code.delx.au - gnu-emacs/blob - lisp/language/cyrillic.el

   1 ;;; cyrillic.el --- support for Cyrillic -*- coding: iso-2022-7bit; -*-
   2
   3 ;; Copyright (C) 1997, 1998, 2001, 2002, 2003, 2004, 2005, 2006, 2007
   4 ;;   Free Software Foundation, Inc.
   5 ;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
   6 ;;   2005, 2006, 2007
   7 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
   8 ;;   Registration Number H14PRO021
   9 ;; Copyright (C) 2003
  10 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
  11 ;;   Registration Number H13PRO009
  12
  13 ;; Author: Kenichi Handa <handa@etl.go.jp>
  14 ;; Keywords: multilingual, Cyrillic, i18n
  15
  16 ;; This file is part of GNU Emacs.
  17
  18 ;; GNU Emacs is free software; you can redistribute it and/or modify
  19 ;; it under the terms of the GNU General Public License as published by
  20 ;; the Free Software Foundation; either version 3, or (at your option)
  21 ;; any later version.
  22
  23 ;; GNU Emacs is distributed in the hope that it will be useful,
  24 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  25 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  26 ;; GNU General Public License for more details.
  27
  28 ;; You should have received a copy of the GNU General Public License
  29 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  30 ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  31 ;; Boston, MA 02110-1301, USA.
  32
  33 ;;; Commentary:
  34
  35 ;; The character set ISO8859-5 is supported.  KOI-8 and ALTERNATIVNYJ
  36 ;; are converted to Unicode internally.  See
  37 ;; <URL:http://www.ecma.ch/ecma1/STAND/ECMA-113.HTM>.  For more info
  38 ;; on Cyrillic charsets, see
  39 ;; <URL:http://czyborra.com/charsets/cyrillic.html>.  The KOI and
  40 ;; Alternativnyj coding systems should live in code-pages.el, but
  41 ;; they've always been preloaded and the coding system autoload
  42 ;; mechanism didn't get accepted, so they have to stay here and
  43 ;; duplicate code-pages stuff.
  44
  45 ;; Note that 8859-5 maps directly onto the Unicode Cyrillic block,
  46 ;; apart from codepoints 160 (NBSP, c.f. U+0400), 173 (soft hyphen,
  47 ;; c.f. U+04OD) and 253 (section sign, c.f U+045D).  The KOI-8 and
  48 ;; Alternativnyj coding systems encode both 8859-5 and Unicode.
  49 ;; ucs-tables.el provides unification for cyrillic-iso-8bit.
  50
  51 ;; Customizing `utf-fragment-on-decoding' allows decoding characters
  52 ;; from KOI and Alternativnyj into 8859-5 where that's possible.
  53 ;; cyrillic-iso8859-5 characters take half as much space in the buffer
  54 ;; as the mule-unicode-0100-24ff equivalents, though that's probably
  55 ;; not normally a big deal.
  56
  57 ;;; Code:
  58
  59 ;; Cyrillic (general)
  60
  61 ;; ISO-8859-5 stuff
  62
  63 (define-coding-system 'cyrillic-iso-8bit
  64   "ISO 2022 based 8-bit encoding for Cyrillic script (MIME:ISO-8859-5)."
  65   :coding-type 'charset
  66   :mnemonic ?5
  67   :charset-list '(iso-8859-5)
  68   :mime-charset 'iso-8859-5)
  69
  70 (define-coding-system-alias 'iso-8859-5 'cyrillic-iso-8bit)
  71
  72 (set-language-info-alist
  73  "Cyrillic-ISO" '((charset iso-8859-5)
  74                   (coding-system cyrillic-iso-8bit)
  75                   (coding-priority cyrillic-iso-8bit)
  76                   (input-method . "cyrillic-yawerty") ; fixme
  77                   (nonascii-translation . iso-8859-5)
  78                   (unibyte-display . cyrillic-iso-8bit)
  79                   (features cyril-util)
  80                   (sample-text . "Russian (\e,L@caaZXY\e(B)       \e,L7T`PRabRcYbU\e(B!")
  81                   (documentation . "Support for Cyrillic ISO-8859-5."))
  82  '("Cyrillic"))
  83
  84 ;; KOI-8R stuff
  85
  86 (define-coding-system 'cyrillic-koi8
  87   "KOI8 8-bit encoding for Cyrillic (MIME: KOI8-R)."
  88   :coding-type 'charset
  89   ;; We used to use ?K.  It is true that ?K is more strictly correct,
  90   ;; but it is also used for Korean.  So people who use koi8 for
  91   ;; languages other than Russian will have to forgive us.
  92   :mnemonic ?R
  93   :charset-list '(koi8)
  94   :mime-charset 'koi8-r)
  95
  96 (define-coding-system-alias 'koi8-r 'cyrillic-koi8)
  97 (define-coding-system-alias 'koi8 'cyrillic-koi8)
  98 (define-coding-system-alias 'cp878 'cyrillic-koi8)
  99
 100 (set-language-info-alist
 101  "Cyrillic-KOI8" `((charset koi8)
 102                    (coding-system cyrillic-koi8)
 103                    (coding-priority cyrillic-koi8 cyrillic-iso-8bit)
 104                    (ctext-non-standard-encodings "koi8-r")
 105                    (nonascii-translation . koi8)
 106                    (input-method . "russian-typewriter")
 107                    (features cyril-util)
 108                    (unibyte-display . cyrillic-koi8)
 109                    (sample-text . "Russian (\e,L@caaZXY\e(B)      \e,L7T`PRabRcYbU\e(B!")
 110                    (documentation . "Support for Cyrillic KOI8-R."))
 111  '("Cyrillic"))
 112
 113 (set-language-info-alist
 114  "Russian" `((charset cyrillic-iso8859-5)
 115              (nonascii-translation
 116               . ,(get 'cyrillic-koi8-r-nonascii-translation-table
 117                       'translation-table))
 118              (coding-system cyrillic-koi8)
 119              (coding-priority cyrillic-koi8 cyrillic-iso-8bit)
 120              (input-method . "russian-computer")
 121              (features cyril-util)
 122              (unibyte-display . cyrillic-koi8)
 123              (sample-text . "Russian (\e,L@caaZXY\e(B)    \e,L7T`PRabRcYbU\e(B!")
 124              (documentation . "\
 125 Support for Russian using koi8-r and the russian-computer input method.")
 126              (tutorial . "TUTORIAL.ru"))
 127  '("Cyrillic"))
 128
 129 (define-coding-system 'koi8-u
 130   "KOI8-U 8-bit encoding for Cyrillic (MIME: KOI8-U)"
 131   :coding-type 'charset
 132   :mnemonic ?U
 133   :charset-list '(koi8-u)
 134   :mime-charset 'koi8-u)
 135
 136 (set-language-info-alist
 137  "Ukrainian" `((charset koi8-u)
 138                (coding-system koi8-u)
 139                (coding-priority koi8-u)
 140                (nonascii-translation . koi8-u)
 141                (input-method . "ukrainian-computer")
 142                (documentation
 143                 . "Support for Ukrainian with KOI8-U character set."))
 144  '("Cyrillic"))
 145
 146 ;;; ALTERNATIVNYJ stuff
 147
 148 (define-coding-system 'cyrillic-alternativnyj
 149   "ALTERNATIVNYJ 8-bit encoding for Cyrillic."
 150   :coding-type 'charset
 151   :mnemonic ?A
 152   :charset-list '(alternativnyj))
 153
 154 (define-coding-system-alias 'alternativnyj 'cyrillic-alternativnyj)
 155
 156 (set-language-info-alist
 157  "Cyrillic-ALT" `((charset alternativnyj)
 158                   (coding-system cyrillic-alternativnyj)
 159                   (coding-priority cyrillic-alternativnyj)
 160                   (nonascii-translation . alternativnyj)
 161                   (input-method . "russian-typewriter")
 162                   (features cyril-util)
 163                   (unibyte-display . cyrillic-alternativnyj)
 164                   (sample-text . "Russian (\e,L@caaZXY\e(B)       \e,L7T`PRabRcYbU\e(B!")
 165                   (documentation . "Support for Cyrillic ALTERNATIVNYJ."))
 166  '("Cyrillic"))
 167
 168 (define-coding-system 'cp866
 169   "CP866 encoding for Cyrillic."
 170   :coding-type 'charset
 171   :mnemonic ?*
 172   :charset-list '(ibm866)
 173   :mime-charset 'cp866)
 174
 175 (define-coding-system 'koi8-u
 176   "KOI8-U 8-bit encoding for Cyrillic (MIME: KOI8-U)"
 177   :coding-type 'charset
 178   :mnemonic ?U
 179   :charset-list '(koi8-u)
 180   :mime-charset 'koi8-u)
 181
 182 (define-coding-system 'koi8-t
 183   "KOI8-T 8-bit encoding for Cyrillic"
 184   :coding-type 'charset
 185   :mnemonic ?*
 186   :charset-list '(koi8-t)
 187   :mime-charset 'koi8-t)
 188
 189 (define-coding-system 'windows-1251
 190   "windows-1251 8-bit encoding for Cyrillic (MIME: WINDOWS-1251)"
 191   :coding-type 'charset
 192   :mnemonic ?b
 193   :charset-list '(windows-1251)
 194   :mime-charset 'windows-1251)
 195 (define-coding-system-alias 'cp1251 'windows-1251)
 196
 197 (define-coding-system 'cp1125
 198   "cp1125 8-bit encoding for Cyrillic"
 199   :coding-type 'charset
 200   :mnemonic ?*
 201   :charset-list '(cp1125))
 202 (define-coding-system-alias 'ruscii 'cp1125)
 203 ;; Original name for cp1125, says Serhii Hlodin <hlodin@lutsk.bank.gov.ua>
 204 (define-coding-system-alias 'cp866u 'cp1125)
 205
 206 (define-coding-system 'cp855
 207   "DOS codepage 855 (Russian)"
 208   :coding-type 'charset
 209   :mnemonic ?D
 210   :charset-list '(cp855)
 211   :mime-charset 'cp855)
 212 (define-coding-system-alias 'ibm855 'cp855)
 213
 214 (define-coding-system 'mik
 215   "Bulgarian DOS codepage"
 216   :coding-type 'charset
 217   :mnemonic ?D
 218   :charset-list '(mik))
 219
 220 (define-coding-system 'pt154
 221   "Parattype Asian Cyrillic codepage"
 222   :coding-type 'charset
 223   :mnemonic ?D
 224   :charset-list '(pt154))
 225
 226 ;; (set-language-info-alist
 227 ;;  "Windows-1251" `((coding-system windows-1251)
 228 ;;                (coding-priority windows-1251)
 229 ;;                (input-method . "russian-typewriter") ; fixme?
 230 ;;                (features code-pages)
 231 ;;                (documentation . "Support for windows-1251 character set."))
 232 ;;  '("Cyrillic"))
 233
 234 (set-language-info-alist
 235  "Tajik" `((coding-system koi8-t)
 236            (coding-priority koi8-t)
 237            (nonascii-translation . cyrillic-koi8-t)
 238            (charset koi8-t)
 239            (input-method . "russian-typewriter") ; fixme?
 240            (features code-pages)
 241            (documentation . "Support for Tajik using KOI8-T."))
 242  '("Cyrillic"))
 243
 244 (let ((elt `("microsoft-cp1251" windows-1251 1
 245              ,(get 'encode-windows-1251 'translation-table)))
 246       (slot (assoc "microsoft-cp1251" ctext-non-standard-encodings-alist)))
 247   (if slot
 248       (setcdr slot (cdr elt))
 249     (push elt ctext-non-standard-encodings-alist)))
 250
 251 (set-language-info-alist
 252  "Bulgarian" `((coding-system windows-1251)
 253                (coding-priority windows-1251)
 254                (nonascii-translation . windows-1251)
 255                (charset windows-1251)
 256                (ctext-non-standard-encodings "microsoft-cp1251")
 257                (input-method . "bulgarian-bds")
 258                (documentation
 259                 . "Support for Bulgrian with windows-1251 character set."))
 260  '("Cyrillic"))
 261
 262 (set-language-info-alist
 263  "Belarusian" `((coding-system windows-1251)
 264                 (coding-priority windows-1251)
 265                 (nonascii-translation . windows-1251)
 266                 (charset windows-1251)
 267                 (ctext-non-standard-encodings "microsoft-cp1251")
 268                 (input-method . "belarusian")
 269                 (documentation
 270                  . "Support for Belarusian with windows-1251 character set.
 271 \(The name Belarusian replaced Byelorussian in the early 1990s.)"))
 272  '("Cyrillic"))
 273
 274 (set-language-info-alist
 275  "Ukrainian" `((coding-system koi8-u)
 276                (coding-priority koi8-u)
 277                (input-method . "ukrainian-computer")
 278                (documentation
 279                 . "Support for Ukrainian with koi8-u character set."))
 280  '("Cyrillic"))
 281
 282 (provide 'cyrillic)
 283
 284 ;;; arch-tag: bda71ae0-ba41-4cb6-a6e0-1dff542313d3
 285 ;;; cyrillic.el ends here