Mercurial > emacs
annotate lisp/language/thai-util.el @ 89407:a9b9211d250a
*** empty log message ***
author | Kenichi Handa <handa@m17n.org> |
---|---|
date | Sat, 12 Apr 2003 02:37:18 +0000 |
parents | b1774ad77cc7 |
children | 2f877ed80fa6 |
rev | line source |
---|---|
36683
b82a6fbaae16
Add coding: iso-2022-7bit tag.
Kenichi Handa <handa@m17n.org>
parents:
29360
diff
changeset
|
1 ;;; thai-util.el --- utilities for Thai -*- coding: iso-2022-7bit; -*- |
17052 | 2 |
3 ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN. | |
18377
8b4a66c66dd6
Change copyright notice.
Richard M. Stallman <rms@gnu.org>
parents:
18309
diff
changeset
|
4 ;; Licensed to the Free Software Foundation. |
88761 | 5 ;; Copyright (C) 2001 Free Software Foundation, Inc. |
17052 | 6 |
88761 | 7 ;; Keywords: mule, multilingual, Thai, i18n |
17052 | 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 | |
17071 | 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. | |
17052 | 25 |
38414
67b464da13ec
Some fixes to follow coding conventions.
Pavel Janík <Pavel@Janik.cz>
parents:
36683
diff
changeset
|
26 ;;; Commentary: |
67b464da13ec
Some fixes to follow coding conventions.
Pavel Janík <Pavel@Janik.cz>
parents:
36683
diff
changeset
|
27 |
17052 | 28 ;;; Code: |
29 | |
30 ;; Setting information of Thai characters. | |
31 | |
26892
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
32 (defconst thai-category-table (make-category-table)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
33 (define-category ?c "Thai consonant" thai-category-table) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
34 (define-category ?v "Thai upper/lower vowel" thai-category-table) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
35 (define-category ?t "Thai tone" thai-category-table) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
36 |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
37 ;; The general composing rules are as follows: |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
38 ;; |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
39 ;; T |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
40 ;; V T V T |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
41 ;; CV -> C, CT -> C, CVT -> C, Cv -> C, CvT -> C |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
42 ;; v v |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
43 ;; |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
44 ;; where C: consonant, V: vowel upper, v: vowel lower, T: tone mark. |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
45 |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
46 (defvar thai-composition-pattern "\\cc\\(\\ct\\|\\cv\\ct?\\)" |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
47 "Regular expression matching a Thai composite sequence.") |
23546
d180b9c69e4a
(thai-pre-write-conversion): Cancel previous
Kenichi Handa <handa@m17n.org>
parents:
23523
diff
changeset
|
48 |
17052 | 49 (let ((l '((?,T!(B consonant "LETTER KO KAI") ; 0xA1 |
50 (?,T"(B consonant "LETTER KHO KHAI") ; 0xA2 | |
51 (?,T#(B consonant "LETTER KHO KHUAT") ; 0xA3 | |
52 (?,T$(B consonant "LETTER KHO KHWAI") ; 0xA4 | |
53 (?,T%(B consonant "LETTER KHO KHON") ; 0xA5 | |
54 (?,T&(B consonant "LETTER KHO RAKHANG") ; 0xA6 | |
55 (?,T'(B consonant "LETTER NGO NGU") ; 0xA7 | |
56 (?,T((B consonant "LETTER CHO CHAN") ; 0xA8 | |
57 (?,T)(B consonant "LETTER CHO CHING") ; 0xA9 | |
58 (?,T*(B consonant "LETTER CHO CHANG") ; 0xAA | |
59 (?,T+(B consonant "LETTER SO SO") ; 0xAB | |
60 (?,T,(B consonant "LETTER CHO CHOE") ; 0xAC | |
61 (?,T-(B consonant "LETTER YO YING") ; 0xAD | |
62 (?,T.(B consonant "LETTER DO CHADA") ; 0xAE | |
63 (?,T/(B consonant "LETTER TO PATAK") ; 0xAF | |
64 (?,T0(B consonant "LETTER THO THAN") ; 0xB0 | |
65 (?,T1(B consonant "LETTER THO NANGMONTHO") ; 0xB1 | |
66 (?,T2(B consonant "LETTER THO PHUTHAO") ; 0xB2 | |
67 (?,T3(B consonant "LETTER NO NEN") ; 0xB3 | |
68 (?,T4(B consonant "LETTER DO DEK") ; 0xB4 | |
69 (?,T5(B consonant "LETTER TO TAO") ; 0xB5 | |
70 (?,T6(B consonant "LETTER THO THUNG") ; 0xB6 | |
71 (?,T7(B consonant "LETTER THO THAHAN") ; 0xB7 | |
72 (?,T8(B consonant "LETTER THO THONG") ; 0xB8 | |
73 (?,T9(B consonant "LETTER NO NU") ; 0xB9 | |
74 (?,T:(B consonant "LETTER BO BAIMAI") ; 0xBA | |
75 (?,T;(B consonant "LETTER PO PLA") ; 0xBB | |
76 (?,T<(B consonant "LETTER PHO PHUNG") ; 0xBC | |
77 (?,T=(B consonant "LETTER FO FA") ; 0xBD | |
78 (?,T>(B consonant "LETTER PHO PHAN") ; 0xBE | |
79 (?,T?(B consonant "LETTER FO FAN") ; 0xBF | |
80 (?,T@(B consonant "LETTER PHO SAMPHAO") ; 0xC0 | |
81 (?,TA(B consonant "LETTER MO MA") ; 0xC1 | |
82 (?,TB(B consonant "LETTER YO YAK") ; 0xC2 | |
83 (?,TC(B consonant "LETTER RO RUA") ; 0xC3 | |
84 (?,TD(B vowel-base "LETTER RU (Pali vowel letter)") ; 0xC4 | |
85 (?,TE(B consonant "LETTER LO LING") ; 0xC5 | |
86 (?,TF(B vowel-base "LETTER LU (Pali vowel letter)") ; 0xC6 | |
87 (?,TG(B consonant "LETTER WO WAEN") ; 0xC7 | |
88 (?,TH(B consonant "LETTER SO SALA") ; 0xC8 | |
89 (?,TI(B consonant "LETTER SO RUSI") ; 0xC9 | |
90 (?,TJ(B consonant "LETTER SO SUA") ; 0xCA | |
91 (?,TK(B consonant "LETTER HO HIP") ; 0xCB | |
92 (?,TL(B consonant "LETTER LO CHULA") ; 0xCC | |
93 (?,TM(B consonant "LETTER O ANG") ; 0xCD | |
94 (?,TN(B consonant "LETTER HO NOK HUK") ; 0xCE | |
95 (?,TO(B special "PAI YAN NOI (abbreviation)") ; 0xCF | |
96 (?,TP(B vowel-base "VOWEL SIGN SARA A") ; 0xD0 | |
97 (?,TQ(B vowel-upper "VOWEL SIGN MAI HAN-AKAT N/S-T") ; 0xD1 | |
98 (?,TR(B vowel-base "VOWEL SIGN SARA AA") ; 0xD2 | |
99 (?,TS(B vowel-base "VOWEL SIGN SARA AM") ; 0xD3 | |
100 (?,TT(B vowel-upper "VOWEL SIGN SARA I N/S-T") ; 0xD4 | |
101 (?,TU(B vowel-upper "VOWEL SIGN SARA II N/S-T") ; 0xD5 | |
102 (?,TV(B vowel-upper "VOWEL SIGN SARA UE N/S-T") ; 0xD6 | |
103 (?,TW(B vowel-upper "VOWEL SIGN SARA UEE N/S-T") ; 0xD7 | |
104 (?,TX(B vowel-lower "VOWEL SIGN SARA U N/S-B") ; 0xD8 | |
105 (?,TY(B vowel-lower "VOWEL SIGN SARA UU N/S-B") ; 0xD9 | |
106 (?,TZ(B vowel-lower "VOWEL SIGN PHINTHU N/S-B (Pali virama)") ; 0xDA | |
107 (?,T[(B invalid nil) ; 0xDA | |
108 (?,T\(B invalid nil) ; 0xDC | |
109 (?,T](B invalid nil) ; 0xDC | |
110 (?,T^(B invalid nil) ; 0xDC | |
111 (?,T_(B special "BAHT SIGN (currency symbol)") ; 0xDF | |
112 (?,T`(B vowel-base "VOWEL SIGN SARA E") ; 0xE0 | |
113 (?,Ta(B vowel-base "VOWEL SIGN SARA AE") ; 0xE1 | |
114 (?,Tb(B vowel-base "VOWEL SIGN SARA O") ; 0xE2 | |
115 (?,Tc(B vowel-base "VOWEL SIGN SARA MAI MUAN") ; 0xE3 | |
116 (?,Td(B vowel-base "VOWEL SIGN SARA MAI MALAI") ; 0xE4 | |
117 (?,Te(B vowel-base "LAK KHANG YAO") ; 0xE5 | |
118 (?,Tf(B special "MAI YAMOK (repetion)") ; 0xE6 | |
119 (?,Tg(B vowel-upper "VOWEL SIGN MAI TAI KHU N/S-T") ; 0xE7 | |
120 (?,Th(B tone "TONE MAI EK N/S-T") ; 0xE8 | |
121 (?,Ti(B tone "TONE MAI THO N/S-T") ; 0xE9 | |
122 (?,Tj(B tone "TONE MAI TRI N/S-T") ; 0xEA | |
123 (?,Tk(B tone "TONE MAI CHATTAWA N/S-T") ; 0xEB | |
124 (?,Tl(B tone "THANTHAKHAT N/S-T (cancellation mark)") ; 0xEC | |
125 (?,Tm(B tone "NIKKHAHIT N/S-T (final nasal)") ; 0xED | |
126 (?,Tn(B vowel-upper "YAMAKKAN N/S-T") ; 0xEE | |
127 (?,To(B special "FONRMAN") ; 0xEF | |
128 (?,Tp(B special "DIGIT ZERO") ; 0xF0 | |
129 (?,Tq(B special "DIGIT ONE") ; 0xF1 | |
130 (?,Tr(B special "DIGIT TWO") ; 0xF2 | |
131 (?,Ts(B special "DIGIT THREE") ; 0xF3 | |
132 (?,Tt(B special "DIGIT FOUR") ; 0xF4 | |
133 (?,Tu(B special "DIGIT FIVE") ; 0xF5 | |
134 (?,Tv(B special "DIGIT SIX") ; 0xF6 | |
135 (?,Tw(B special "DIGIT SEVEN") ; 0xF7 | |
136 (?,Tx(B special "DIGIT EIGHT") ; 0xF8 | |
137 (?,Ty(B special "DIGIT NINE") ; 0xF9 | |
138 (?,Tz(B special "ANGKHANKHU (ellipsis)") ; 0xFA | |
139 (?,T{(B special "KHOMUT (beginning of religious texts)") ; 0xFB | |
140 (?,T|(B invalid nil) ; 0xFC | |
141 (?,T}(B invalid nil) ; 0xFD | |
142 (?,T~(B invalid nil) ; 0xFE | |
143 )) | |
144 elm) | |
145 (while l | |
26892
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
146 (setq elm (car l) l (cdr l)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
147 (let ((char (car elm)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
148 (ptype (nth 1 elm))) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
149 (put-char-code-property char 'phonetic-type ptype) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
150 (cond ((eq ptype 'consonant) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
151 (modify-category-entry char ?c thai-category-table)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
152 ((memq ptype '(vowel-upper vowel-lower)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
153 (modify-category-entry char ?v thai-category-table)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
154 ((eq ptype 'tone) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
155 (modify-category-entry char ?t thai-category-table))) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
156 (put-char-code-property char 'name (nth 2 elm))))) |
23546
d180b9c69e4a
(thai-pre-write-conversion): Cancel previous
Kenichi Handa <handa@m17n.org>
parents:
23523
diff
changeset
|
157 |
17052 | 158 ;;;###autoload |
159 (defun thai-compose-region (beg end) | |
160 "Compose Thai characters in the region. | |
161 When called from a program, expects two arguments, | |
162 positions (integers or markers) specifying the region." | |
163 (interactive "r") | |
164 (save-restriction | |
165 (narrow-to-region beg end) | |
166 (goto-char (point-min)) | |
26892
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
167 (with-category-table thai-category-table |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
168 (while (re-search-forward thai-composition-pattern nil t) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
169 (compose-region (match-beginning 0) (match-end 0)))))) |
17052 | 170 |
171 ;;;###autoload | |
23994
9ae5e1af3518
(thai-compose-string): New function.
Kenichi Handa <handa@m17n.org>
parents:
23992
diff
changeset
|
172 (defun thai-compose-string (string) |
9ae5e1af3518
(thai-compose-string): New function.
Kenichi Handa <handa@m17n.org>
parents:
23992
diff
changeset
|
173 "Compose Thai characters in STRING and return the resulting string." |
26892
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
174 (with-category-table thai-category-table |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
175 (let ((idx 0)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
176 (while (setq idx (string-match thai-composition-pattern string idx)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
177 (compose-string string idx (match-end 0)) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
178 (setq idx (match-end 0))))) |
fce3871ada53
(thai-category-table): Use
Kenichi Handa <handa@m17n.org>
parents:
23994
diff
changeset
|
179 string) |
23994
9ae5e1af3518
(thai-compose-string): New function.
Kenichi Handa <handa@m17n.org>
parents:
23992
diff
changeset
|
180 |
9ae5e1af3518
(thai-compose-string): New function.
Kenichi Handa <handa@m17n.org>
parents:
23992
diff
changeset
|
181 ;;;###autoload |
17052 | 182 (defun thai-compose-buffer () |
183 "Compose Thai characters in the current buffer." | |
184 (interactive) | |
185 (thai-compose-region (point-min) (point-max))) | |
186 | |
187 ;;;###autoload | |
89302
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
188 (defun thai-composition-function (pos &optional string) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
189 (setq pos (1- pos)) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
190 (let ((pattern "[,T!(B-,TCEG(B-,TN!(B-,TCEG(B-,TN(B][,TQT(B-,TWgnX(B-,TZQT(B-,TWgnX(B-,TZ(B]?[,Th(B-,Tmh(B-,Tm(B]?")) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
191 (if string |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
192 (if (and (>= pos 0) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
193 (eq (string-match pattern string pos) pos)) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
194 (prog1 (match-end 0) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
195 (compose-string string pos (match-end 0)))) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
196 (if (>= pos (point-min)) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
197 (progn |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
198 (goto-char pos) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
199 (if (looking-at pattern) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
200 (prog1 (match-end 0) |
b1774ad77cc7
(thai-post-read-conversion): Delete this
Kenichi Handa <handa@m17n.org>
parents:
88761
diff
changeset
|
201 (compose-region pos (match-end 0))))))))) |
17052 | 202 |
203 ;; | |
18309
bd8b521f5218
Provide XXX-util instead of
Kenichi Handa <handa@m17n.org>
parents:
18203
diff
changeset
|
204 (provide 'thai-util) |
17052 | 205 |
206 ;;; thai-util.el ends here |