* url-handlers.el: Re-order file to avoid recursive load.
[emacs.git] / leim / quail / hangul.el
blob2ce55a57107a52960e931b05a6b1fe19aed9d634
1 ;;; hangul.el --- Korean Hangul input method
3 ;; Copyright (C) 2008-2012 Free Software Foundation, Inc.
5 ;; Author: Jihyun Cho <jihyun.jo@gmail.com>
6 ;; Keywords: multilingual, input method, Korean, Hangul
8 ;; This file is part of GNU Emacs.
10 ;; GNU Emacs is free software: you can redistribute it and/or modify
11 ;; it under the terms of the GNU General Public License as published by
12 ;; the Free Software Foundation, either version 3 of the License, or
13 ;; (at your option) any later version.
15 ;; GNU Emacs is distributed in the hope that it will be useful,
16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 ;; GNU General Public License for more details.
20 ;; You should have received a copy of the GNU General Public License
21 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
23 ;;; Commentary:
25 ;; This file is to implement the following hangul automata:
26 ;; - Hangul 2-Bulsik input method
27 ;; - Hangul 3-Bulsik final input method
28 ;; - Hangul 3-Bulsik 390 input method
30 ;;; Code:
32 (require 'quail)
33 (eval-when-compile (require 'cl)) ; for setf
34 (require 'hanja-util)
36 ;; Hangul double Jamo table.
37 ;; The format is an alist of JAMO-TYPE vs. DOUBLE-JAMO-TABLE.
39 ;; JAMO-TYPE is a symbol `cho' for Choseong, `jung' for Jungseong, and
40 ;; `jong' for Jongseong.
42 ;; DOUBLE-JAMO-TABLE is an alist of Jamo index vs. the vector of Jamo
43 ;; indies that can be combined with the car part.
45 ;; Jamo index is a relative index in `hangul Compatibility Jamo' area
46 ;; of the Unicode (i.e. 1 for U+3131).
48 (defconst hangul-djamo-table
49 '((cho . ((1 . [1]) ; Choseong
50 (7 . [7])
51 (18 . [18])
52 (21 . [21])
53 (24 . [24])))
54 (jung . ((39 . [31 32 51]) ; Jungseong
55 (44 . [35 36 51])
56 (49 . [51])))
57 (jong . ((1 . [1 21]) ; Jongseong
58 (4 . [24 30])
59 (9 . [1 17 18 21 28 29 30])
60 (18 . [18 21])
61 (21 . [21])))))
63 ;; Hangul 2-Bulsik keymap.
64 ;; It converts an ASCII code A-Z, a-z, to the corresponding hangul
65 ;; Jamo index.
67 (defconst hangul2-keymap
68 [17 48 26 23 7 9 30 39 33 35 31 51 49 44 32 36 18 1 4 21 37 29 24 28 43 27])
70 ;; Hangul 3-Bulsik final keymap. 3-Bulsik use full keyboard layout.
71 ;; Therefore, we must map all printable ASCII codes (`!' to `~')
72 ;; to Hangul 3-Bulsik codes.
73 ;; Other parts are the same as `hangul2-keymap'.
74 (defconst hangul3-keymap
75 [2 183 24 15 14 8220 120 39 126 8221 43 44 41 46 74 119 30 22 18 78 83
76 68 73 85 79 52 110 44 62 46 33 10 7 63 27 12 5 11 69 48 55 49 50 51
77 34 45 56 57 29 16 6 13 54 3 28 20 53 26 40 58 60 61 59 42 23 79 71
78 86 72 66 84 96 109 115 93 116 122 113 118 121 21 67 4 70 99 74 9 1
79 101 17 37 92 47 8251])
81 ;; Hangul 3-Bulsik 390 keymap.
82 ;; The role is the same as `hangul3-keymap'.
83 (defconst hangul390-keymap
84 [24 34 35 36 37 38 120 40 41 42 43 44 45 46 73 119 30 22 18 77 82 67 72
85 84 78 58 110 50 61 51 63 64 7 33 11 10 27 2 47 39 56 52 53 54 49 48
86 57 62 29 68 6 59 55 16 28 20 60 26 91 92 93 94 95 96 23 78 70 85 71
87 65 83 90 109 115 87 116 122 113 118 121 21 66 4 69 99 73 9 1 101 17
88 123 124 125 126])
90 (defvar hangul-im-keymap
91 (let ((map (make-sparse-keymap)))
92 (define-key map "\d" 'hangul-delete-backward-char)
93 (define-key map [f9] 'hangul-to-hanja-conversion)
94 (define-key map [Hangul_Hanja] 'hangul-to-hanja-conversion)
95 map)
96 "Keymap for Hangul method. It is used by all Hangul input methods.")
98 ;; Current input character buffer. Store separated hangul character.
99 ;; The first and second are Choseong position.
100 ;; The third and forth are Jungseong position.
101 ;; The fifth and sixth are Jongseong position.
102 ;; The second, forth and sixth are double Jamo position.
103 (defvar hangul-queue
104 (make-vector 6 0))
106 (defsubst notzerop (number)
107 (not (zerop number)))
109 (defsubst alphabetp (char)
110 (or (and (>= char ?A) (<= char ?Z))
111 (and (>= char ?a) (<= char ?z))))
113 (defun hangul-character (cho jung jong)
114 "Convert CHO, JUNG, JONG to the precomposed `Hangul Syllables' character.
115 CHO, JUNG, JONG are relative indices in `Hangul Compatibility Jamo' of Unicode.
116 Return a zero-length string if the conversion fails."
118 (decode-char
119 'ucs
120 (if (and (/= cho 0) (/= jung 0))
121 (+ #xac00
122 (* 588
123 (- cho
124 (cond ((< cho 3) 1)
125 ((< cho 5) 2)
126 ((< cho 10) 4)
127 ((< cho 20) 11)
128 (t 12))))
129 (* 28 (- jung 31))
130 (- jong
131 (cond ((< jong 8) 0)
132 ((< jong 19) 1)
133 ((< jong 25) 2)
134 (t 3))))
135 (+ #x3130
136 (cond ((/= cho 0) cho)
137 ((/= jung 0) jung)
138 ((/= jong 0) jong)))))
139 ""))
141 (defun hangul-insert-character (&rest queues)
142 "Insert characters generated from QUEUES.
143 Each queue has the same form as `hangul-queue'.
144 Setup `quail-overlay' to the last character."
145 (if (and mark-active transient-mark-mode)
146 (progn
147 (delete-region (region-beginning) (region-end))
148 (deactivate-mark)))
149 (quail-delete-region)
150 (let ((first (car queues)))
151 (insert
152 (hangul-character
153 (+ (aref first 0) (hangul-djamo 'cho (aref first 0) (aref first 1)))
154 (+ (aref first 2) (hangul-djamo 'jung (aref first 2) (aref first 3)))
155 (+ (aref first 4) (hangul-djamo 'jong (aref first 4) (aref first 5))))))
156 (move-overlay quail-overlay (overlay-start quail-overlay) (point))
157 (dolist (queue (cdr queues))
158 (insert
159 (hangul-character
160 (+ (aref queue 0) (hangul-djamo 'cho (aref queue 0) (aref queue 1)))
161 (+ (aref queue 2) (hangul-djamo 'jung (aref queue 2) (aref queue 3)))
162 (+ (aref queue 4) (hangul-djamo 'jong (aref queue 4) (aref queue 5)))))
163 (move-overlay quail-overlay (1+ (overlay-start quail-overlay)) (point))))
165 (defun hangul-djamo (jamo char1 char2)
166 "Return the double Jamo index calculated from the arguments.
167 JAMO is a type of Hangul Jamo; `cho', `jung', or `jong'.
168 CHAR1 and CHAR2 are Hangul Jamo indices.
169 Return nil if CHAR1 and CHAR2 can not be combined."
170 (let* ((jamo (cdr (assoc jamo hangul-djamo-table)))
171 (char1 (cdr (assoc char1 jamo))))
172 (if char1
173 (let ((i (length char1)))
174 (or (catch 'found
175 (while (> i 0)
176 (if (= char2 (aref char1 (1- i)))
177 (throw 'found i))
178 (setf i (1- i))))
180 0)))
182 (defsubst hangul2-input-method-jaum (char)
183 "Store Hangul Jamo indice CHAR in `hangul-queue'.
184 It is a Hangul 2-Bulsik Jaum.
185 This function processes a Hangul 2-Bulsik Jaum.
186 The Hangul 2-Bulsik is composed of a Jaum and a Moum.
187 The Jaum can be located in a Choseong position and a Jongseong position.
188 Unless the function inserts CHAR to `hangul-queue',
189 commit current `hangul-queue' and then set a new `hangul-queue',
190 and insert CHAR to new `hangul-queue'."
191 (if (cond ((zerop (aref hangul-queue 0))
192 (aset hangul-queue 0 char))
193 ((and (zerop (aref hangul-queue 1))
194 (zerop (aref hangul-queue 2))
195 (notzerop (hangul-djamo 'cho (aref hangul-queue 0) char)))
196 (aset hangul-queue 1 char))
197 ((and (zerop (aref hangul-queue 4))
198 (notzerop (aref hangul-queue 2))
199 (/= char 8)
200 (/= char 19)
201 (/= char 25)
202 (numberp
203 (hangul-character
204 (+ (aref hangul-queue 0)
205 (hangul-djamo
206 'cho
207 (aref hangul-queue 0)
208 (aref hangul-queue 1)))
209 (+ (aref hangul-queue 2)
210 (hangul-djamo
211 'jung
212 (aref hangul-queue 2)
213 (aref hangul-queue 3)))
214 char)))
215 (aset hangul-queue 4 char))
216 ((and (zerop (aref hangul-queue 5))
217 (notzerop (hangul-djamo 'jong (aref hangul-queue 4) char))
218 (numberp
219 (hangul-character
220 (+ (aref hangul-queue 0)
221 (hangul-djamo
222 'cho
223 (aref hangul-queue 0)
224 (aref hangul-queue 1)))
225 (+ (aref hangul-queue 2)
226 (hangul-djamo
227 'jung
228 (aref hangul-queue 2)
229 (aref hangul-queue 3)))
230 (+ (aref hangul-queue 4)
231 (hangul-djamo
232 'jong
233 (aref hangul-queue 4)
234 char)))))
235 (aset hangul-queue 5 char)))
236 (hangul-insert-character hangul-queue)
237 (hangul-insert-character hangul-queue
238 (setq hangul-queue (vector char 0 0 0 0 0)))))
240 (defsubst hangul2-input-method-moum (char)
241 "Store Hangul Jamo indice CHAR in `hangul-queue'.
242 It is a Hangul 2-Bulsik Moum.
243 This function processes a Hangul 2-Bulsik Moum.
244 The Moum can be located in a Jungseong position.
245 Other parts are the same as a `hangul2-input-method-jaum'."
246 (if (cond ((zerop (aref hangul-queue 2))
247 (aset hangul-queue 2 char))
248 ((and (zerop (aref hangul-queue 3))
249 (zerop (aref hangul-queue 4))
250 (notzerop (hangul-djamo 'jung (aref hangul-queue 2) char)))
251 (aset hangul-queue 3 char)))
252 (hangul-insert-character hangul-queue)
253 (let ((next-char (vector 0 0 char 0 0 0)))
254 (cond ((notzerop (aref hangul-queue 5))
255 (aset next-char 0 (aref hangul-queue 5))
256 (aset hangul-queue 5 0))
257 ((notzerop (aref hangul-queue 4))
258 (aset next-char 0 (aref hangul-queue 4))
259 (aset hangul-queue 4 0)))
260 (hangul-insert-character hangul-queue
261 (setq hangul-queue next-char)))))
263 (defsubst hangul3-input-method-cho (char)
264 "Store Hangul Jamo indice CHAR in `hangul-queue'.
265 It is a Hangul 3-Bulsik Choseong.
266 This function processes a Hangul 3-Bulsik Choseong.
267 The Hangul 3-Bulsik is composed of a Choseong, a Jungseong and a Jongseong.
268 The Choseong can be located in a Choseong position.
269 Other parts are the same as a `hangul2-input-method-jaum'."
270 (if (cond ((and (zerop (aref hangul-queue 0))
271 (zerop (aref hangul-queue 4)))
272 (aset hangul-queue 0 char))
273 ((and (zerop (aref hangul-queue 1))
274 (zerop (aref hangul-queue 2))
275 (notzerop (hangul-djamo 'cho (aref hangul-queue 0) char)))
276 (aset hangul-queue 1 char)))
277 (hangul-insert-character hangul-queue)
278 (hangul-insert-character hangul-queue
279 (setq hangul-queue (vector char 0 0 0 0 0)))))
281 (defsubst hangul3-input-method-jung (char)
282 "Store Hangul Jamo indice CHAR in `hangul-queue'.
283 It is a Hangul 3-Bulsik Jungseong.
284 This function processes a Hangul 3-Bulsik Jungseong.
285 The Jungseong can be located in a Jungseong position.
286 Other parts are the same as a `hangul3-input-method-cho'."
287 (if (cond ((and (zerop (aref hangul-queue 2))
288 (zerop (aref hangul-queue 4)))
289 (aset hangul-queue 2 char))
290 ((and (zerop (aref hangul-queue 3))
291 (notzerop (hangul-djamo 'jung (aref hangul-queue 2) char)))
292 (aset hangul-queue 3 char)))
293 (hangul-insert-character hangul-queue)
294 (hangul-insert-character hangul-queue
295 (setq hangul-queue (vector 0 0 char 0 0 0)))))
297 (defsubst hangul3-input-method-jong (char)
298 "Store Hangul Jamo indice CHAR in `hangul-queue'.
299 It is a Hangul 3-Bulsik Jongseong.
300 This function processes a Hangul 3-Bulsik Jongseong.
301 The Jongseong can be located in a Jongseong position.
302 Other parts are the same as a `hangul3-input-method-cho'."
303 (if (cond ((and (zerop (aref hangul-queue 4))
304 (notzerop (aref hangul-queue 0))
305 (notzerop (aref hangul-queue 2))
306 (numberp
307 (hangul-character
308 (+ (aref hangul-queue 0)
309 (hangul-djamo
310 'cho
311 (aref hangul-queue 0)
312 (aref hangul-queue 1)))
313 (+ (aref hangul-queue 2)
314 (hangul-djamo
315 'jung
316 (aref hangul-queue 2)
317 (aref hangul-queue 3)))
318 char)))
319 (aset hangul-queue 4 char))
320 ((and (zerop (aref hangul-queue 5))
321 (notzerop (hangul-djamo 'jong (aref hangul-queue 4) char))
322 (numberp
323 (hangul-character
324 (+ (aref hangul-queue 0)
325 (hangul-djamo
326 'cho
327 (aref hangul-queue 0)
328 (aref hangul-queue 1)))
329 (+ (aref hangul-queue 2)
330 (hangul-djamo
331 'jung
332 (aref hangul-queue 2)
333 (aref hangul-queue 3)))
334 (+ (aref hangul-queue 4)
335 (hangul-djamo
336 'jong
337 (aref hangul-queue 4)
338 char)))))
339 (aset hangul-queue 5 char)))
340 (hangul-insert-character hangul-queue)
341 (if (zerop (apply '+ (append hangul-queue nil)))
342 (hangul-insert-character (setq hangul-queue (vector 0 0 0 0 char 0)))
343 (hangul-insert-character hangul-queue
344 (setq hangul-queue (vector 0 0 0 0 char 0))))))
346 (defun hangul-delete-backward-char ()
347 "Delete the previous hangul character by Jaso units."
348 (interactive)
349 (let ((i 5))
350 (while (and (> i 0) (zerop (aref hangul-queue i)))
351 (setq i (1- i)))
352 (aset hangul-queue i 0))
353 (if (notzerop (apply '+ (append hangul-queue nil)))
354 (hangul-insert-character hangul-queue)
355 (delete-backward-char 1)))
357 (defun hangul-to-hanja-conversion ()
358 "Convert the previous hangul character to the corresponding hanja character.
359 When a Korean input method is off, convert the following hangul character."
360 (interactive)
361 (let ((echo-keystrokes 0)
362 delete-func
363 hanja-character)
364 (if (and (overlayp quail-overlay) (overlay-start quail-overlay))
365 (progn
366 (setq hanja-character (hangul-to-hanja-char (preceding-char)))
367 (setq delete-func (lambda () (delete-backward-char 1))))
368 (setq hanja-character (hangul-to-hanja-char (following-char)))
369 (setq delete-func (lambda () (delete-char 1))))
370 (when hanja-character
371 (funcall delete-func)
372 (insert hanja-character)
373 (setq hangul-queue (make-vector 6 0))
374 (if (and (overlayp quail-overlay) (overlay-start quail-overlay))
375 (move-overlay quail-overlay (point) (point))))))
377 ;; Support function for `hangul2-input-method'. Actually, this
378 ;; function handles the Hangul 2-Bulsik. KEY is an entered key code
379 ;; used for looking up `hangul2-keymap'."
380 (defun hangul2-input-method-internal (key)
381 (let ((char (+ (aref hangul2-keymap (1- (% key 32)))
382 (cond ((or (= key ?O) (= key ?P)) 2)
383 ((or (= key ?E) (= key ?Q) (= key ?R)
384 (= key ?T) (= key ?W)) 1)
385 (t 0)))))
386 (if (< char 31)
387 (hangul2-input-method-jaum char)
388 (hangul2-input-method-moum char))))
390 (defun hangul2-input-method (key)
391 "2-Bulsik input method."
392 (if (or buffer-read-only (not (alphabetp key)))
393 (list key)
394 (quail-setup-overlays nil)
395 (let ((input-method-function nil)
396 (echo-keystrokes 0)
397 (help-char nil))
398 (setq hangul-queue (make-vector 6 0))
399 (hangul2-input-method-internal key)
400 (unwind-protect
401 (catch 'exit-input-loop
402 (while t
403 (let* ((seq (read-key-sequence nil))
404 (cmd (lookup-key hangul-im-keymap seq))
405 key)
406 (cond ((and (stringp seq)
407 (= 1 (length seq))
408 (setq key (aref seq 0))
409 (alphabetp key))
410 (hangul2-input-method-internal key))
411 ((commandp cmd)
412 (call-interactively cmd))
414 (setq unread-command-events (listify-key-sequence seq))
415 (throw 'exit-input-loop nil))))))
416 (quail-delete-overlays)))))
418 ;; Support function for `hangul3-input-method'. Actually, this
419 ;; function handles the Hangul 3-Bulsik final. KEY is an entered key
420 ;; code used for looking up `hangul3-keymap'."
421 (defun hangul3-input-method-internal (key)
422 (let ((char (aref hangul3-keymap (- key 33))))
423 (cond ((and (> char 92) (< char 123))
424 (hangul3-input-method-cho (- char 92)))
425 ((and (> char 65) (< char 87))
426 (hangul3-input-method-jung (- char 35)))
427 ((< char 31)
428 (hangul3-input-method-jong char))
430 (setq hangul-queue (make-vector 6 0))
431 (insert (decode-char 'ucs char))
432 (move-overlay quail-overlay (point) (point))))))
434 (defun hangul3-input-method (key)
435 "3-Bulsik final input method."
436 (if (or buffer-read-only (< key 33) (>= key 127))
437 (list key)
438 (quail-setup-overlays nil)
439 (let ((input-method-function nil)
440 (echo-keystrokes 0)
441 (help-char nil))
442 (setq hangul-queue (make-vector 6 0))
443 (hangul3-input-method-internal key)
444 (unwind-protect
445 (catch 'exit-input-loop
446 (while t
447 (let* ((seq (read-key-sequence nil))
448 (cmd (lookup-key hangul-im-keymap seq))
449 key)
450 (cond ((and (stringp seq)
451 (= 1 (length seq))
452 (setq key (aref seq 0))
453 (and (>= key 33) (< key 127)))
454 (hangul3-input-method-internal key))
455 ((commandp cmd)
456 (call-interactively cmd))
458 (setq unread-command-events (listify-key-sequence seq))
459 (throw 'exit-input-loop nil))))))
460 (quail-delete-overlays)))))
462 ;; Support function for `hangul390-input-method'. Actually, this
463 ;; function handles the Hangul 3-Bulsik 390. KEY is an entered key
464 ;; code used for looking up `hangul390-keymap'."
465 (defun hangul390-input-method-internal (key)
466 (let ((char (aref hangul390-keymap (- key 33))))
467 (cond ((or (and (> char 86) (< char 91))
468 (and (> char 96) (< char 123)))
469 (hangul3-input-method-cho (- char (if (< char 97) 86 92))))
470 ((and (> char 64) (< char 86))
471 (hangul3-input-method-jung (- char 34)))
472 ((< char 31)
473 (hangul3-input-method-jong char))
475 (setq hangul-queue (make-vector 6 0))
476 (insert (decode-char 'ucs char))
477 (move-overlay quail-overlay (point) (point))))))
479 (defun hangul390-input-method (key)
480 "3-Bulsik 390 input method."
481 (if (or buffer-read-only (< key 33) (>= key 127))
482 (list key)
483 (quail-setup-overlays nil)
484 (let ((input-method-function nil)
485 (echo-keystrokes 0)
486 (help-char nil))
487 (setq hangul-queue (make-vector 6 0))
488 (hangul390-input-method-internal key)
489 (unwind-protect
490 (catch 'exit-input-loop
491 (while t
492 (let* ((seq (read-key-sequence nil))
493 (cmd (lookup-key hangul-im-keymap seq))
494 key)
495 (cond ((and (stringp seq)
496 (= 1 (length seq))
497 (setq key (aref seq 0))
498 (and (>= key 33) (< key 127)))
499 (hangul390-input-method-internal key))
500 ((commandp cmd)
501 (call-interactively cmd))
503 (setq unread-command-events (listify-key-sequence seq))
504 (throw 'exit-input-loop nil))))))
505 (quail-delete-overlays)))))
507 ;; Text shown by describe-input-method. Set to a proper text by
508 ;; hangul-input-method-activate.
509 (defvar hangul-input-method-help-text nil)
510 (make-variable-buffer-local 'hangul-input-method-help-text)
512 (defun hangul-input-method-activate (input-method func help-text &rest args)
513 "Activate Hangul input method INPUT-METHOD.
514 FUNC is a function to handle input key.
515 HELP-TEXT is a text set in `hangul-input-method-help-text'."
516 (setq inactivate-current-input-method-function 'hangul-input-method-inactivate
517 describe-current-input-method-function 'hangul-input-method-help
518 hangul-input-method-help-text help-text)
519 (quail-delete-overlays)
520 (if (eq (selected-window) (minibuffer-window))
521 (add-hook 'minibuffer-exit-hook 'quail-exit-from-minibuffer))
522 (set (make-local-variable 'input-method-function) func))
524 (defun hangul-input-method-inactivate ()
525 "Inactivate the current Hangul input method."
526 (interactive)
527 (unwind-protect
528 (progn
529 (quail-hide-guidance)
530 (quail-delete-overlays)
531 (setq describe-current-input-method-function nil))
532 (kill-local-variable 'input-method-function)))
534 (defun hangul-input-method-help ()
535 "Describe the current Hangul input method."
536 (interactive)
537 (with-output-to-temp-buffer "*Help*"
538 (princ hangul-input-method-help-text)))
540 (provide 'hangul)
542 ;;; hangul.el ends here