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, 2008, 2009, 2010, 2011, 2012
   4 ;;   Free Software Foundation, Inc.
   5 ;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
   6 ;;   2005, 2006, 2007, 2008, 2009, 2010, 2011
   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 of the License, or
  21 ;; (at your option) 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.  If not, see <http://www.gnu.org/licenses/>.
  30
  31 ;;; Commentary:
  32
  33 ;; The character set ISO8859-5 is supported.  KOI-8 and ALTERNATIVNYJ
  34 ;; are converted to Unicode internally.  See
  35 ;; <URL:http://www.ecma.ch/ecma1/STAND/ECMA-113.HTM>.  For more info
  36 ;; on Cyrillic charsets, see
  37 ;; <URL:http://czyborra.com/charsets/cyrillic.html>.  The KOI and
  38 ;; Alternativnyj coding systems should live in code-pages.el, but
  39 ;; they've always been preloaded and the coding system autoload
  40 ;; mechanism didn't get accepted, so they have to stay here and
  41 ;; duplicate code-pages stuff.
  42
  43 ;; Note that 8859-5 maps directly onto the Unicode Cyrillic block,
  44 ;; apart from codepoints 160 (NBSP, c.f. U+0400), 173 (soft hyphen,
  45 ;; c.f. U+04OD) and 253 (section sign, c.f U+045D).  The KOI-8 and
  46 ;; Alternativnyj coding systems encode both 8859-5 and Unicode.
  47 ;; ucs-tables.el provides unification for cyrillic-iso-8bit.
  48
  49 ;; Customizing `utf-fragment-on-decoding' allows decoding characters
  50 ;; from KOI and Alternativnyj into 8859-5 where that's possible.
  51 ;; cyrillic-iso8859-5 characters take half as much space in the buffer
  52 ;; as the mule-unicode-0100-24ff equivalents, though that's probably
  53 ;; not normally a big deal.
  54
  55 ;;; Code:
  56
  57 ;; Cyrillic (general)
  58
  59 ;; ISO-8859-5 stuff
  60
  61 (define-coding-system 'cyrillic-iso-8bit
  62   "ISO 2022 based 8-bit encoding for Cyrillic script (MIME:ISO-8859-5)."
  63   :coding-type 'charset
  64   :mnemonic ?5
  65   :charset-list '(iso-8859-5)
  66   :mime-charset 'iso-8859-5)
  67
  68 (define-coding-system-alias 'iso-8859-5 'cyrillic-iso-8bit)
  69
  70 (set-language-info-alist
  71  "Cyrillic-ISO" '((charset iso-8859-5)
  72                   (coding-system cyrillic-iso-8bit)
  73                   (coding-priority cyrillic-iso-8bit)
  74                   (input-method . "cyrillic-yawerty") ; fixme
  75                   (nonascii-translation . iso-8859-5)
  76                   (unibyte-display . cyrillic-iso-8bit)
  77                   (features cyril-util)
  78                   (sample-text . "Russian (\e,L@caaZXY\e(B)       \e,L7T`PRabRcYbU\e(B!")
  79                   (documentation . "Support for Cyrillic ISO-8859-5."))
  80  '("Cyrillic"))
  81
  82 ;; KOI-8R stuff
  83
  84 (define-coding-system 'cyrillic-koi8
  85   "KOI8 8-bit encoding for Cyrillic (MIME: KOI8-R)."
  86   :coding-type 'charset
  87   ;; We used to use ?K.  It is true that ?K is more strictly correct,
  88   ;; but it is also used for Korean.  So people who use koi8 for
  89   ;; languages other than Russian will have to forgive us.
  90   :mnemonic ?R
  91   :charset-list '(koi8)
  92   :mime-charset 'koi8-r)
  93
  94 (define-coding-system-alias 'koi8-r 'cyrillic-koi8)
  95 (define-coding-system-alias 'koi8 'cyrillic-koi8)
  96 (define-coding-system-alias 'cp878 'cyrillic-koi8)
  97
  98 (set-language-info-alist
  99  "Cyrillic-KOI8" `((charset koi8)
 100                    (coding-system cyrillic-koi8)
 101                    (coding-priority cyrillic-koi8 cyrillic-iso-8bit)
 102                    (ctext-non-standard-encodings "koi8-r")
 103                    (nonascii-translation . koi8)
 104                    (input-method . "russian-typewriter")
 105                    (features cyril-util)
 106                    (unibyte-display . cyrillic-koi8)
 107                    (sample-text . "Russian (\e,L@caaZXY\e(B)      \e,L7T`PRabRcYbU\e(B!")
 108                    (documentation . "Support for Cyrillic KOI8-R."))
 109  '("Cyrillic"))
 110
 111 (set-language-info-alist
 112  "Russian" `((charset cyrillic-iso8859-5)
 113              (nonascii-translation
 114               . ,(get 'cyrillic-koi8-r-nonascii-translation-table
 115                       'translation-table))
 116              (coding-system cyrillic-koi8)
 117              (coding-priority cyrillic-koi8 cyrillic-iso-8bit)
 118              (input-method . "russian-computer")
 119              (features cyril-util)
 120              (unibyte-display . cyrillic-koi8)
 121              (sample-text . "Russian (\e,L@caaZXY\e(B)    \e,L7T`PRabRcYbU\e(B!")
 122              (documentation . "\
 123 Support for Russian using koi8-r and the russian-computer input method.")
 124              (tutorial . "TUTORIAL.ru"))
 125  '("Cyrillic"))
 126
 127 (define-coding-system 'koi8-u
 128   "KOI8-U 8-bit encoding for Cyrillic (MIME: KOI8-U)"
 129   :coding-type 'charset
 130   :mnemonic ?U
 131   :charset-list '(koi8-u)
 132   :mime-charset 'koi8-u)
 133
 134 (set-language-info-alist
 135  "Ukrainian" `((charset koi8-u)
 136                (coding-system koi8-u)
 137                (coding-priority koi8-u)
 138                (nonascii-translation . koi8-u)
 139                (input-method . "ukrainian-computer")
 140                (documentation
 141                 . "Support for Ukrainian with KOI8-U character set."))
 142  '("Cyrillic"))
 143
 144 ;;; ALTERNATIVNYJ stuff
 145
 146 (define-coding-system 'cyrillic-alternativnyj
 147   "ALTERNATIVNYJ 8-bit encoding for Cyrillic."
 148   :coding-type 'charset
 149   :mnemonic ?A
 150   :charset-list '(alternativnyj))
 151
 152 (define-coding-system-alias 'alternativnyj 'cyrillic-alternativnyj)
 153
 154 (set-language-info-alist
 155  "Cyrillic-ALT" `((charset alternativnyj)
 156                   (coding-system cyrillic-alternativnyj)
 157                   (coding-priority cyrillic-alternativnyj)
 158                   (nonascii-translation . alternativnyj)
 159                   (input-method . "russian-typewriter")
 160                   (features cyril-util)
 161                   (unibyte-display . cyrillic-alternativnyj)
 162                   (sample-text . "Russian (\e,L@caaZXY\e(B)       \e,L7T`PRabRcYbU\e(B!")
 163                   (documentation . "Support for Cyrillic ALTERNATIVNYJ."))
 164  '("Cyrillic"))
 165
 166 (define-coding-system 'cp866
 167   "CP866 encoding for Cyrillic."
 168   :coding-type 'charset
 169   :mnemonic ?*
 170   :charset-list '(ibm866)
 171   :mime-charset 'cp866)
 172
 173 (define-coding-system 'koi8-u
 174   "KOI8-U 8-bit encoding for Cyrillic (MIME: KOI8-U)"
 175   :coding-type 'charset
 176   :mnemonic ?U
 177   :charset-list '(koi8-u)
 178   :mime-charset 'koi8-u)
 179
 180 (define-coding-system 'koi8-t
 181   "KOI8-T 8-bit encoding for Cyrillic"
 182   :coding-type 'charset
 183   :mnemonic ?*
 184   :charset-list '(koi8-t)
 185   :mime-charset 'koi8-t)
 186
 187 (define-coding-system 'windows-1251
 188   "windows-1251 8-bit encoding for Cyrillic (MIME: WINDOWS-1251)"
 189   :coding-type 'charset
 190   :mnemonic ?b
 191   :charset-list '(windows-1251)
 192   :mime-charset 'windows-1251)
 193 (define-coding-system-alias 'cp1251 'windows-1251)
 194
 195 (define-coding-system 'cp1125
 196   "cp1125 8-bit encoding for Cyrillic"
 197   :coding-type 'charset
 198   :mnemonic ?*
 199   :charset-list '(cp1125))
 200 (define-coding-system-alias 'ruscii 'cp1125)
 201 ;; Original name for cp1125, says Serhii Hlodin <hlodin@lutsk.bank.gov.ua>
 202 (define-coding-system-alias 'cp866u 'cp1125)
 203
 204 (define-coding-system 'cp855
 205   "DOS codepage 855 (Russian)"
 206   :coding-type 'charset
 207   :mnemonic ?D
 208   :charset-list '(cp855)
 209   :mime-charset 'cp855)
 210 (define-coding-system-alias 'ibm855 'cp855)
 211
 212 (define-coding-system 'mik
 213   "Bulgarian DOS codepage"
 214   :coding-type 'charset
 215   :mnemonic ?D
 216   :charset-list '(mik))
 217
 218 (define-coding-system 'pt154
 219   "Parattype Asian Cyrillic codepage"
 220   :coding-type 'charset
 221   :mnemonic ?D
 222   :charset-list '(pt154))
 223
 224 ;; (set-language-info-alist
 225 ;;  "Windows-1251" `((coding-system windows-1251)
 226 ;;                (coding-priority windows-1251)
 227 ;;                (input-method . "russian-typewriter") ; fixme?
 228 ;;                (features code-pages)
 229 ;;                (documentation . "Support for windows-1251 character set."))
 230 ;;  '("Cyrillic"))
 231
 232 (set-language-info-alist
 233  "Tajik" `((coding-system koi8-t)
 234            (coding-priority koi8-t)
 235            (nonascii-translation . cyrillic-koi8-t)
 236            (charset koi8-t)
 237            (input-method . "russian-typewriter") ; fixme?
 238            (features code-pages)
 239            (documentation . "Support for Tajik using KOI8-T."))
 240  '("Cyrillic"))
 241
 242 (set-language-info-alist
 243  "Bulgarian" `((coding-system windows-1251)
 244                (coding-priority windows-1251)
 245                (nonascii-translation . windows-1251)
 246                (charset windows-1251)
 247                (ctext-non-standard-encodings "microsoft-cp1251")
 248                (input-method . "bulgarian-bds")
 249                (documentation
 250                 . "Support for Bulgrian with windows-1251 character set."))
 251  '("Cyrillic"))
 252
 253 (set-language-info-alist
 254  "Belarusian" `((coding-system windows-1251)
 255                 (coding-priority windows-1251)
 256                 (nonascii-translation . windows-1251)
 257                 (charset windows-1251)
 258                 (ctext-non-standard-encodings "microsoft-cp1251")
 259                 (input-method . "belarusian")
 260                 (documentation
 261                  . "Support for Belarusian with windows-1251 character set.
 262 \(The name Belarusian replaced Byelorussian in the early 1990s.)"))
 263  '("Cyrillic"))
 264
 265 (set-language-info-alist
 266  "Ukrainian" `((coding-system koi8-u)
 267                (coding-priority koi8-u)
 268                (input-method . "ukrainian-computer")
 269                (documentation
 270                 . "Support for Ukrainian with koi8-u character set."))
 271  '("Cyrillic"))
 272
 273 (provide 'cyrillic)
 274
 275 ;; arch-tag: bda71ae0-ba41-4cb6-a6e0-1dff542313d3
 276 ;;; cyrillic.el ends here