Re: Proposal about adding cp1251 (Russian Windoz) encoding in Emacs

[Kenichi Handa]

In article <7CF3E427AE54FC42A26E68C5A89C186910C763@zeal.ufg.com>, "Krjukov 
Victor" <VKryukov@ufg.com> writes:
> Greetings, Emacs developers.
> MULE support in Emacs contains several encodings for Russian language: 
> ISO 8859-5, used as main encoding (it's called a 'standard' encoding but
> _nobody_ here actually use it);
> koi8-r (de-facto standard in email transmission and standard in *nix
> systems - almost all *nix systems in Russia use it);
> and Alternativnyj (also known as cp866) - an ancient encoding come from
> DOS world.

> But Emacs has no support for cp1251 (standard Russian encoding in
> Windows), though we have and activelly use Emacs for Windows.

Thank you for the report.  We have already supported cp1251
in the latest CVS code.  Could you please try it?  When you
set the language environment to "Windows-1251", you can
start using the coding system windows-1251 (alias cp1251).

But, I don't know if Emacs automatically switch to that
lang. env. when started on Windows.  Does anyone know that?

> I've made several changes in several *el files from Emacs 21.2
> distribution, but it seems to me that you need to recompile Emacs to
> include additional language in MULE support (and I have no C compiler on
> my Windows system).

> Could someone check attached files and probably implement them in future
> Emacs version? This would be a great help to russian windows-emacs
> users.

It seems that Emacs already has all the features you want,
so I haven't read through your code.  We greatly appreciate
if you try the latest Emacs code and report problems if any.

Dave, you have improved cyrillic.el greatly.  Are the
following problem already fixed?  Could you please comment
on that?

> PS. I've also made some changes in lisp/language/cyrillic.el: fix some
> typo (staff instead of stuff ;) and fix some bugs (e.g.
> cyrillic-encode-koi8-r-char instead of
> cyrillic-encode-alternativnyj-char at the bottom of folloving excerpt in
> lines 203-217 of cyrillic.el looks really strange).

> I have also another question: comparing sections in cyrillic.el for
> koi8-r and for alternativnyj you may notice that one line       
> (if (r0 == ,(charset-id 'cyrillic-iso8859-5))
> only present for define-ccl-program ccl-encode-koi8 and not in
> ccl-encode-alternativnyj - is it a bug or a feature?
> ---
> (make-coding-system
>  'cyrillic-alternativnyj 4 ?A
>  "ALTERNATIVNYJ 8-bit encoding for Cyrillic"
>  '(ccl-decode-alternativnyj . ccl-encode-alternativnyj)
>  `((safe-chars . ,(let ((table (make-char-table 'safe-chars))
>                       (i 0))
>                   (while (< i 256)
>                     (aset table (aref
> cyrillic-alternativnyj-decode-table i)
>                           t)
>                     (setq i (1+ i)))
>                   table))
>    (valid-codes (0 . 175) (224 . 241) 255)
>    (charset-origin-alist (cyrillic-iso8859-5 "ALTERNATIVNYJ"
        
> cyrillic-encode-koi8-r-char))))
> ---

---
Ken'ichi HANDA
handa@etl.go.jp