1 ;;; composite.el --- support character composition
3 ;; Copyright (C) 1999 Electrotechnical Laboratory, JAPAN.
4 ;; Licensed to the Free Software Foundation.
6 ;; Keywords: mule, multilingual, character composition
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 2, or (at your option)
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; see the file COPYING. If not, write to the
22 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
23 ;; Boston, MA 02111-1307, USA.
29 (defconst reference-point-alist
30 '((tl . 0) (tc . 1) (tr . 2)
31 (Bl . 3) (Bc . 4) (Br . 5)
32 (bl . 6) (bc . 7) (br . 8)
33 (cl . 9) (cc . 10) (cr . 11)
34 (top-left . 0) (top-center . 1) (top-right . 2)
35 (base-left . 3) (base-center . 4) (base-right . 5)
36 (bottom-left . 6) (bottom-center . 7) (bottom-right . 8)
37 (center-left . 9) (center-center . 10) (center-right . 11)
38 ;; For backward compatibility...
39 (ml . 3) (mc . 10) (mr . 5)
40 (mid-left . 3) (mid-center . 10) (mid-right . 5))
41 "Alist of symbols vs integer codes of glyph reference points.
42 A glyph reference point symbol is to be used to specify a composition
43 rule in COMPONENTS argument to such functions as `compose-region'.
45 Meanings of glyph reference point codes are as follows:
47 0----1----2 <---- ascent 0:tl or top-left
48 | | 1:tc or top-center
50 | | 3:Bl or base-left 9:cl or center-left
51 9 10 11 <---- center 4:Bc or base-center 10:cc or center-center
52 | | 5:Br or base-right 11:cr or center-right
53 --3----4----5-- <-- baseline 6:bl or bottom-left
54 | | 7:bc or bottom-center
55 6----7----8 <---- descent 8:br or bottom-right
57 Glyph reference point symbols are to be used to specify composition
58 rule of the form \(GLOBAL-REF-POINT . NEW-REF-POINT), where
59 GLOBAL-REF-POINT is a reference point in the overall glyphs already
60 composed, and NEW-REF-POINT is a reference point in the new glyph to
63 For instance, if GLOBAL-REF-POINT is `br' (bottom-right) and
64 NEW-REF-POINT is `tc' (top-center), the overall glyph is updated as
65 follows (the point `*' corresponds to both reference points):
67 +-------+--+ <--- new ascent
71 -- | | |-- <--- baseline \(doesn't change)
75 +----+-----+ <--- new descent
78 ;; Encode composition rule RULE into an integer value. RULE is a cons
79 ;; of global and new reference point symbols.
80 ;; This must be compatible with C macro COMPOSITION_ENCODE_RULE
81 ;; defined in composite.h.
83 (defun encode-composition-rule (rule)
84 (if (and (integerp rule) (< rule 144))
88 (error "Invalid composition rule: %S" rule))
89 (let ((gref (car rule))
92 (setq gref (cdr (assq gref reference-point-alist))))
94 (setq nref (cdr (assq nref reference-point-alist))))
95 (or (and (>= gref 0) (< gref 12) (>= nref 0) (< nref 12))
96 (error "Invalid composition rule: %S" rule))
97 (+ (* gref 12) nref))))
99 ;; Decode encoded composition rule RULE-CODE. The value is a cons of
100 ;; global and new reference point symbols.
101 ;; This must be compatible with C macro COMPOSITION_DECODE_RULE
102 ;; defined in composite.h.
104 (defun decode-composition-rule (rule-code)
105 (or (and (natnump rule-code) (< rule-code 144))
106 (error "Invalid encoded composition rule: %S" rule-code))
107 (let ((gref (car (rassq (/ rule-code 12) reference-point-alist)))
108 (nref (car (rassq (% rule-code 12) reference-point-alist))))
109 (or (and gref (symbolp gref) nref (symbolp nref))
110 (error "Invalid composition rule code: %S" rule-code))
113 ;; Encode composition rules in composition components COMPONENTS. The
114 ;; value is a copy of COMPONENTS, where composition rules (cons of
115 ;; global and new glyph reference point symbols) are replaced with
116 ;; encoded composition rules. Optional 2nd argument NOCOPY non-nil
117 ;; means don't make a copy but modify COMPONENTS directly.
119 (defun encode-composition-components (components &optional nocopy)
121 (setq components (copy-sequence components)))
122 (if (vectorp components)
123 (let ((len (length components))
127 (encode-composition-rule (aref components i)))
129 (let ((tail (cdr components)))
132 (encode-composition-rule (car tail)))
133 (setq tail (nthcdr 2 tail)))))
136 ;; Decode composition rule codes in composition components COMPONENTS.
137 ;; The value is a copy of COMPONENTS, where composition rule codes are
138 ;; replaced with composition rules (cons of global and new glyph
139 ;; reference point symbols). Optional 2nd argument NOCOPY non-nil
140 ;; means don't make a copy but modify COMPONENTS directly.
141 ;; It is assumed that COMPONENTS is a vector and is for rule-base
142 ;; composition, thus (2N+1)th elements are rule codes.
144 (defun decode-composition-components (components &optional nocopy)
146 (setq components (copy-sequence components)))
147 (let ((len (length components))
151 (decode-composition-rule (aref components i)))
155 (defun compose-region (start end &optional components modification-func)
156 "Compose characters in the current region.
158 Characters are composed relatively, i.e. composed by overstricking or
159 stacking depending on ascent, descent and other properties.
161 When called from a program, expects these four arguments.
163 First two arguments START and END are positions (integers or markers)
164 specifying the region.
166 Optional 3rd argument COMPONENTS, if non-nil, is a character, a string
167 Optional 3rd argument COMPONENTS, if non-nil, is a character, a string
168 or a vector or list of integers and rules.
170 If it is a character, it is an alternate character to display instead
171 of the text in the region.
173 If it is a string, the elements are alternate characters.
175 If it is a vector or list, it is a sequence of alternate characters and
176 composition rules, where (2N)th elements are characters and (2N+1)th
177 elements are composition rules to specify how to compose (2N+2)th
178 elements with previously composed N glyphs.
180 A composition rule is a cons of global and new glyph reference point
181 symbols. See the documentation of `reference-point-alist' for more
184 Optional 4th argument MODIFICATION-FUNC is a function to call to
185 adjust the composition when it gets invalid because of a change of
186 text in the composition."
188 (let ((modified-p (buffer-modified-p))
189 (buffer-read-only nil))
190 (if (or (vectorp components) (listp components))
191 (setq components (encode-composition-components components)))
192 (compose-region-internal start end components modification-func)
193 (set-buffer-modified-p modified-p)))
195 (defun decompose-region (start end)
196 "Decompose text in the current region.
198 When called from a program, expects two arguments,
199 positions (integers or markers) specifying the region."
201 (let ((modified-p (buffer-modified-p))
202 (buffer-read-only nil))
203 (remove-text-properties start end '(composition nil))
204 (set-buffer-modified-p modified-p)))
206 (defun compose-string (string &optional start end components modification-func)
207 "Compose characters in string STRING.
209 The return value is STRING with the `composition' property put on all
210 the characters in it.
212 Optional 2nd and 3rd arguments START and END specify the range of
213 STRING to be composed. They default to the beginning and the end of
216 Optional 4th argument COMPONENTS, if non-nil, is a character or a
217 sequence (vector, list, or string) of integers. See the function
218 `compose-region' for more detail.
220 Optional 5th argument MODIFICATION-FUNC is a function to call to
221 adjust the composition when it gets invalid because of a change of
222 text in the composition."
223 (if (or (vectorp components) (listp components))
224 (setq components (encode-composition-components components)))
225 (or start (setq start 0))
226 (or end (setq end (length string)))
227 (compose-string-internal string start end components modification-func)
230 (defun decompose-string (string)
231 "Return STRING where `composition' property is removed."
232 (remove-text-properties 0 (length string) '(composition nil) string)
235 (defun compose-chars (&rest args)
236 "Return a string from arguments in which all characters are composed.
237 For relative composition, arguments are characters.
238 For rule-based composition, Mth \(where M is odd) arguments are
239 characters, and Nth \(where N is even) arguments are composition rules.
240 A composition rule is a cons of glyph reference points of the form
241 \(GLOBAL-REF-POINT . NEW-REF-POINT). See the documentation of
242 `reference-point-alist' for more detail."
243 (let (str components)
244 (if (consp (car (cdr args)))
245 ;; Rule-base composition.
246 (let ((len (length args))
247 (tail (encode-composition-components args 'nocopy)))
250 (setq str (cons (car tail) str))
251 (setq tail (nthcdr 2 tail)))
252 (setq str (concat (nreverse str))
254 ;; Relative composition.
255 (setq str (concat args)))
256 (compose-string-internal str 0 (length str) components)))
258 (defun find-composition (pos &optional limit string detail-p)
259 "Return information about a composition at or nearest to buffer position POS.
261 If the character at POS has `composition' property, the value is a list
262 of FROM, TO, and VALID-P.
264 FROM and TO specify the range of text that has the same `composition'
265 property, VALID-P is non-nil if and only if this composition is valid.
267 If there's no composition at POS, and the optional 2nd argument LIMIT
268 is non-nil, search for a composition toward LIMIT.
270 If no composition is found, return nil.
272 Optional 3rd argument STRING, if non-nil, is a string to look for a
273 composition in; nil means the current buffer.
275 If a valid composition is found and the optional 4th argument DETAIL-P
276 is non-nil, the return value is a list of FROM, TO, COMPONENTS,
277 RELATIVE-P, MOD-FUNC, and WIDTH.
279 COMPONENTS is a vector of integers, the meaning depends on RELATIVE-P.
281 RELATIVE-P is t if the composition method is relative, else nil.
283 If RELATIVE-P is t, COMPONENTS is a vector of characters to be
284 composed. If RELATIVE-P is nil, COMPONENTS is a vector of characters
285 and composition rules as described in `compose-region'.
287 MOD-FUNC is a modification function of the composition.
289 WIDTH is a number of columns the composition occupies on the screen."
290 (let ((result (find-composition-internal pos limit string detail-p)))
291 (if (and detail-p result (nth 2 result) (not (nth 3 result)))
292 ;; This is a valid rule-base composition.
293 (decode-composition-components (nth 2 result) 'nocopy))
297 (defun compose-chars-after (pos &optional limit object)
298 "Compose characters in current buffer after position POS.
300 It looks up the char-table `composition-function-table' (which see) by
301 a character after POS. If non-nil value is found, the format of the
302 value should be an alist of PATTERNs vs FUNCs, where PATTERNs are
303 regular expressions and FUNCs are functions. If the text after POS
304 matches one of PATTERNs, call the corresponding FUNC with three
305 arguments POS, TO, and PATTERN, where TO is the end position of text
306 matching PATTERN, and return what FUNC returns. Otherwise, return
309 FUNC is responsible for composing the text properly. The return value
311 nil -- if no characters were composed.
312 CHARS (integer) -- if CHARS characters were composed.
314 Optional 2nd arg LIMIT, if non-nil, limits the matching of text.
316 Optional 3rd arg OBJECT, if non-nil, is a string that contains the
317 text to compose. In that case, POS and LIMIT index into the string.
319 This function is the default value of `compose-chars-after-function'."
320 (let ((tail (aref composition-function-table (char-after pos)))
325 (while (and tail (not func))
326 (setq pattern (car (car tail))
327 func (cdr (car tail)))
330 (and (re-search-forward pattern limit t)
331 (= (match-beginning 0) pos))
332 (looking-at pattern))
333 (setq result (funcall func pos (match-end 0) pattern nil))
334 (setq func nil tail (cdr tail)))))))
337 (defun compose-last-chars (args)
338 "Compose last characters.
339 The argument is a parameterized event of the form
340 \(compose-last-chars N COMPONENTS),
341 where N is the number of characters before point to compose,
342 COMPONENTS, if non-nil, is the same as the argument to `compose-region'
343 \(which see). If it is nil, `compose-chars-after' is called,
344 and that function find a proper rule to compose the target characters.
345 This function is intended to be used from input methods.
346 The global keymap binds special event `compose-last-chars' to this
347 function. Input method may generate an event (compose-last-chars N COMPONENTS)
348 after a sequence character events."
350 (let ((chars (nth 1 args)))
351 (if (and (numberp chars)
352 (>= (- (point) (point-min)) chars))
354 (compose-region (- (point) chars) (point) (nth 2 args))
355 (compose-chars-after (- (point) chars) (point))))))
357 (global-set-key [compose-last-chars] 'compose-last-chars)
360 ;;; Automatic character composition.
362 (defvar composition-function-table
363 (make-char-table nil)
364 "Char table of functions for automatic character composition.
365 For each character that has to be composed automatically with
366 preceding and/or following characters, this char table contains
367 a function to call to compose that character.
369 Each function is called with two arguments, POS and STRING.
371 If STRING is nil, POS is a position in the current buffer, and the
372 function has to compose a character at POS with surrounding characters
373 in the current buffer.
375 Otherwise, STRING is a string, and POS is an index into the string. In
376 this case, the function has to compose a character at POS with
377 surrounding characters in the string.
379 See also the command `toggle-auto-composition'.")
381 ;; Copied from font-lock.el.
383 ;; Borrowed from lazy-lock.el.
384 ;; We use this to preserve or protect things when modifying text properties.
385 (defmacro save-buffer-state (varlist &rest body)
386 "Bind variables according to VARLIST and eval BODY restoring buffer state."
387 `(let* ,(append varlist
388 '((modified (buffer-modified-p)) (buffer-undo-list t)
389 (inhibit-read-only t) (inhibit-point-motion-hooks t)
390 (inhibit-modification-hooks t)
391 deactivate-mark buffer-file-name buffer-file-truename))
394 (restore-buffer-modified-p nil))))
395 (put 'save-buffer-state 'lisp-indent-function 1)
396 ;; Fixme: This makes bootstrapping fail with this error.
397 ;; Symbol's function definition is void: eval-defun
398 ;;(def-edebug-spec save-buffer-state let)
401 (defvar auto-composition-chunk-size 500
402 "*Automatic composition uses chunks of this many characters, or smaller.")
404 (defun auto-compose-chars (pos string)
405 "Compose characters after the buffer position POS.
406 If STRING is non-nil, it is a string, and POS is an index into the string.
407 In that case, compose characters in the string.
409 This function is the default value of `auto-composition-function' (which see)."
410 (save-buffer-state nil
415 (end (if string (length string) (point-max)))
416 (limit (next-single-property-change pos 'auto-composed string
420 (if (> (- limit start) auto-composition-chunk-size)
421 (setq limit (+ start auto-composition-chunk-size)))
422 (while (and (< pos end)
423 (setq ch (if string (aref string pos)
427 (setq func (aref composition-function-table ch))
429 (setq newpos (funcall func pos string)
430 pos (if (and (integerp newpos) (> newpos pos))
433 (setq pos (1+ pos))))
436 (put-text-property start pos 'auto-composed t string)))))))
438 (setq auto-composition-function 'auto-compose-chars)
440 (defun toggle-auto-composition (&optional arg)
441 "Change whether automatic character composition is enabled in this buffer.
442 With arg, enable it iff arg is positive."
444 (let ((enable (if (null arg) (not auto-composition-function)
445 (> (prefix-numeric-value arg) 0))))
447 (kill-local-variable 'auto-composition-function)
448 (make-local-variable 'auto-composition-function)
449 (setq auto-composition-function nil)
450 (save-buffer-state nil
453 (decompose-region (point-min) (point-max)))))
455 (save-buffer-state nil
458 (put-text-property (point-min) (point-max) 'auto-composed nil)))))
460 ;;; The following codes are only for backward compatibility with Emacs
461 ;;; 20.4 and earlier.
463 (defun decompose-composite-char (char &optional type with-composition-rule)
464 "Convert CHAR to string.
466 If optional 2nd arg TYPE is non-nil, it is `string', `list', or
467 `vector'. In this case, CHAR is converted to string, list of CHAR, or
468 vector of CHAR respectively.
469 Optional 3rd arg WITH-COMPOSITION-RULE is ignored."
470 (cond ((or (null type) (eq type 'string)) (char-to-string char))
471 ((eq type 'list) (list char))
474 (make-obsolete 'decompose-composite-char 'char-to-string "21.1")
477 ;;; composite.el ends here