Mercurial > hg > xemacs-beta

comparison lisp/language/indian.el @ 159:3bb7ccffb0c0 r20-3b6

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Import from CVS: tag r20-3b6
author cvs
date Mon, 13 Aug 2007 09:41:43 +0200
parents
children
comparison
equal deleted inserted replaced
158:558dfa75ffb3 159:3bb7ccffb0c0
1 ;;; indian.el --- Support for Indian Languages
2
3 ;; Copyright (C) 1995 Free Software Foundation, Inc.
4
5 ;; Author: KAWABATA, Taichi <kawabata@is.s.u-tokyo.ac.jp>
6
7 ;; Keywords: multilingual, Indian
8
9 ;; This file is part of GNU Emacs.
10
11 ;; GNU Emacs is free software; you can redistribute it and/or modify
12 ;; it under the terms of the GNU General Public License as published by
13 ;; the Free Software Foundation; either version 2, or (at your option)
14 ;; any later version.
15
16 ;; GNU Emacs is distributed in the hope that it will be useful,
17 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
18 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 ;; GNU General Public License for more details.
20
21 ;; You should have received a copy of the GNU General Public License
22 ;; along with GNU Emacs; see the file COPYING. If not, write to the
23 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
24 ;; Boston, MA 02111-1307, USA.
25
26 ;;; Commentary:
27
28 ;; History:
29 ;; 1996.10.18 written by KAWABATA, Taichi <kawabata@is.s.u-tokyo.ac.jp>
30
31 ;; For Indian, the character set IS 13194 is supported.
32 ;;
33 ;; IS 13194 does not specifically assign glyphs for each characters.
34 ;; Following code is not specific to each Indian language.
35 ;;
36 ;; Eventually, this code will support generic information about
37 ;; following scripts.
38 ;;
39 ;; Devanagari
40 ;; Bengali
41 ;; Gurmukhi
42 ;; Gujarati
43 ;; Oriya
44 ;; Tamil
45 ;; Telgu
46 ;; Kannada
47 ;; Malayalam
48 ;;
49 ;; In this file, charsets other than charset-ascii and charset-indian-is13194
50 ;; should not be used except in the comment.
51
52 ;;; Code:
53
54 (defun describe-indian-support ()
55 "Describe how Emacs support Indian languages."
56 (interactive)
57 (describe-language-support-internal "Indian"))
58
59 (set-language-info-alist
60 "Indian" '((describe-function . describe-indian-support)
61 (charset . (indian-is13194))
62 (documentation . "\
63 Among Indian languages, such languages using Devanagari scripts as
64 Hindi and Marathi are supproted. For them, please specify Devanagari
65 for more detail.")
66 ))
67
68 ;; Followings are what you see when you refer to the Emacs
69 ;; representations of IS 13194 charcters. However, this is merely
70 ;; tentative apperance, and you must convert them by
71 ;; indian-to-xxxxxx(specific script) function to use them.
72 ;; Devanagari is not an exception of this rule.
73
74 ;; 0xa0 //(5!"#$%&'()*+,-./(B
75 ;; 0xb0 (50123456789:;<=>?(B
76 ;; 0xc0 (5@ABCDEFGHIJKLMNO(B
77 ;; 0xd0 (5PQRSTUVWXYZ[\]^_(B
78 ;; 0xe0 (5`abcdefghijklmno(B
79 ;; 0xf0 (5pqrstuvwxyz{|}~(B//
80
81 ;; Note - In IS 13194, several symbols are obtained by special
82 ;; combination of several characters and Nukta sign.
83 ;;
84 ;; Sanskrit Vowel R -> (5*(B + (5i(B
85 ;; Sanskrit Vowel L -> (5&(B + (5i(B
86 ;; Sanskrit Vowel LL -> (5'(B + (5i(B
87 ;; Sanskrit Avagrah -> (5j(B + (5i(B
88 ;; OM -> (5!(B + (5i(B
89 ;;
90 ;; Note - IS 13194 defines ATR(0xEF) and EXT(0xF0), but they are
91 ;; not used in Emacs.
92 ;;
93 ;; Note - the above characters DO NOT represent any script. For
94 ;; example, if you want to obtain Devanagari character, you must do
95 ;; something like the following.
96 ;;
97 ;; (char-to-string (indian-to-devanagari ?(5$(B))
98 ;; "$(5!$(B"
99
100 (let ((deflist
101 '(;; chars syntax category
102 ("(5!"#(B" "w" ?7) ; vowel-modifying diacritical mark
103 ; chandrabindu, anuswar, visarga
104 ("(5$(B-(52(B" "w" ?5) ; independent vowel
105 ("(53(B-(5X(B" "w" ?0) ; consonant
106 ("(5Z(B-(5g(B" "w" ?8) ; matra
107 ("(5q(B-(5z(B" "w" ?6) ; digit
108 ))
109 elm chars len syntax category to ch i)
110 (while deflist
111 (setq elm (car deflist))
112 (setq chars (car elm)
113 len (length chars)
114 syntax (nth 1 elm)
115 category (nth 2 elm)
116 i 0)
117 (while (< i len)
118 (if (= (aref chars i) ?-)
119 (setq i (1+ i)
120 to (sref chars i))
121 (setq ch (sref chars i)
122 to ch))
123 (while (<= ch to)
124 (modify-syntax-entry ch syntax)
125 (modify-category-entry ch category)
126 (setq ch (1+ ch)))
127 (setq i (+ i (char-bytes to))))
128 (setq deflist (cdr deflist))))
129
130
131 ;;; ITRANS
132 ;;
133 ;; ITRANS is one of the most popular method to exchange indian scripts
134 ;; electronically. Here is the table to convert between ITRANS code and
135 ;; IS 13194 code.
136
137 (defvar indian-itrans-consonant-alist
138 '(
139 ("k" . "(53(B")
140 ("kh" . "(54(B")
141 ("g" . "(55(B")
142 ("gh" . "(56(B")
143 ("N^" . "(57(B")
144 ("ch" . "(58(B")
145 ("chh" . "(59(B")
146 ("j" . "(5:(B")
147 ("jh" . "(5;(B")
148 ("JN" . "(5<(B")
149 ("T" . "(5=(B")
150 ("Th" . "(5>(B")
151 ("D" . "(5?(B")
152 ("Dh" . "(5@(B")
153 ("N" . "(5A(B")
154 ("t" . "(5B(B")
155 ("th" . "(5C(B")
156 ("d" . "(5D(B")
157 ("dh" . "(5E(B")
158 ("n" . "(5F(B")
159 ("nh" . "(5G(B") ; For transcription of non-Devanagari Languages.
160 ("p" . "(5H(B")
161 ("ph" . "(5I(B")
162 ("b" . "(5J(B")
163 ("bh" . "(5K(B")
164 ("m" . "(5L(B")
165 ("y" . "(5M(B")
166 ("yh" . "(5N(B") ; For transcription of non-Devanagari Languages.
167 ("r" . "(5O(B")
168 ("rh" . "(5P(B") ; For transcription of non-Devanagari Languages.
169 ("l" . "(5Q(B")
170 ("v" . "(5T(B")
171 ("sh" . "(5U(B")
172 ("shh" . "(5V(B")
173 ("s" . "(5W(B")
174 ("h" . "(5X(B")
175 ("ld" . "(5R(B")
176 ("L" . "(5R(B")
177 ("ksh" . "$(5!3!h!V(B")
178 ("GY" . "***GY***") ; Must check out later.
179 ;; special consonants
180 ("q" . "(53i(B")
181 ("K" . "(54i(B")
182 ("G" . "(55i(B")
183 ("z" . "(5:i(B")
184 ("f" . "(5Ii(B")
185 (".D" . "(5?i(B")
186 (".Dh" . "(5@i(B")
187 ))
188
189 (defvar indian-itrans-vowel-sign-alist
190 '(
191 ;; Special treatment unique to IS 13194 Transliteration
192 ("" . "(5h(B")
193 ("a" . "")
194 ;; Matra (Vowel Sign)
195 ("aa" . "(5Z(B")
196 ("A" . "(5Z(B")
197 ("i" . "(5[(B")
198 ("ii" . "(5\(B")
199 ("I" . "(5\(B")
200 ("u" . "(5](B")
201 ("uu" . "(5^(B")
202 ("U" . "(5^(B")
203 ("R^i" . "(5_(B") ; These must be checked out later.
204 ("R^I" . "(5_i(B")
205 ("L^i" . "(5[i(B")
206 ("L^I" . "(5\i(B")
207 ("E" . "(5`(B") ; For transcription of non-Devanangri Languages.
208 ("e" . "(5a(B")
209 ("ai" . "(5b(B")
210 ;; ("e.c" . "(5c(B") ; Tentatively suppressed.
211 ("O" . "(5d(B") ; For transcription of non-Devanagari Languages.
212 ("o" . "(5e(B")
213 ("au" . "(5f(B")
214 ;; ("o.c" . "(5g(B") ; Tentatively suppressed.
215 ))
216
217 ;;
218 ;; Independent vowels and other signs.
219 ;;
220
221 (defvar indian-itrans-other-letters-alist
222 '(
223 ("a" . "(5$(B")
224 ("aa" . "(5%(B")
225 ("A" . "(5%(B")
226 ("i" . "(5&(B")
227 ("ii" . "(5'(B")
228 ("I" . "(5'(B")
229 ("u" . "(5((B")
230 ("uu" . "(5)(B")
231 ("U" . "(5)(B")
232 ("R^i" . "(5*(B")
233 ("R^I" . "(5*i(B")
234 ("L^i" . "(5&i(B")
235 ("L^I" . "(5'i(B")
236 ("E" . "(5+(B") ; For transcription of non-Devanagari Languages.
237 ("e" . "(5,(B")
238 ("ai" . "(5-(B")
239 ;; ("e.c" . "(5.(B") ; Candra E
240 ("O" . "(5/(B") ; For transcription of non-Devanagari Languages.
241 ("o" . "(50(B")
242 ("au" . "(51(B")
243 ;; ("o.c" . "(52(B") ; Candra O
244 ("M" . "(5$(B")
245 ("H" . "(5#(B")
246 ("AUM" . "(5!i(B")
247 ("OM" . "(5!i(B")
248 (".r" . "(5Oh(B")
249 (".n" . "(5"(B")
250 (".N" . "(5!(B")
251 (".h" . "(5h(B") ; Halant
252 (".." . "(5j(B")
253 (".a" . "(5ji(B") ; Avagrah
254 ("0" . "(5q(B")
255 ("1" . "(5r(B")
256 ("2" . "(5s(B")
257 ("3" . "(5t(B")
258 ("4" . "(5u(B")
259 ("5" . "(5v(B")
260 ("6" . "(5w(B")
261 ("7" . "(5x(B")
262 ("8" . "(5y(B")
263 ("9" . "(5z(B")
264 ))
265
266 ;; Regular expression matching single Indian character represented
267 ;; by ITRANS.
268
269 (defvar indian-itrans-regexp
270 (let ((consonant "\\([cs]hh?\\)\\|[kgjTDnpbyr]h?\\|\\(N\\^?\\)\\|\\(jN\\)\\|[mvqKGzfs]\\|\\(ld?\\)\\|\\(ksh\\)\\|\\(GY\\)\\|\\(\\.Dh?\\)")
271 (vowel "\\(a[aiu]\\)\\|\\(ii\\)\\|\\(uu\\)\\|\\([RL]\\^[iI]\\)\\|[AIEOeoaiu]")
272 (misc "[MH0-9]\\|\\(AUM\\)\\|\\(OM\\)\\|\\(\\.[rnNh\\.a]\\)")
273 (lpre "\\(") (rpre "\\)") (orre "\\|"))
274 (concat lpre misc rpre orre
275 lpre lpre consonant rpre "?" lpre vowel rpre rpre orre
276 lpre consonant rpre )))
277
278 ;;
279 ;; Regular expression matching single ITRANS unit for IS 13194 characters.
280 ;;
281
282 (defvar itrans-indian-regexp
283 (let ((vowel "[(5$(B-(52(B]")
284 (consonant "[(53(B-(5X(B]")
285 (matra "[(5Z(B-(5g(B]")
286 (misc "[(5q(B-(5z(B]")
287 (lpre "\\(") (rpre "\\)") (orre "\\|"))
288 (concat misc orre
289 lpre consonant matra "?" rpre orre
290 vowel)))
291
292 ;;
293 ;; IS13194 - ITRANS conversion table for string matching above regexp.
294 ;;
295
296 (defvar indian-itrans-alist
297 (let ((cl indian-itrans-consonant-alist)
298 (ml indian-itrans-other-letters-alist) rules)
299 (while cl
300 (let ((vl indian-itrans-vowel-sign-alist))
301 (while vl
302 (setq rules
303 (cons (cons (concat (car (car cl)) (car (car vl)))
304 (concat (cdr (car cl)) (cdr (car vl))))
305 rules))
306 (setq vl (cdr vl))))
307 (setq cl (cdr cl)))
308 (while ml
309 (setq rules (cons (cons (car (car ml))
310 (cdr (car ml)))
311 rules))
312 (setq ml (cdr ml)))
313 rules))
314
315 ;;
316 ;; Utility program to convert from ITRANS to IS 13194 in specified region.
317 ;;
318
319 (defun indian-decode-itrans-region (from to)
320 "Convert `ITRANS' mnemonics of the current region to Indian characters.
321 When called from a program, expects two arguments,
322 positions (integers or markers) specifying the stretch of the region."
323 (interactive "r")
324 (save-restriction
325 (narrow-to-region from to)
326 (goto-char (point-min))
327 (while (re-search-forward indian-itrans-regexp nil t)
328 (let* ((itrans (buffer-substring (match-beginning 0) (match-end 0)))
329 (ch (cdr (assoc itrans indian-itrans-alist))))
330 (if ch
331 (progn
332 (delete-region (match-beginning 0) (match-end 0))
333 (insert ch)))))
334 (goto-char (point-min))
335 (while (re-search-forward "\\((5h(B\\)[^\\c0]" nil t)
336 (delete-region (match-beginning 1) (match-end 1)))))
337
338 ;;
339 ;; Utility program to convert from IS 13194 to ITRANS in specified region.
340 ;;
341
342 (defun indian-encode-itrans-region (from to)
343 "Convert indian region to ITRANS mnemonics."
344 (interactive "r")
345 (save-restriction
346 (narrow-to-region from to)
347 (goto-char (point-min))
348 (while (re-search-forward itrans-indian-regexp nil t)
349 (let* ((indian (buffer-substring (match-beginning 0) (match-end 0)))
350 (ch (car (rassoc indian indian-itrans-alist))))
351 (if ch
352 (progn
353 (delete-region (match-beginning 0) (match-end 0))
354 (insert ch)))))
355 (goto-char (point-min))))
356
357 ;;; indian.el ends here