lisp/language/japanese.el

   1 ;;; japanese.el --- support for Japanese -*- coding: iso-2022-7bit; no-byte-compile: t -*-
   2
   3 ;; Copyright (C) 1997, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
   4 ;;   Free Software Foundation, Inc.
   5 ;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
   6 ;;   2005, 2006, 2007, 2008
   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, Japanese
  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, or (at your option)
  20 ;; 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; see the file COPYING.  If not, write to the
  29 ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  30 ;; Boston, MA 02110-1301, USA.
  31
  32 ;;; Commentary:
  33
  34 ;; For Japanese, character sets JISX0201, JISX0208, JISX0212 are
  35 ;; supported.
  36
  37 ;;; Code:
  38
  39 ;;; Load translation tables for CP932.
  40 (load "international/cp51932")
  41 (load "international/eucjp-ms")
  42
  43 (define-coding-system 'iso-2022-jp
  44   "ISO 2022 based 7bit encoding for Japanese (MIME:ISO-2022-JP)."
  45   :coding-type 'iso-2022
  46   :mnemonic ?J
  47   :designation [(ascii japanese-jisx0208-1978 japanese-jisx0208
  48                        latin-jisx0201)
  49                 nil nil nil]
  50   :flags '(short ascii-at-eol ascii-at-cntl 7-bit designation)
  51   :charset-list '(ascii japanese-jisx0208
  52                         japanese-jisx0208-1978 latin-jisx0201)
  53   :mime-charset 'iso-2022-jp
  54   :suitable-for-keyboard t)
  55
  56 (define-coding-system-alias 'junet 'iso-2022-jp)
  57
  58 (define-coding-system 'iso-2022-jp-2
  59   "ISO 2022 based 7bit encoding for CJK, Latin-1, Greek (MIME:ISO-2022-JP-2)."
  60   :coding-type 'iso-2022
  61   :mnemonic ?J
  62   :designation [(ascii japanese-jisx0208-1978 japanese-jisx0208
  63                        latin-jisx0201 japanese-jisx0212
  64                        chinese-gb2312 korean-ksc5601)
  65                 nil
  66                 (nil latin-iso8859-1 greek-iso8859-7)
  67                 nil]
  68   :flags '(short ascii-at-eol ascii-at-cntl 7-bit designation single-shift
  69                  init-at-bol)
  70   :charset-list '(ascii japanese-jisx0208 japanese-jisx0212
  71                         latin-jisx0201 japanese-jisx0208-1978
  72                         chinese-gb2312 korean-ksc5601
  73                         latin-iso8859-1 greek-iso8859-7)
  74   :mime-charset 'iso-2022-jp-2
  75   :suitable-for-keyboard t)
  76
  77 (let ((map                      ; JIS           vs      CP932
  78        '((#x301C . #xFF5E)      ; WAVE DASH             FULLWIDTH TILDE
  79          (#x2014 . #x2015)      ; EM DASH               HORIZONTAL BAR
  80          (#x2016 . #x2225)      ; DOUBLE VERTICAL LINE  PARALLEL TO
  81          (#x2212 . #xFF0D)      ; MINUS SIGN            FULLWIDTH HYPHEN-MINUS
  82          (#x00A2 . #xFFE0)      ; CENT SIGN             FULLWIDTH CENT SIGN
  83          (#x00A3 . #xFFE1)      ; POUND SIGN            FULLWIDTH POUND SIGN
  84          (#x00AC . #xFFE2)      ; NOT SIGN              FULLWIDTH NOT SIGN
  85          (#x00A6 . #xFFE4)      ; BROKEN LINE           FULLWIDTH BROKEN LINE
  86          )))
  87   (define-translation-table 'japanese-ucs-jis-to-cp932-map map)
  88   (mapc #'(lambda (x) (let ((tmp (car x)))
  89                         (setcar x (cdr x)) (setcdr x tmp)))
  90         map)
  91   (define-translation-table 'japanese-ucs-cp932-to-jis-map map))
  92
  93 ;; U+2014 (EM DASH) vs U+2015 (HORIZONTAL BAR)
  94 (define-translation-table 'japanese-ucs-glibc-to-jis-map '((#x2015 . #x2014)))
  95 (define-translation-table 'japanese-ucs-jis-to-glibc-map '((#x2014 . #x2015)))
  96
  97 (define-coding-system 'japanese-shift-jis
  98   "Shift-JIS 8-bit encoding for Japanese (MIME:SHIFT_JIS)"
  99   :coding-type 'shift-jis
 100   :mnemonic ?S
 101   :charset-list '(ascii katakana-jisx0201 japanese-jisx0208)
 102   :mime-charset 'shift_jis)
 103
 104 (define-coding-system-alias 'shift_jis 'japanese-shift-jis)
 105 (define-coding-system-alias 'sjis 'japanese-shift-jis)
 106 (define-coding-system-alias 'cp932 'japanese-shift-jis)
 107
 108 (define-coding-system 'japanese-cp932
 109   "CP932 (Microsoft shift-jis)"
 110   :coding-type 'charset
 111   :mnemonic ?S
 112   :charset-list '(ascii katakana-sjis cp932-2-byte))
 113
 114 (define-coding-system-alias 'cp932 'japanese-cp932)
 115
 116 (define-coding-system 'japanese-iso-7bit-1978-irv
 117   "ISO 2022 based 7-bit encoding for Japanese JISX0208-1978 and JISX0201-Roman."
 118   :coding-type 'iso-2022
 119   :mnemonic ?j
 120   :designation [(latin-jisx0201 japanese-jisx0208-1978 japanese-jisx0208
 121                                 japanese-jisx0212 katakana-jisx0201)
 122                 nil nil nil]
 123   :flags '(short ascii-at-eol ascii-at-cntl 7-bit designation
 124                  use-roman use-oldjis)
 125   :charset-list '(ascii latin-jisx0201 japanese-jisx0208-1978 japanese-jisx0208
 126                         japanese-jisx0212))
 127
 128 (define-coding-system-alias 'iso-2022-jp-1978-irv 'japanese-iso-7bit-1978-irv)
 129 (define-coding-system-alias 'old-jis 'japanese-iso-7bit-1978-irv)
 130
 131 (define-coding-system 'japanese-iso-8bit
 132   "ISO 2022 based EUC encoding for Japanese (MIME:EUC-JP)."
 133   :coding-type 'iso-2022
 134   :mnemonic ?E
 135   :designation [ascii japanese-jisx0208 katakana-jisx0201 japanese-jisx0212]
 136   :flags '(short ascii-at-eol ascii-at-cntl single-shift)
 137   :charset-list '(ascii latin-jisx0201 japanese-jisx0208
 138                         katakana-jisx0201 japanese-jisx0212
 139                         japanese-jisx0208-1978)
 140   :mime-charset 'euc-jp)
 141
 142 (define-coding-system-alias 'euc-japan-1990 'japanese-iso-8bit)
 143 (define-coding-system-alias 'euc-japan 'japanese-iso-8bit)
 144 (define-coding-system-alias 'euc-jp 'japanese-iso-8bit)
 145
 146 (define-coding-system 'eucjp-ms
 147   "eucJP-ms (like EUC-JP but with CP932 extension).
 148 eucJP-ms is defined in <http://www.opengroup.or.jp/jvc/cde/appendix.html>."
 149   :coding-type 'iso-2022
 150   :mnemonic ?E
 151   :designation [ascii japanese-jisx0208 katakana-jisx0201 japanese-jisx0212]
 152   :flags '(short ascii-at-eol ascii-at-cntl single-shift)
 153   :charset-list '(ascii latin-jisx0201 japanese-jisx0208
 154                         katakana-jisx0201 japanese-jisx0212)
 155   :decode-translation-table 'eucjp-ms-decode
 156   :encode-translation-table 'eucjp-ms-encode)
 157
 158 (define-coding-system 'iso-2022-jp-2004
 159   "ISO 2022 based 7bit encoding for JIS X 0213:2004 (MIME:ISO-2022-JP-2004)."
 160   :coding-type 'iso-2022
 161   :mnemonic ?J
 162   :designation [(ascii japanese-jisx0208 japanese-jisx0213.2004-1
 163                        japanese-jisx0213-1 japanese-jisx0213-2)
 164                 nil nil nil]
 165   :flags '(short ascii-at-eol ascii-at-cntl 7-bit designation)
 166                  ;; init-at-bol)
 167   :charset-list '(ascii japanese-jisx0208 japanese-jisx0213.2004-1
 168                         japanese-jisx0213-1 japanese-jisx0213-2)
 169   :mime-charset 'iso-2022-jp-2004
 170   :suitable-for-keyboard t)
 171
 172 (define-coding-system-alias 'iso-2022-jp-3 'iso-2022-jp-2004)
 173
 174 (define-coding-system 'euc-jis-2004
 175   "ISO 2022 based EUC encoding for JIS X 0213 (MIME:EUC-JIS-2004)."
 176   :coding-type 'iso-2022
 177   :mnemonic ?E
 178   :designation [ascii japanese-jisx0213.2004-1 katakana-jisx0201
 179                       japanese-jisx0213-2]
 180   :flags '(short ascii-at-eol ascii-at-cntl single-shift)
 181   :charset-list '(ascii latin-jisx0201 japanese-jisx0213.2004-1
 182                         japanese-jisx0213-1 katakana-jisx0201
 183                         japanese-jisx0213-2)
 184   :mime-charset 'euc-jis-2004)
 185
 186 (define-coding-system-alias 'euc-jisx0213 'euc-jis-2004)
 187
 188 (define-coding-system 'japanese-shift-jis-2004
 189   "Shift_JIS 8-bit encodinf for Japanese (MIME:SHIFT_JIS-2004)"
 190   :coding-type 'shift-jis
 191   :mnemonic ?S
 192   :charset-list '(ascii katakana-jisx0201
 193                         japanese-jisx0213.2004-1 japanese-jisx0213-2))
 194
 195 (define-coding-system-alias 'shift_jis-2004 'japanese-shift-jis-2004)
 196
 197 (set-language-info-alist
 198  "Japanese" '((setup-function . setup-japanese-environment-internal)
 199               (tutorial . "TUTORIAL.ja")
 200               (charset japanese-jisx0208
 201                        japanese-jisx0212 latin-jisx0201 katakana-jisx0201
 202                        japanese-jisx0213.2004-1 japanese-jisx0213-1
 203                        japanese-jisx0213-2 japanese-jisx0208-1978)
 204               (coding-system iso-2022-jp japanese-iso-8bit
 205                              japanese-shift-jis japanese-iso-7bit-1978-irv
 206                              iso-2022-jp-2004 japanese-shift-jis-2004
 207                              euc-jis-2004)
 208               (coding-priority iso-2022-jp japanese-iso-8bit
 209                                japanese-shift-jis
 210                                iso-2022-jp-2004 euc-jis-2004
 211                                japanese-shift-jis-2004
 212                                iso-2022-jp-2)
 213               (input-method . "japanese")
 214               (features japan-util)
 215               (sample-text . "Japanese (\e$BF|K\8l\e(B)   \e$B$3$s$K$A$O\e(B, \e(I:]FAJ\e(B")
 216               (documentation . t)))
 217
 218 (let ((map
 219        ;; JISX0213-1 vs Unicode
 220        '((#x2477 . [#x304B #x309A])
 221          (#x2478 . [#x304D #x309A])
 222          (#x2479 . [#x304F #x309A])
 223          (#x247a . [#x3051 #x309A])
 224          (#x247b . [#x3053 #x309A])
 225          (#x2577 . [#x30AB #x309A])
 226          (#x2578 . [#x30AD #x309A])
 227          (#x2579 . [#x30AF #x309A])
 228          (#x257a . [#x30B1 #x309A])
 229          (#x257b . [#x30B3 #x309A])
 230          (#x257c . [#x30BB #x309A])
 231          (#x257d . [#x30C4 #x309A])
 232          (#x257e . [#x30C8 #x309A])
 233          (#x2678 . [#x31F7 #x309A])
 234          (#x2b44 . [#x00E6 #x0300])
 235          (#x2b48 . [#x0254 #x0300])
 236          (#x2b49 . [#x0254 #x0301])
 237          (#x2b4a . [#x028C #x0300])
 238          (#x2b4b . [#x028C #x0301])
 239          (#x2b4c . [#x0259 #x0300])
 240          (#x2b4d . [#x0259 #x0301])
 241          (#x2b4e . [#x025A #x0300])
 242          (#x2b4f . [#x025A #x0301])
 243          (#x2b65 . [#x02E9 #x02E5])
 244          (#x2b66 . [#x02E5 #x02E9])))
 245       table)
 246   (dolist (elt map)
 247     (setcar elt (decode-char 'japanese-jisx0213-1 (car elt))))
 248   (setq table (make-translation-table-from-alist map))
 249   (define-translation-table 'jisx0213-to-unicode table)
 250   (define-translation-table 'unicode-to-jisx0213
 251     (char-table-extra-slot table 0)))
 252
 253 (provide 'japanese)
 254
 255 ;;; arch-tag: 450f5537-9d53-4d5e-b731-4cf116d8cbc9
 256 ;;; japanese.el ends here