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

   1 ;;; chinese.el --- support for Chinese -*- coding: iso-2022-7bit; -*-
   2
   3 ;; Copyright (C) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
   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 ;; Keywords: multilingual, Chinese
  14
  15 ;; This file is part of GNU Emacs.
  16
  17 ;; GNU Emacs is free software: you can redistribute it and/or modify
  18 ;; it under the terms of the GNU General Public License as published by
  19 ;; the Free Software Foundation, either version 3 of the License, or
  20 ;; (at your option) any later version.
  21
  22 ;; GNU Emacs is distributed in the hope that it will be useful,
  23 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  24 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  25 ;; GNU General Public License for more details.
  26
  27 ;; You should have received a copy of the GNU General Public License
  28 ;; along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.
  29
  30 ;;; Commentary:
  31
  32 ;; For Chinese, three character sets GB2312, BIG5, and CNS11643 are
  33 ;; supported.
  34
  35 ;;; Code:
  36
  37 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
  38 ;;; Chinese (general)
  39 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
  40
  41
  42 (define-coding-system 'iso-2022-cn
  43   "ISO 2022 based 7bit encoding for Chinese GB and CNS (MIME:ISO-2022-CN)."
  44   :coding-type 'iso-2022
  45   :mnemonic ?C
  46   :charset-list '(ascii chinese-gb2312 chinese-cns11643-1 chinese-cns11643-2)
  47   :designation [ascii
  48                 (nil chinese-gb2312 chinese-cns11643-1)
  49                 (nil chinese-cns11643-2)
  50                 nil]
  51   :flags '(ascii-at-eol ascii-at-cntl 7-bit
  52                         designation locking-shift single-shift init-at-bol)
  53   :mime-charset 'iso-2022-cn
  54   :suitable-for-keyboard t)
  55
  56 (define-coding-system-alias 'chinese-iso-7bit 'iso-2022-cn)
  57
  58 (define-coding-system 'iso-2022-cn-ext
  59   "ISO 2022 based 7bit encoding for Chinese GB and CNS (MIME:ISO-2022-CN-EXT)."
  60   :coding-type 'iso-2022
  61   :mnemonic ?C
  62   :charset-list '(ascii
  63                   chinese-gb2312 chinese-cns11643-1
  64                   chinese-cns11643-2 chinese-cns11643-3 chinese-cns11643-4
  65                   chinese-cns11643-5 chinese-cns11643-6 chinese-cns11643-7)
  66   :designation '[ascii
  67                  (nil chinese-gb2312 chinese-cns11643-1)
  68                  (nil chinese-cns11643-2)
  69                  (nil chinese-cns11643-3 chinese-cns11643-4 chinese-cns11643-5
  70                       chinese-cns11643-6 chinese-cns11643-7)]
  71   :flags '(ascii-at-eol ascii-at-cntl 7-bit
  72                         designation locking-shift single-shift init-at-bol)
  73   :mime-charset 'iso-2022-cn-ext
  74   :suitable-for-keyboard t)
  75
  76 \f
  77 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
  78 ;;; Chinese GB2312 (simplified)
  79 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
  80
  81 (define-coding-system 'chinese-iso-8bit
  82   "ISO 2022 based EUC encoding for Chinese GB2312 (MIME:GB2312)."
  83   :coding-type 'iso-2022
  84   :mnemonic ?c
  85   :charset-list '(ascii chinese-gb2312)
  86   :designation [ascii chinese-gb2312 nil nil]
  87   :mime-charset 'gb2312)
  88
  89 (define-coding-system-alias 'cn-gb-2312 'chinese-iso-8bit)
  90 (define-coding-system-alias 'euc-china 'chinese-iso-8bit)
  91 (define-coding-system-alias 'euc-cn 'chinese-iso-8bit)
  92 (define-coding-system-alias 'cn-gb 'chinese-iso-8bit)
  93 (define-coding-system-alias 'gb2312 'chinese-iso-8bit)
  94
  95 (define-coding-system 'chinese-hz
  96   "Hz/ZW 7-bit encoding for Chinese GB2312 (MIME:HZ-GB-2312)."
  97   :coding-type 'utf-8
  98   :mnemonic ?z
  99   :charset-list '(ascii chinese-gb2312)
 100   :mime-charset 'hz-gb-2312
 101   :post-read-conversion 'post-read-decode-hz
 102   :pre-write-conversion 'pre-write-encode-hz)
 103
 104 (define-coding-system-alias 'hz-gb-2312 'chinese-hz)
 105 (define-coding-system-alias 'hz 'chinese-hz)
 106
 107 (set-language-info-alist
 108  "Chinese-GB" '((charset chinese-gb2312 chinese-sisheng)
 109                 (iso639-language . zh)
 110                 (setup-function . (lambda ()
 111                                     (use-cjk-char-width-table 'zh_CN)))
 112                 (exit-function . use-default-char-width-table)
 113                 (coding-system chinese-iso-8bit iso-2022-cn chinese-hz)
 114                 (coding-priority chinese-iso-8bit chinese-big5 iso-2022-cn)
 115                 (input-method . "chinese-py-punct")
 116                 (features china-util)
 117                 (sample-text . "Chinese (\e$AVPND\e(B,\e$AFUM(;0\e(B,\e$A::So\e(B)    \e$ADc:C\e(B")
 118                 (documentation . "Support for Chinese GB2312 character set.")
 119                 (tutorial . "TUTORIAL.cn"))
 120  '("Chinese"))
 121
 122 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 123 ;; Chinese BIG5 (traditional)
 124 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 125
 126 (define-coding-system 'chinese-big5
 127   "BIG5 8-bit encoding for Chinese (MIME:Big5)"
 128   :coding-type 'big5
 129   :mnemonic ?B
 130   :charset-list '(ascii big5)
 131   :mime-charset 'big5)
 132
 133 (define-coding-system-alias 'big5 'chinese-big5)
 134 (define-coding-system-alias 'cn-big5 'chinese-big5)
 135 (define-coding-system-alias 'cp950 'chinese-big5)
 136
 137 (set-language-info-alist
 138  "Chinese-BIG5" '((charset chinese-big5-1 chinese-big5-2)
 139                   (iso639-language . zh)
 140                   (setup-function . (lambda ()
 141                                       (use-cjk-char-width-table 'zh_HK)))
 142                   (exit-function . use-default-char-width-table)
 143                   (coding-system chinese-big5 chinese-iso-7bit)
 144                   (coding-priority chinese-big5 iso-2022-cn chinese-iso-8bit)
 145                   (input-method . "chinese-py-punct-b5")
 146                   (ctext-non-standard-encodings "big5-0")
 147                   (features china-util)
 148                   (sample-text . "Cantonese (\e$(0GnM$\e(B,\e$(0N]0*Hd\e(B) \e$(0*/=(\e(B, \e$(0+$)p\e(B")
 149                   (documentation . "Support for Chinese Big5 character set.")
 150                   (tutorial . "TUTORIAL.zh"))
 151  '("Chinese"))
 152
 153 (define-coding-system 'chinese-big5-hkscs
 154   "BIG5-HKSCS 8-bit encoding for Chinese, Hong Kong supplement (MIME:Big5-HKSCS)"
 155   :coding-type 'charset
 156   :mnemonic ?B
 157   :charset-list '(ascii big5-hkscs)
 158   :mime-charset 'big5-hkscs)
 159 (define-coding-system-alias 'big5-hkscs 'chinese-big5-hkscs)
 160 (define-coding-system-alias 'cn-big5-hkscs 'chinese-big5-hkscs)
 161
 162 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 163 ;; Chinese CNS11643 (traditional)
 164 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 165
 166 (define-coding-system 'euc-tw
 167   "ISO 2022 based EUC encoding for Chinese CNS11643."
 168   :coding-type 'iso-2022
 169   :mnemonic ?Z
 170   :charset-list '(ascii
 171                   chinese-cns11643-1
 172                   chinese-cns11643-2
 173                   chinese-cns11643-3
 174                   chinese-cns11643-4
 175                   chinese-cns11643-5
 176                   chinese-cns11643-6
 177                   chinese-cns11643-7)
 178   :designation [ascii chinese-cns11643-1 (chinese-cns11643-1
 179                                           chinese-cns11643-2
 180                                           chinese-cns11643-3
 181                                           chinese-cns11643-4
 182                                           chinese-cns11643-5
 183                                           chinese-cns11643-6
 184                                           chinese-cns11643-7) nil]
 185   :mime-charset 'euc-tw)
 186
 187 (define-coding-system-alias 'euc-taiwan 'euc-tw)
 188
 189 (set-language-info-alist
 190  "Chinese-CNS" '((charset chinese-cns11643-1 chinese-cns11643-2
 191                           chinese-cns11643-3 chinese-cns11643-4
 192                           chinese-cns11643-5 chinese-cns11643-6
 193                           chinese-cns11643-7)
 194                  (iso639-language . zh)
 195                  (setup-function . (lambda ()
 196                                      (use-cjk-char-width-table 'zh_TW)))
 197                  (exit-function . use-default-char-width-table)
 198                  (coding-system iso-2022-cn euc-tw)
 199                  (coding-priority iso-2022-cn euc-tw chinese-big5
 200                                   chinese-iso-8bit)
 201                  (features china-util)
 202                  (input-method . "chinese-cns-quick")
 203                  ;; Fixme: presumably it won't accept big5 now.
 204                  (documentation . "\
 205 Support for Chinese CNS character sets.  Note that the EUC-TW coding system
 206 accepts Big5 for input also (which is then converted to CNS)."))
 207  '("Chinese"))
 208
 209 (set-language-info-alist
 210  "Chinese-EUC-TW" '((charset chinese-cns11643-1 chinese-cns11643-2
 211                              chinese-cns11643-3 chinese-cns11643-4
 212                              chinese-cns11643-5 chinese-cns11643-6
 213                              chinese-cns11643-7 chinese-big5-1 chinese-big5-2)
 214                     (iso639-language . zh)
 215                     (setup-function . (lambda ()
 216                                         (use-cjk-char-width-table 'zh_TW)))
 217                     (exit-function . use-default-char-width-table)
 218                     (coding-system euc-tw iso-2022-cn)
 219                     (coding-priority euc-tw chinese-big5 iso-2022-cn
 220                                      chinese-iso-8bit)
 221                     (features china-util)
 222                     (input-method . "chinese-cns-quick")
 223                     (documentation . "\
 224 Support for Chinese, prefering the EUC-TW character set.  Note that
 225 the EUC-TW coding system accepts Big5 for input also (which is then
 226 converted to CNS)."))
 227  '("Chinese"))
 228
 229
 230 ;;; Chinese GBK
 231
 232 (define-coding-system 'chinese-gbk
 233   "GBK encoding for Chinese (MIME:GBK)."
 234   :coding-type 'charset
 235   :mnemonic ?c
 236   :charset-list '(ascii chinese-gbk)
 237   :mime-charset 'gbk)
 238 (define-coding-system-alias 'gbk 'chinese-gbk)
 239 (define-coding-system-alias 'cp936 'chinese-gbk)
 240 (define-coding-system-alias 'windows-936 'chinese-gbk)
 241
 242 (set-language-info-alist
 243  "Chinese-GBK" '((charset chinese-gbk)
 244                  (iso639-language . zh)
 245                  (setup-function . (lambda ()
 246                                      (use-cjk-char-width-table 'zh_CN)))
 247                  (exit-function . use-default-char-width-table)
 248                  (coding-system chinese-gbk)
 249                  (coding-priority gbk iso-2022-cn chinese-big5
 250                                   chinese-iso-8bit) ; fixme?
 251                  (ctext-non-standard-encodings "gbk-0")
 252                  (input-method . "chinese-py-punct") ; fixme?
 253                  (sample-text . "Chinese (\e$BCfJ8\e(B,\e$BIaDL\e$A;0\e(B,\e$A::So\e(B) \e$(D95\e$B9%\e(B")
 254                  (features china-util)
 255                  (documentation . "Support for Chinese GBK character set.")
 256                  (tutorial . "TUTORIAL.cn"))
 257  '("Chinese"))
 258
 259 ;;; Chinese GB18030
 260
 261 (define-coding-system 'chinese-gb18030
 262   "GB18030 encoding for Chinese (MIME:GB18030)."
 263   :coding-type 'charset
 264   :mnemonic ?c
 265   :charset-list '(ascii gb18030-2-byte
 266                         gb18030-4-byte-bmp gb18030-4-byte-smp
 267                         gb18030-4-byte-ext-1 gb18030-4-byte-ext-2)
 268   :mime-charset 'gb18030)
 269
 270 (define-coding-system-alias 'gb18030 'chinese-gb18030)
 271
 272 (set-language-info-alist
 273  "Chinese-GB18030" '((charset gb18030)
 274                      (iso639-language . zh)
 275                      (coding-system chinese-gb18030)
 276                      (coding-priority gb18030 gbk iso-2022-cn chinese-big5
 277                                       chinese-iso-8bit) ; fixme?
 278                      (input-method . "chinese-py-punct") ; fixme?
 279                      (sample-text . "Chinese (\e$BCfJ8\e(B,\e$BIaDL\e$A;0\e(B,\e$A::So\e(B) \e$(D0_\e$B9%\e(B")
 280                      (features china-util)
 281                      (documentation
 282                       . "Support for Chinese GB18030 character set.")
 283                      (tutorial . "TUTORIAL.cn"))
 284  '("Chinese"))
 285
 286 ;; Fixme: add HKSCS
 287
 288 (provide 'chinese)
 289
 290 ;; arch-tag: b82fcf7a-84f6-4e0b-b38c-1742dac0e09f
 291 ;;; chinese.el ends here