Sync to HEAD
[bpt/emacs.git] / lisp / international / iso-acc.el
1 ;;; iso-acc.el --- minor mode providing electric accent keys
2
3 ;; Copyright (C) 1993, 1994, 1996, 2001 Free Software Foundation, Inc.
4
5 ;; Author: Johan Vromans
6 ;; Maintainer: FSF
7 ;; Keywords: i18n
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 ;; Function `iso-accents-mode' activates a minor mode in which
29 ;; typewriter "dead keys" are emulated. The purpose of this emulation
30 ;; is to provide a simple means for inserting accented characters
31 ;; according to the ISO-8859-1...3 character sets.
32 ;;
33 ;; In `iso-accents-mode', pseudo accent characters are used to
34 ;; introduce accented keys. The pseudo-accent characters are:
35 ;;
36 ;; ' (minute) -> actue accent
37 ;; ` (backtick) -> grave accent
38 ;; " (second) -> diaeresis
39 ;; ^ (caret) -> circumflex
40 ;; ~ (tilde) -> tilde over the character
41 ;; / (slash) -> slash through the character.
42 ;; Also: /A is A-with-ring and /E is AE ligature.
43 ;; These two are enabled only if you set iso-accents-enable
44 ;; to include them:
45 ;; . (period) -> dot over the character (some languages only)
46 ;; , (cedilla) -> cedilla under the character (some languages only)
47 ;;
48 ;; The action taken depends on the key that follows the pseudo accent.
49 ;; In general:
50 ;;
51 ;; pseudo-accent + appropriate letter -> accented letter
52 ;; pseudo-accent + space -> pseudo-accent (except comma and period)
53 ;; pseudo-accent + pseudo-accent -> accent (if available)
54 ;; pseudo-accent + other -> pseudo-accent + other
55 ;;
56 ;; If the pseudo-accent is followed by anything else than a
57 ;; self-insert-command, the dead-key code is terminated, the
58 ;; pseudo-accent inserted 'as is' and the bell is rung to signal this.
59 ;;
60 ;; Function `iso-accents-mode' can be used to enable the iso accents
61 ;; minor mode, or disable it.
62
63 ;; If you want only some of these characters to serve as accents,
64 ;; add a language to `iso-languages' which specifies the accent characters
65 ;; that you want, then select the language with `iso-accents-customize'.
66 \f
67 ;;; Code:
68
69 (provide 'iso-acc)
70
71 (defgroup iso-acc nil
72 "Minor mode providing electric accent keys."
73 :prefix "iso-accents-"
74 :group 'i18n)
75
76 (defcustom iso-accents-insert-offset nonascii-insert-offset
77 "*Offset added by ISO Accents mode to character codes 0200 and above."
78 :type 'integer
79 :group 'iso-acc)
80
81 (defvar iso-languages
82 '(("catalan"
83 ;; Note this includes some extra characters used in Spanish,
84 ;; on the idea that someone who uses Catalan is likely to use Spanish
85 ;; as well.
86 (?' (?A . ?\301) (?E . ?\311) (?I . ?\315) (?O . ?\323) (?U . ?\332)
87 (?a . ?\341) (?e . ?\351) (?i . ?\355) (?o . ?\363) (?u . ?\372)
88 (?\ . ?'))
89 (?` (?A . ?\300) (?E . ?\310) (?O . ?\322)
90 (?a . ?\340) (?e . ?\350) (?o . ?\362)
91 (?\ . ?`))
92 (?\" (?I . ?\317) (?U . ?\334) (?i . ?\357) (?u . ?\374)
93 (?\ . ?\"))
94 (?~ (?C . ?\307) (?N . ?\321) (?c . ?\347) (?n . ?\361)
95 (?> . ?\273) (?< . ?\253) (?! . ?\241) (?? . ?\277)
96 (?\ . ?\~)))
97
98 ("esperanto"
99 (?^ (?H . ?\246) (?J . ?\254) (?h . ?\266) (?j . ?\274) (?C . ?\306)
100 (?G . ?\330) (?S . ?\336) (?c . ?\346) (?g . ?\370) (?s . ?\376)
101 (?^ . ?^) (?\ . ?^))
102 (?~ (?U . ?\335) (?u . ?\375) (?\ . ?~)))
103
104 ("french"
105 (?' (?E . ?\311) (?C . ?\307) (?e . ?\351) (?c . ?\347)
106 (?\ . ?'))
107 (?` (?A . ?\300) (?E . ?\310) (?U . ?\331)
108 (?a . ?\340) (?e . ?\350) (?u . ?\371)
109 (?\ . ?`))
110 (?^ (?A . ?\302) (?E . ?\312) (?I . ?\316) (?O . ?\324) (?U . ?\333)
111 (?a . ?\342) (?e . ?\352) (?i . ?\356) (?o . ?\364) (?u . ?\373)
112 (?\ . ?^))
113 (?\" (?E . ?\313) (?I . ?\317)
114 (?e . ?\353) (?i . ?\357)
115 (?\ . ?\"))
116 (?~ (?< . ?\253) (?> . ?\273) (?C . ?\307) (?c . ?\347)
117 (?\ . ?~))
118 (?, (?C . ?\307) (?c . ?\347) (?\ . ?\,)))
119
120 ("german"
121 (?\" (?A . ?\304) (?O . ?\326) (?U . ?\334)
122 (?a . ?\344) (?o . ?\366) (?u . ?\374) (?s . ?\337) (?\ . ?\")))
123
124 ("irish"
125 (?' (?A . ?\301) (?E . ?\311) (?I . ?\315) (?O . ?\323) (?U . ?\332)
126 (?a . ?\341) (?e . ?\351) (?i . ?\355) (?o . ?\363) (?u . ?\372)
127 (?\ . ?')))
128
129 ("portuguese"
130 (?' (?A . ?\301) (?E . ?\311) (?I . ?\315) (?O . ?\323) (?U . ?\332)
131 (?C . ?\307) (?a . ?\341) (?e . ?\351) (?i . ?\355) (?o . ?\363)
132 (?u . ?\372) (?c . ?\347)
133 (?\ . ?'))
134 (?` (?A . ?\300) (?a . ?\340)
135 (?\ . ?`))
136 (?^ (?A . ?\302) (?E . ?\312) (?O . ?\324)
137 (?a . ?\342) (?e . ?\352) (?o . ?\364)
138 (?\ . ?^))
139 (?\" (?U . ?\334) (?u . ?\374)
140 (?\ . ?\"))
141 (?~ (?A . ?\303) (?O . ?\325) (?a . ?\343) (?o . ?\365)
142 (?C . ?\307) (?N . ?\321) (?c . ?\347) (?n . ?\361)
143 (?\ . ?~))
144 (?, (?c . ?\347) (?C . ?\307) (?, . ?,)))
145
146 ("spanish"
147 (?' (?A . ?\301) (?E . ?\311) (?I . ?\315) (?O . ?\323) (?U . ?\332)
148 (?a . ?\341) (?e . ?\351) (?i . ?\355) (?o . ?\363) (?u . ?\372)
149 (?\ . ?'))
150 (?\" (?U . ?\334) (?u . ?\374) (?\ . ?\"))
151 (?\~ (?N . ?\321) (?n . ?\361) (?> . ?\273) (?< . ?\253) (?! . ?\241)
152 (?? . ?\277) (?\ . ?\~)))
153
154 ("latin-1"
155 (?' (?A . ?\301) (?E . ?\311) (?I . ?\315) (?O . ?\323) (?U . ?\332)
156 (?Y . ?\335) (?a . ?\341) (?e . ?\351) (?i . ?\355) (?o . ?\363)
157 (?u . ?\372) (?y . ?\375) (?' . ?\264)
158 (?\ . ?'))
159 (?` (?A . ?\300) (?E . ?\310) (?I . ?\314) (?O . ?\322) (?U . ?\331)
160 (?a . ?\340) (?e . ?\350) (?i . ?\354) (?o . ?\362) (?u . ?\371)
161 (?` . ?`) (?\ . ?`))
162 (?^ (?A . ?\302) (?E . ?\312) (?I . ?\316) (?O . ?\324) (?U . ?\333)
163 (?a . ?\342) (?e . ?\352) (?i . ?\356) (?o . ?\364) (?u . ?\373)
164 (?^ . ?^) (?\ . ?^))
165 (?\" (?A . ?\304) (?E . ?\313) (?I . ?\317) (?O . ?\326) (?U . ?\334)
166 (?a . ?\344) (?e . ?\353) (?i . ?\357) (?o . ?\366) (?s . ?\337)
167 (?u . ?\374) (?y . ?\377)
168 (?\" . ?\250) (?\ . ?\"))
169 (?~ (?A . ?\303) (?C . ?\307) (?D . ?\320) (?N . ?\321) (?O . ?\325)
170 (?T . ?\336) (?a . ?\343) (?c . ?\347) (?d . ?\360) (?n . ?\361)
171 (?o . ?\365) (?t . ?\376)
172 (?> . ?\273) (?< . ?\253) (?! . ?\241) (?? . ?\277)
173 (?\~ . ?\270) (?\ . ?~))
174 (?/ (?A . ?\305) (?E . ?\306) (?O . ?\330) (?a . ?\345) (?e . ?\346)
175 (?o . ?\370)
176 (?/ . ?\260) (?\ . ?/)))
177
178 ("latin-2" latin-iso8859-2
179 (?' (?A . ?\301) (?C . ?\306) (?D . ?\320) (?E . ?\311) (?I . ?\315)
180 (?L . ?\305) (?N . ?\321) (?O . ?\323) (?R . ?\300) (?S . ?\246)
181 (?U . ?\332) (?Y . ?\335) (?Z . ?\254)
182 (?a . ?\341) (?c . ?\346) (?d . ?\360) (?e . ?\351) (?i . ?\355)
183 (?l . ?\345) (?n . ?\361) (?o . ?\363) (?r . ?\340) (?s . ?\266)
184 (?u . ?\372) (?y . ?\375) (?z . ?\274)
185 (?' . ?\264) (?\ . ?'))
186 (?` (?A . ?\241) (?C . ?\307) (?E . ?\312) (?L . ?\243) (?S . ?\252)
187 (?T . ?\336) (?Z . ?\257)
188 (?a . ?\261) (?l . ?\263) (?c . ?\347) (?e . ?\352) (?s . ?\272)
189 (?t . ?\376) (?z . ?\277)
190 (?` . ?\252)
191 (?. . ?\377) (?\ . ?`))
192 (?^ (?A . ?\302) (?I . ?\316) (?O . ?\324)
193 (?a . ?\342) (?i . ?\356) (?o . ?\364)
194 (?^ . ?^) ; no special code?
195 (?\ . ?^))
196 (?\" (?A . ?\304) (?E . ?\313) (?O . ?\326) (?U . ?\334)
197 (?a . ?\344) (?e . ?\353) (?o . ?\366) (?s . ?\337) (?u . ?\374)
198 (?\" . ?\250)
199 (?\ . ?\"))
200 (?~ (?A . ?\303) (?C . ?\310) (?D . ?\317) (?L . ?\245) (?N . ?\322)
201 (?O . ?\325) (?R . ?\330) (?S . ?\251) (?T . ?\253) (?U . ?\333)
202 (?Z . ?\256)
203 (?a . ?\343) (?c . ?\350) (?d . ?\357) (?l . ?\265) (?n . ?\362)
204 (?o . ?\365) (?r . ?\370) (?s . ?\271) (?t . ?\273) (?u . ?\373)
205 (?z . ?\276)
206 (?v . ?\242) ; v accent
207 (?\~ . ?\242) ; v accent
208 (?\. . ?\270) ; cedilla accent
209 (?\ . ?~)))
210
211 ("latin-3" latin-iso8859-3
212 (?' (?A . ?\301) (?E . ?\311) (?I . ?\315) (?O . ?\323) (?U . ?\332)
213 (?a . ?\341) (?e . ?\351) (?i . ?\355) (?o . ?\363) (?u . ?\372)
214 (?' . ?\264) (?\ . ?'))
215 (?` (?A . ?\300) (?E . ?\310) (?I . ?\314) (?O . ?\322) (?U . ?\331)
216 (?a . ?\340) (?e . ?\350) (?i . ?\354) (?o . ?\362) (?u . ?\371)
217 (?` . ?`) (?\ . ?`))
218 (?^ (?A . ?\302) (?C . ?\306) (?E . ?\312) (?G . ?\330) (?H . ?\246)
219 (?I . ?\316) (?J . ?\254) (?O . ?\324) (?S . ?\336) (?U . ?\333)
220 (?a . ?\342) (?c . ?\346) (?e . ?\352) (?g . ?\370) (?h . ?\266)
221 (?i . ?\356) (?j . ?\274) (?o . ?\364) (?s . ?\376) (?u . ?\373)
222 (?^ . ?^) (?\ . ?^))
223 (?\" (?A . ?\304) (?E . ?\313) (?I . ?\317) (?O . ?\326) (?U . ?\334)
224 (?a . ?\344) (?e . ?\353) (?i . ?\357) (?o . ?\366) (?u . ?\374)
225 (?s . ?\337)
226 (?\" . ?\250) (?\ . ?\"))
227 (?~ (?A . ?\303) (?C . ?\307) (?D . ?\320) (?N . ?\321) (?O . ?\325)
228 (?a . ?\343) (?c . ?\347) (?d . ?\360) (?n . ?\361) (?o . ?\365)
229 (?$ . ?\245) (?S . ?\252) (?s . ?\272) (?G . ?\253) (?g . ?\273)
230 (?U . ?\335) (?u . ?\375) (?` . ?\242)
231 (?~ . ?\270) (?\ . ?~))
232 (?/ (?C . ?\305) (?G . ?\325) (?H . ?\241) (?I . ?\251) (?Z . ?\257)
233 (?c . ?\345) (?g . ?\365) (?h . ?\261) (?i . ?\271) (?z . ?\277)
234 (?r . ?\256)
235 (?. . ?\377) (?# . ?\243) (?$ . ?\244)
236 (?/ . ?\260) (?\ . ?/))
237 (?. (?C . ?\305) (?G . ?\325) (?I . ?\251) (?Z . ?\257)
238 (?c . ?\345) (?g . ?\365) (?z . ?\277))))
239 "List of language-specific customizations for the ISO Accents mode.
240
241 Each element of the list is of the form
242
243 (LANGUAGE [CHARSET]
244 (PSEUDO-ACCENT MAPPINGS)
245 (PSEUDO-ACCENT MAPPINGS)
246 ...)
247
248 LANGUAGE is a string naming the language.
249 CHARSET (which may be omitted) is the symbol name
250 of the character set used in this language.
251 If CHARSET is omitted, latin-iso8859-1 is the default.
252 PSEUDO-ACCENT is a char specifying an accent key.
253 MAPPINGS are cons cells of the form (CHAR . ISO-CHAR).
254
255 The net effect is that the key sequence PSEUDO-ACCENT CHAR is mapped
256 to ISO-CHAR on input.")
257
258 (defvar iso-language nil
259 "Language for which ISO Accents mode is currently customized.
260 Change it with the `iso-accents-customize' function.")
261
262 (defvar iso-accents-list nil
263 "Association list for ISO accent combinations, for the chosen language.")
264
265 (defcustom iso-accents-mode nil
266 "*Non-nil enables ISO Accents mode.
267 Setting this variable makes it local to the current buffer.
268 See the function `iso-accents-mode'."
269 :type 'boolean
270 :group 'iso-acc)
271 (make-variable-buffer-local 'iso-accents-mode)
272
273 (defcustom iso-accents-enable '(?' ?` ?^ ?\" ?~ ?/)
274 "*List of accent keys that become prefixes in ISO Accents mode.
275 The default is (?' ?` ?^ ?\" ?~ ?/), which contains all the supported
276 accent keys. If you set this variable to a list in which some of those
277 characters are missing, the missing ones do not act as accents.
278
279 Note that if you specify a language with `iso-accents-customize',
280 that can also turn off certain prefixes (whichever ones are not needed in
281 the language you choose)."
282 :type '(repeat character)
283 :group 'iso-acc)
284
285 (defun iso-accents-accent-key (prompt)
286 "Modify the following character by adding an accent to it."
287 ;; Pick up the accent character.
288 (if (and iso-accents-mode
289 (memq last-input-char iso-accents-enable))
290 (iso-accents-compose prompt)
291 (vector last-input-char)))
292
293 (defun iso-accents-compose (prompt)
294 (let* ((first-char last-input-char)
295 (list (assq first-char iso-accents-list))
296 ;; Wait for the second key and look up the combination.
297 (second-char (if (or prompt
298 (not (eq (key-binding "a")
299 'self-insert-command))
300 ;; Not at start of a key sequence.
301 (> (length (this-single-command-keys)) 1)
302 ;; Called from anything but the command loop.
303 this-command)
304 (progn
305 (message "%s%c"
306 (or prompt "Compose with ")
307 first-char)
308 (read-event))
309 (insert first-char)
310 (prog1 (read-event)
311 (delete-region (1- (point)) (point)))))
312 (entry (cdr (assq second-char list))))
313 (if entry
314 ;; Found it: return the mapped char
315 (vector
316 (if (and enable-multibyte-characters
317 (>= entry ?\200))
318 (+ iso-accents-insert-offset entry)
319 entry))
320 ;; Otherwise, advance and schedule the second key for execution.
321 (push second-char unread-command-events)
322 (vector first-char))))
323
324 ;; It is a matter of taste if you want the minor mode indicated
325 ;; in the mode line...
326 ;; If so, uncomment the next four lines.
327 ;; (or (assq 'iso-accents-mode minor-mode-alist)
328 ;; (setq minor-mode-alist
329 ;; (append minor-mode-alist
330 ;; '((iso-accents-mode " ISO-Acc")))))
331
332 ;;;###autoload
333 (defun iso-accents-mode (&optional arg)
334 "Toggle ISO Accents mode, in which accents modify the following letter.
335 This permits easy insertion of accented characters according to ISO-8859-1.
336 When Iso-accents mode is enabled, accent character keys
337 \(`, ', \", ^, / and ~) do not self-insert; instead, they modify the following
338 letter key so that it inserts an ISO accented letter.
339
340 You can customize ISO Accents mode to a particular language
341 with the command `iso-accents-customize'.
342
343 Special combinations: ~c gives a c with cedilla,
344 ~d gives an Icelandic eth (d with dash).
345 ~t gives an Icelandic thorn.
346 \"s gives German sharp s.
347 /a gives a with ring.
348 /e gives an a-e ligature.
349 ~< and ~> give guillemots.
350 ~! gives an inverted exclamation mark.
351 ~? gives an inverted question mark.
352
353 With an argument, a positive argument enables ISO Accents mode,
354 and a negative argument disables it."
355
356 (interactive "P")
357
358 (if (if arg
359 ;; Negative arg means switch it off.
360 (<= (prefix-numeric-value arg) 0)
361 ;; No arg means toggle.
362 iso-accents-mode)
363 (setq iso-accents-mode nil)
364
365 ;; Enable electric accents.
366 (setq iso-accents-mode t)))
367
368 (defun iso-accents-customize (language)
369 "Customize the ISO accents machinery for a particular language.
370 It selects the customization based on the specifications in the
371 `iso-languages' variable."
372 (interactive (list (completing-read "Language: " iso-languages nil t)))
373 (let ((table (cdr (assoc language iso-languages)))
374 all-accents tail)
375 (if (not table)
376 (error "Unknown language `%s'" language)
377 (setq iso-accents-insert-offset (- (make-char (if (symbolp (car table))
378 (car table)
379 'latin-iso8859-1))
380 128))
381 (if (symbolp (car table))
382 (setq table (cdr table)))
383 (setq iso-language language
384 iso-accents-list table)
385 (if key-translation-map
386 (substitute-key-definition
387 'iso-accents-accent-key nil key-translation-map)
388 (setq key-translation-map (make-sparse-keymap)))
389 ;; Set up translations for all the characters that are used as
390 ;; accent prefixes in this language.
391 (setq tail iso-accents-list)
392 (while tail
393 (define-key key-translation-map (vector (car (car tail)))
394 'iso-accents-accent-key)
395 (setq tail (cdr tail))))))
396
397 (defun iso-accentuate (start end)
398 "Convert two-character sequences in region into accented characters.
399 Noninteractively, this operates on text from START to END.
400 This uses the same conversion that ISO Accents mode uses for type-in."
401 (interactive "r")
402 (save-excursion
403 (save-restriction
404 (narrow-to-region start end)
405 (goto-char start)
406 (forward-char 1)
407 (let (entry)
408 (while (< (point) end)
409 (if (and (memq (preceding-char) iso-accents-enable)
410 (setq entry (cdr (assq (following-char) (assq (preceding-char) iso-accents-list)))))
411 (progn
412 (forward-char -1)
413 (delete-char 2)
414 (insert entry)
415 (setq end (1- end)))
416 (forward-char 1)))))))
417
418 (defun iso-accent-rassoc-unit (value alist)
419 (let (elt acc)
420 (while (and alist (not elt))
421 (setq acc (car (car alist))
422 elt (car (rassq value (cdr (car alist))))
423 alist (cdr alist)))
424 (if elt
425 (cons acc elt))))
426
427 (defun iso-unaccentuate (start end)
428 "Convert accented characters in the region into two-character sequences.
429 Noninteractively, this operates on text from START to END.
430 This uses the opposite of the conversion done by ISO Accents mode for type-in."
431 (interactive "r")
432 (save-excursion
433 (save-restriction
434 (narrow-to-region start end)
435 (goto-char start)
436 (let (entry)
437 (while (< (point) end)
438 (if (and (> (following-char) 127)
439 (setq entry (iso-accent-rassoc-unit (following-char)
440 iso-accents-list)))
441 (progn
442 (delete-char 1)
443 (insert (car entry) (cdr entry))
444 (setq end (1+ end)))
445 (forward-char 1)))))))
446
447 (defun iso-deaccentuate (start end)
448 "Convert accented characters in the region into unaccented characters.
449 Noninteractively, this operates on text from START to END."
450 (interactive "r")
451 (save-excursion
452 (save-restriction
453 (narrow-to-region start end)
454 (goto-char start)
455 (let (entry)
456 (while (< (point) end)
457 (if (and (> (following-char) 127)
458 (setq entry (iso-accent-rassoc-unit (following-char)
459 iso-accents-list)))
460 (progn
461 (delete-char 1)
462 (insert (cdr entry)))
463 (forward-char 1)))))))
464
465 ;; Set up the default settings.
466 (iso-accents-customize "latin-1")
467
468 ;; Use Iso-Accents mode in the minibuffer
469 ;; if it was in use in the previous buffer.
470 (defun iso-acc-minibuf-setup ()
471 (setq iso-accents-mode
472 (save-excursion
473 (set-buffer (window-buffer minibuffer-scroll-window))
474 iso-accents-mode)))
475
476 (add-hook 'minibuffer-setup-hook 'iso-acc-minibuf-setup)
477
478 ;;; arch-tag: 149ff409-7c3e-4574-9b5d-ac038939c0a6
479 ;;; iso-acc.el ends here