   save it using the coding system which is specified to the
   `nnrss-file-coding-system' variable (see below).  I made a
   simple command for doing that.  It is attached at the end of
   this message.

   OTOH, removing all existing rss groups, removing all files
   under the ~/News/rss/ directory, restarting Gnus and creating
   again those groups is a good idea. ;-)

3. If you are an XEmacs user and want to use non-ASCII group
   names, you have to specify the value for the
   `nnmail-pathname-coding-system' variable.  It is used to
   encode and decode local rss filenames which will contain
   non-ASCII group names.  The value should be the coding system
   which is used as your computer's default.  For example:

   (setq nnmail-pathname-coding-system 'iso-8859-1)

   Although Emacs users don't have to bother to specify it since
   Emacs should set the `default-file-name-coding-system'
   variable properly, you can override it in the same way.

4. If you are using both Emacs and XEmacs-MULE, you have to
   specify the value for the `nnrss-file-coding-system' variable
   properly.  It is used to encode text contents which are saved
   in local rss files.  The default value is `emacs-mule' in
   Emacs or is `escape-quoted' in XEmacs-MULE, those are
   incompatible.  The possible values include `iso-2022-7bit',
   `utf-8', `ctext', etc.  For example:

   (setq nnrss-file-coding-system 'iso-2022-7bit)


OT: nnshimbun.el contained in the emacs-w3m package is also
useful for reading news sites in the world.  There are modules
for BBC News, Heise, Laut, N24, Rediff, Spiegel, The Register,
Welt, Zeit, etc.

\f

[-- Attachment #2: Type: application/emacs-lisp, Size: 1603 bytes --]

Re: improving nnrss.el

[Katsumi Yamaoka]

Hi,

Tsuyoshi AKIHO has sent a signed paper to FSF and I've installed
the new nnrss.el in the Gnus trunk.

>>>>> In <b9y1xdiziat.fsf@jpl.org>
>>>>>	Katsumi Yamaoka <yamaoka@jpl.org> wrote:

> The new nnrss.el supports multilingual text and non-ASCII group
> names.  You who are already subscribing to rss groups will need to
> do some tedious tasks before using the new nnrss.el.  I'm sorry for
> that but it will give you a real advantage.

Regards,