Commit | Line | Data |
---|---|---|
4ed46869 KH |
1 | ;;; mule.el --- basic commands for mulitilingual environment |
2 | ||
3 | ;; Copyright (C) 1995 Free Software Foundation, Inc. | |
4 | ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN. | |
5 | ||
6 | ;; Keywords: mule, multilingual, character set, coding system | |
7 | ||
8 | ;; This file is part of GNU Emacs. | |
9 | ||
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) | |
13 | ;; any later version. | |
14 | ||
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. | |
19 | ||
20 | ;; You should have received a copy of the GNU General Public License | |
369314dc KH |
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. | |
4ed46869 KH |
24 | |
25 | ;;; Code: | |
26 | ||
27 | (defconst mule-version "3.0 (MOMIJINOGA)" "\ | |
28 | Version number and name of this version of MULE (multilingual environment).") | |
29 | ||
30 | (defconst mule-version-date "1998.1.1" "\ | |
31 | Distribution date of this version of MULE (multilingual environment).") | |
32 | ||
33 | (defun load-with-code-conversion (fullname file &optional noerror nomessage) | |
34 | "Execute a file of Lisp code named FILE whose absolute path is FULLNAME. | |
35 | The FILE is decoded before evaluation if necessary. | |
36 | If optional second arg NOERROR is non-nil, | |
37 | report no error if FILE doesn't exist. | |
38 | Print messages at start and end of loading unless | |
39 | optional third arg NOMESSAGE is non-nil. | |
40 | Return t if file exists." | |
41 | (if (null (file-readable-p fullname)) | |
42 | (and (null noerror) | |
43 | (signal 'file-error (list "Cannot open load file" file))) | |
44 | ;; Read file with code conversion, and then eval. | |
45 | (let* ((buffer | |
46 | ;; To avoid any autoloading, set default-major-mode to | |
47 | ;; fundamental-mode. | |
48 | (let ((default-major-mode 'fundamental-mode)) | |
49 | ;; We can't use `generate-new-buffer' because files.el | |
50 | ;; is not yet loaded. | |
51 | (get-buffer-create (generate-new-buffer-name " *load*")))) | |
52 | (load-in-progress t)) | |
53 | (or nomessage (message "Loading %s..." file)) | |
54 | (unwind-protect | |
55 | (progn | |
56 | (save-excursion | |
57 | (set-buffer buffer) | |
58 | (insert-file-contents fullname) | |
59 | ;; We must set `buffer-file-name' for `eval-buffer' and | |
60 | ;; `load-history'. | |
61 | (setq buffer-file-name file) | |
62 | ;; Make `kill-buffer' quiet. | |
63 | (set-buffer-modified-p nil)) | |
64 | ;; Eval in the original buffer. | |
65 | (eval-buffer buffer)) | |
66 | (kill-buffer buffer)) | |
67 | (let ((hook (assoc file after-load-alist))) | |
68 | (if hook | |
69 | (mapcar (function eval) (cdr hook)))) | |
70 | (or nomessage noninteractive | |
71 | (message "Loading %s...done" file)) | |
72 | t))) | |
73 | ||
74 | ;; API (Application Program Interface) for charsets. | |
75 | ||
76 | ;; Return t if OBJ is a quoted symbol. | |
77 | (defsubst quoted-symbol-p (obj) | |
78 | (and (listp obj) (eq (car obj) 'quote))) | |
79 | ||
80 | (defsubst charsetp (object) | |
81 | "T is OBJECT is a charset." | |
82 | (and (symbolp object) (vectorp (get object 'charset)))) | |
83 | ||
84 | (defsubst charset-info (charset) | |
85 | "Return a vector of information of CHARSET. | |
86 | The elements of the vector are: | |
87 | CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION, | |
88 | LEADING-CODE-BASE, LEADING-CODE-EXT, | |
89 | ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE, | |
90 | REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION, | |
91 | PLIST, | |
92 | where | |
93 | CHARSET-ID (integer) is the identification number of the charset. | |
94 | DIMENSION (integer) is the number of bytes to represent a character of | |
95 | the charset: 1 or 2. | |
96 | CHARS (integer) is the number of characters in a dimension: 94 or 96. | |
97 | BYTE (integer) is the length of multi-byte form of a character in | |
98 | the charset: one of 1, 2, 3, and 4. | |
99 | WIDTH (integer) is the number of columns a character in the charset | |
100 | occupies on the screen: one of 0, 1, and 2. | |
101 | DIRECTION (integer) is the rendering direction of characters in the | |
102 | charset when rendering. If 0, render from right to left, else | |
103 | render from left to right. | |
104 | LEADING-CODE-BASE (integer) is the base leading-code for the | |
105 | charset. | |
106 | LEADING-CODE-EXT (integer) is the extended leading-code for the | |
107 | charset. All charsets of less than 0xA0 has the value 0. | |
108 | ISO-FINAL-CHAR (character) is the final character of the | |
109 | corresponding ISO 2022 charset. | |
110 | ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked | |
111 | while encoding to variants of ISO 2022 coding system, one of the | |
112 | following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR). | |
113 | REVERSE-CHARSET (integer) is the charset which differs only in | |
114 | LEFT-TO-RIGHT value from the charset. If there's no such a | |
115 | charset, the value is -1. | |
116 | SHORT-NAME (string) is the short name to refer to the charset. | |
117 | LONG-NAME (string) is the long name to refer to the charset | |
118 | DESCRIPTION (string) is the description string of the charset. | |
119 | PLIST (property list) may contain any type of information a user | |
120 | want to put and get by functions `put-charset-property' and | |
121 | `get-charset-property' respectively." | |
122 | (get charset 'charset)) | |
123 | ||
124 | (defmacro charset-id (charset) | |
125 | "Return charset identification number of CHARSET." | |
126 | (if (and (listp charset) (eq (car charset) 'quote)) | |
127 | (aref (charset-info (nth 1 charset)) 0) | |
128 | `(aref (charset-info ,charset) 0))) | |
129 | ||
130 | (defmacro charset-bytes (charset) | |
131 | (if (quoted-symbol-p charset) | |
132 | (aref (charset-info (nth 1 charset)) 1) | |
133 | `(aref (charset-info ,charset) 1))) | |
134 | ||
135 | (defmacro charset-dimension (charset) | |
136 | (if (quoted-symbol-p charset) | |
137 | (aref (charset-info (nth 1 charset)) 2) | |
138 | `(aref (charset-info ,charset) 2))) | |
139 | ||
140 | (defmacro charset-chars (charset) | |
141 | (if (quoted-symbol-p charset) | |
142 | (aref (charset-info (nth 1 charset)) 3) | |
143 | `(aref (charset-info ,charset) 3))) | |
144 | ||
145 | (defmacro charset-width (charset) | |
146 | (if (quoted-symbol-p charset) | |
147 | (aref (charset-info (nth 1 charset)) 4) | |
148 | `(aref (charset-info ,charset) 4))) | |
149 | ||
150 | (defmacro charset-direction (charset) | |
151 | (if (quoted-symbol-p charset) | |
152 | (aref (charset-info (nth 1 charset)) 5) | |
153 | `(aref (charset-info ,charset) 5))) | |
154 | ||
155 | (defmacro charset-iso-final-char (charset) | |
156 | (if (quoted-symbol-p charset) | |
157 | (aref (charset-info (nth 1 charset)) 8) | |
158 | `(aref (charset-info ,charset) 8))) | |
159 | ||
160 | (defmacro charset-iso-graphic-plane (charset) | |
161 | (if (quoted-symbol-p charset) | |
162 | (aref (charset-info (nth 1 charset)) 9) | |
163 | `(aref (charset-info ,charset) 9))) | |
164 | ||
165 | (defmacro charset-reverse-charset (charset) | |
166 | (if (quoted-symbol-p charset) | |
167 | (aref (charset-info (nth 1 charset)) 10) | |
168 | `(aref (charset-info ,charset) 10))) | |
169 | ||
170 | (defmacro charset-short-name (charset) | |
171 | (if (quoted-symbol-p charset) | |
172 | (aref (charset-info (nth 1 charset)) 11) | |
173 | `(aref (charset-info ,charset) 11))) | |
174 | ||
175 | (defmacro charset-long-name (charset) | |
176 | (if (quoted-symbol-p charset) | |
177 | (aref (charset-info (nth 1 charset)) 12) | |
178 | `(aref (charset-info ,charset) 12))) | |
179 | ||
180 | (defmacro charset-description (charset) | |
181 | (if (quoted-symbol-p charset) | |
182 | (aref (charset-info (nth 1 charset)) 13) | |
183 | `(aref (charset-info ,charset) 13))) | |
184 | ||
185 | (defmacro charset-plist (charset) | |
186 | (if (quoted-symbol-p charset) | |
f98e2797 | 187 | `(aref ,(charset-info (nth 1 charset)) 14) |
4ed46869 KH |
188 | `(aref (charset-info ,charset) 14))) |
189 | ||
190 | (defun set-charset-plist (charset plist) | |
191 | (aset (charset-info charset) 14 plist)) | |
192 | ||
193 | (defmacro make-char (charset &optional c1 c2) | |
f98e2797 KH |
194 | "Return a character of CHARSET and position-codes CODE1 and CODE2. |
195 | CODE1 and CODE2 are optional, but if you don't supply | |
196 | sufficient position-codes, return a generic character which stands for | |
197 | all characters or group of characters in the character sets. | |
a73a8c89 | 198 | A generic character can be used to index a char table (e.g. syntax-table)." |
4ed46869 KH |
199 | (if (quoted-symbol-p charset) |
200 | `(make-char-internal ,(charset-id (nth 1 charset)) ,c1 ,c2) | |
201 | `(make-char-internal (charset-id ,charset) ,c1 ,c2))) | |
202 | ||
69eba008 KH |
203 | (defmacro charset-list () |
204 | "Return list of charsets ever defined." | |
13d5617d KH |
205 | 'charset-list) |
206 | ||
207 | (defsubst generic-char-p (char) | |
208 | "Return t if and only if CHAR is a generic character. | |
209 | See also the documentation of make-char." | |
210 | (let ((l (split-char char))) | |
211 | (and (or (= (nth 1 l) 0) (eq (nth 2 l) 0)) | |
212 | (not (eq (car l) 'composition))))) | |
69eba008 | 213 | |
4ed46869 KH |
214 | ;; Coding-system staffs |
215 | ||
216 | ;; Coding-system object is a symbol that has the property | |
217 | ;; `coding-system' and `eol-type'. | |
218 | ;; | |
219 | ;; The value of the property `coding-system' is a coding-vector of the | |
220 | ;; format: [TYPE MNEMONIC DOCSTRING NOT-USED-NOW FLAGS]. | |
221 | ;; See comments in src/coding.c for more detail. The property value | |
222 | ;; may be another coding-system, in which case, the coding-vector | |
223 | ;; should be taken from that coding-system. | |
224 | ;; | |
225 | ;; The value of the property `eol-type' is integer 0..2 or a vector of | |
226 | ;; length 3. The integer value 0, 1, and 2 indicate the format of | |
227 | ;; end-of-line LF, CRLF, and CR respectively. The vector value | |
228 | ;; indicates that the format of end-of-line should be detected | |
229 | ;; automatically. Nth element of the vector is the subsidiary | |
230 | ;; coding-system whose `eol-type' property is integer value. | |
231 | ;; | |
232 | ;; Coding-system may also have properties `post-read-conversion' and | |
233 | ;; `pre-write-conversion and the values are functions. | |
234 | ;; | |
235 | ;; The function in `post-read-conversion' is called after some text is | |
236 | ;; inserted and decoded along the coding-system and before any | |
237 | ;; functions in `after-insert-functions' are called. The arguments to | |
238 | ;; this function is the same as those of a function in | |
239 | ;; `after-insert-functions', i.e. LENGTH of a text while putting point | |
240 | ;; at the head of the text to be decoded | |
241 | ;; | |
242 | ;; The function in `pre-write-conversion' is called after all | |
243 | ;; functions in `write-region-annotate-functions' and | |
244 | ;; `buffer-file-format' are called, and before the text is encoded by | |
245 | ;; the coding-system. The arguments to this function is the same as | |
246 | ;; those of a function in `write-region-annotate-functions', i.e. FROM | |
247 | ;; and TO specifying region of a text. | |
248 | ||
249 | (defsubst coding-vector-type (vec) (aref vec 0)) | |
250 | (defsubst coding-vector-mnemonic (vec) (aref vec 1)) | |
251 | (defsubst coding-vector-docstring (vec) (aref vec 2)) | |
252 | (defsubst coding-vector-flags (vec) (aref vec 4)) | |
253 | ||
254 | ;; Return type of CODING-SYSTEM. | |
255 | (defun coding-system-type (coding-system) | |
69eba008 | 256 | (check-coding-system coding-system) |
4ed46869 KH |
257 | (let ((vec (coding-system-vector coding-system))) |
258 | (if vec (coding-vector-type vec)))) | |
259 | ||
260 | ;; Return mnemonic character of CODING-SYSTEM. | |
261 | (defun coding-system-mnemonic (coding-system) | |
69eba008 | 262 | (check-coding-system coding-system) |
4ed46869 KH |
263 | (let ((vec (coding-system-vector coding-system))) |
264 | (if vec (coding-vector-mnemonic vec) | |
265 | ?-))) | |
266 | ||
267 | ;; Return docstring of CODING-SYSTEM. | |
268 | (defun coding-system-docstring (coding-system) | |
69eba008 | 269 | (check-coding-system coding-system) |
4ed46869 KH |
270 | (let ((vec (coding-system-vector coding-system))) |
271 | (if vec (coding-vector-docstring vec)))) | |
272 | ||
273 | ;; Return flags of CODING-SYSTEM. | |
274 | (defun coding-system-flags (coding-system) | |
69eba008 | 275 | (check-coding-system coding-system) |
4ed46869 KH |
276 | (let ((vec (coding-system-vector coding-system))) |
277 | (if vec (coding-vector-flags vec)))) | |
278 | ||
279 | ;; Return eol-type of CODING-SYSTEM. | |
280 | (defun coding-system-eoltype (coding-system) | |
69eba008 | 281 | (check-coding-system coding-system) |
4ed46869 KH |
282 | (and coding-system |
283 | (or (get coding-system 'eol-type) | |
284 | (coding-system-eoltype (get coding-system 'coding-system))))) | |
285 | ||
286 | ;; Return mnemonic character of eol-type of CODING-SYSTEM. | |
287 | (defun coding-system-eoltype-mnemonic (coding-system) | |
288 | (let ((eol-type (coding-system-eoltype coding-system))) | |
289 | (cond ((vectorp eol-type) eol-mnemonic-undecided) | |
290 | ((eq eol-type 0) eol-mnemonic-unix) | |
291 | ((eq eol-type 1) eol-mnemonic-unix) | |
292 | ((eq eol-type 2) eol-mnemonic-unix) | |
293 | (t ?-)))) | |
294 | ||
295 | ;; Return function for post-read-conversion of CODING-SYSTEM. | |
296 | (defun coding-system-post-read-conversion (coding-system) | |
297 | (and coding-system | |
298 | (symbolp coding-system) | |
299 | (or (get coding-system 'post-read-conversion) | |
300 | (coding-system-post-read-conversion | |
301 | (get coding-system 'coding-system))))) | |
302 | ||
303 | ;; Return function for pre-write-conversion of CODING-SYSTEM. | |
304 | (defun coding-system-pre-write-conversion (coding-system) | |
305 | (and coding-system | |
306 | (symbolp coding-system) | |
307 | (or (get coding-system 'pre-write-conversion) | |
308 | (coding-system-pre-write-conversion | |
309 | (get coding-system 'coding-system))))) | |
310 | ||
311 | (defun make-coding-system (coding-system type mnemonic docstring | |
312 | &optional flags) | |
313 | "Define a new CODING-SYSTEM (symbol). | |
314 | Remaining arguments are TYPE, MNEMONIC, DOCSTRING, and FLAGS (optional). | |
315 | TYPE is an integer value indicating the type of coding-system as follows: | |
316 | 0: Emacs internal format, | |
317 | 1: Shift-JIS (or MS-Kanji) used mainly on Japanese PC, | |
318 | 2: ISO-2022 including many variants, | |
319 | 3: Big5 used mainly on Chinese PC, | |
320 | 4: private, CCL programs provide encoding/decoding algorithm. | |
321 | MNEMONIC is a character to be displayed on mode line for the coding-system. | |
322 | DOCSTRING is a documentation string for the coding-system. | |
323 | FLAGS specifies more precise information of each TYPE. | |
324 | If TYPE is 2 (ISO-2022), FLAGS should be a list of: | |
325 | CHARSET0, CHARSET1, CHARSET2, CHARSET3, SHORT-FORM, | |
326 | ASCII-EOL, ASCII-CNTL, SEVEN, LOCKING-SHIFT, SINGLE-SHIFT, | |
a73a8c89 | 327 | USE-ROMAN, USE-OLDJIS, NO-ISO6429, INIT-BOL, DESIGNATION-BOL. |
4ed46869 KH |
328 | CHARSETn are character sets initially designated to Gn graphic registers. |
329 | If CHARSETn is nil, Gn is never used. | |
330 | If CHARSETn is t, Gn can be used but nothing designated initially. | |
331 | If CHARSETn is a list of character sets, those character sets are | |
332 | designated to Gn on output, but nothing designated to Gn initially. | |
333 | SHORT-FORM non-nil means use short designation sequence on output. | |
334 | ASCII-EOL non-nil means designate ASCII to g0 at end of line on output. | |
335 | ASCII-CNTL non-nil means designate ASCII to g0 before control codes and | |
336 | SPACE on output. | |
337 | SEVEN non-nil means use 7-bit code only on output. | |
338 | LOCKING-SHIFT non-nil means use locking-shift. | |
339 | SINGLE-SHIFT non-nil means use single-shift. | |
340 | USE-ROMAN non-nil means designate JIS0201-1976-Roman instead of ASCII. | |
341 | USE-OLDJIS non-nil means designate JIS0208-1976 instead of JIS0208-1983. | |
342 | NO-ISO6429 non-nil means not use ISO6429's direction specification. | |
69eba008 KH |
343 | INIT-BOL non-nil means any designation state is assumed to be reset |
344 | to initial at each beginning of line on output. | |
345 | DESIGNATION-BOL non-nil means designation sequences should be placed | |
346 | at beginning of line on output. | |
4ed46869 KH |
347 | If TYPE is 4 (private), FLAGS should be a cons of CCL programs, |
348 | for encoding and decoding. See the documentation of CCL for more detail." | |
349 | ||
350 | ;; At first, set a value of `coding-system' property. | |
351 | (let ((coding-vector (make-vector 5 nil))) | |
352 | (aset coding-vector 0 type) | |
353 | (aset coding-vector 1 | |
354 | ;; MNEMONIC must be a printable character. | |
355 | (if (and (> mnemonic ? ) (< mnemonic 127)) mnemonic ? )) | |
356 | (aset coding-vector 2 (if (stringp docstring) docstring "")) | |
357 | (aset coding-vector 3 nil) ; obsolete element | |
358 | (cond ((eq type 2) ; ISO2022 | |
359 | (let ((i 0) | |
360 | (vec (make-vector 32 nil))) | |
361 | (while (< i 4) | |
362 | (let ((charset (car flags))) | |
69eba008 KH |
363 | (or (not charset) (eq charset t) (charsetp charset) |
364 | (if (not (listp charset)) | |
365 | (error "Invalid charset: %s" charset) | |
4ed46869 KH |
366 | (let (elt l) |
367 | (while charset | |
368 | (setq elt (car charset)) | |
69eba008 KH |
369 | (or (not elt) (eq elt t) (charsetp elt) |
370 | (error "Invalid charset: %s" elt)) | |
4ed46869 KH |
371 | (setq l (cons elt l)) |
372 | (setq charset (cdr charset))) | |
373 | (setq charset (nreverse l))))) | |
374 | (aset vec i charset)) | |
375 | (setq flags (cdr flags) i (1+ i))) | |
376 | (while (and (< i 32) flags) | |
377 | (aset vec i (car flags)) | |
378 | (setq flags (cdr flags) i (1+ i))) | |
379 | (aset coding-vector 4 vec))) | |
380 | ((eq type 4) ; private | |
381 | (if (and (consp flags) | |
382 | (vectorp (car flags)) | |
383 | (vectorp (cdr flags))) | |
384 | (aset coding-vector 4 flags) | |
385 | (error "Invalid FLAGS argument for TYPE 4 (CCL)"))) | |
386 | (t (aset coding-vector 4 flags))) | |
387 | (put coding-system 'coding-system coding-vector)) | |
388 | ||
389 | ;; Next, set a value of `eol-type' property. The value is a vector | |
390 | ;; of subsidiary coding-systems, each corresponds to a coding-system | |
391 | ;; for the detected end-of-line format. | |
392 | (let ((codings (vector (intern (format "%s-unix" coding-system)) | |
393 | (intern (format "%s-dos" coding-system)) | |
394 | (intern (format "%s-mac" coding-system)))) | |
395 | (i 0)) | |
396 | (while (< i 3) | |
397 | (put (aref codings i) 'coding-system coding-system) | |
398 | (put (aref codings i) 'eol-type i) | |
399 | (setq i (1+ i))) | |
400 | (put coding-system 'eol-type codings)) | |
401 | ) | |
402 | ||
403 | (defun define-coding-system-alias (symbol new-symbol) | |
404 | "Define NEW-SYMBOL as the same coding system as SYMBOL." | |
405 | (check-coding-system symbol) | |
69eba008 KH |
406 | (put new-symbol 'coding-system symbol) |
407 | (let ((eol-type (coding-system-eoltype symbol))) | |
4ed46869 KH |
408 | (if (vectorp eol-type) |
409 | (let* ((name (symbol-name new-symbol)) | |
69eba008 KH |
410 | (new-eol-type (vector (intern (concat name "-unix")) |
411 | (intern (concat name "-dos")) | |
412 | (intern (concat name "-mac"))))) | |
413 | (define-coding-system-alias (aref eol-type 0) (aref new-eol-type 0)) | |
414 | (define-coding-system-alias (aref eol-type 1) (aref new-eol-type 1)) | |
415 | (define-coding-system-alias (aref eol-type 2) (aref new-eol-type 2)) | |
416 | (setq eol-type new-eol-type))) | |
4ed46869 KH |
417 | (put new-symbol 'eol-type eol-type))) |
418 | ||
419 | (defvar buffer-file-coding-system nil | |
420 | "Coding-system of the file which the current-buffer is visiting.") | |
421 | (make-variable-buffer-local 'buffer-file-coding-system) | |
422 | ;; This value should not be reset by changing major mode. | |
423 | (put 'buffer-file-coding-system 'permanent-local t) | |
424 | ||
425 | (defun set-buffer-file-coding-system (coding-system &optional force) | |
426 | "Set buffer-file-coding-system of the current buffer to CODING-SYSTEM. | |
427 | If eol-type of the current buffer-file-coding-system is an integer value N, and | |
a73a8c89 | 428 | eol-type of CODING-SYSTEM is a vector, the Nth element of the vector is used |
4ed46869 KH |
429 | instead of CODING-SYSTEM itself. |
430 | Optional prefix argument FORCE non-nil means CODING-SYSTEM is set | |
431 | regardless of eol-type of the current buffer-file-coding-system." | |
432 | (interactive "zBuffer-file-coding-system: \nP") | |
433 | (check-coding-system coding-system) | |
434 | (if (null force) | |
435 | (let ((x (coding-system-eoltype buffer-file-coding-system)) | |
436 | (y (coding-system-eoltype coding-system))) | |
437 | (if (and (numberp x) (>= x 0) (<= x 2) (vectorp y)) | |
438 | (setq coding-system (aref y x))))) | |
439 | (setq buffer-file-coding-system coding-system) | |
440 | (set-buffer-modified-p t) | |
441 | (force-mode-line-update)) | |
442 | ||
443 | (defun set-current-process-coding-system (input output) | |
444 | (interactive | |
445 | "zCoding-system for process input: \nzCoding-system for process output: ") | |
446 | (let ((proc (get-buffer-process (current-buffer)))) | |
447 | (if (null proc) | |
448 | (error "no process") | |
449 | (check-coding-system input) | |
450 | (check-coding-system output) | |
451 | (set-process-coding-system proc input output))) | |
452 | (force-mode-line-update)) | |
453 | ||
454 | (defvar default-process-coding-system (cons nil nil) | |
455 | "Cons of default values used to read from and write to process.") | |
456 | ||
457 | (defun set-coding-priority (arg) | |
458 | "Set priority of coding-category according to LIST. | |
459 | LIST is a list of coding-categories ordered by priority." | |
460 | (let (l) | |
461 | ;; Put coding-categories listed in ARG to L while checking the | |
462 | ;; validity. We assume that `coding-category-list' contains whole | |
463 | ;; coding-categories. | |
464 | (while arg | |
465 | (if (null (memq (car arg) coding-category-list)) | |
466 | (error "Invalid element in argument: %s" (car arg))) | |
467 | (setq l (cons (car arg) l)) | |
468 | (setq arg (cdr arg))) | |
469 | ;; Put coding-category not listed in ARG to L. | |
470 | (while coding-category-list | |
471 | (if (null (memq (car coding-category-list) l)) | |
472 | (setq l (cons (car coding-category-list) l))) | |
473 | (setq coding-category-list (cdr coding-category-list))) | |
474 | ;; Update `coding-category-list' and return it. | |
475 | (setq coding-category-list (nreverse l)))) | |
476 | ||
477 | ;;; FILE I/O | |
478 | ||
479 | ;; Set buffer-file-coding-system of the current buffer after some text | |
480 | ;; is inserted. | |
481 | (defun after-insert-file-set-buffer-file-coding-system (inserted) | |
482 | (if last-coding-system-used | |
483 | (let ((coding-system | |
484 | (find-new-buffer-file-coding-system last-coding-system-used)) | |
485 | (modified-p (buffer-modified-p))) | |
486 | (if coding-system | |
487 | (set-buffer-file-coding-system coding-system)) | |
488 | (set-buffer-modified-p modified-p))) | |
489 | nil) | |
490 | ||
491 | (setq after-insert-file-functions | |
492 | (cons 'after-insert-file-set-buffer-file-coding-system | |
493 | after-insert-file-functions)) | |
494 | ||
495 | ;; The coding-vector and eol-type of coding-system returned is decided | |
496 | ;; independently in the following order. | |
497 | ;; 1. That of buffer-file-coding-system locally bound. | |
498 | ;; 2. That of CODING. | |
499 | ||
500 | (defun find-new-buffer-file-coding-system (coding) | |
501 | "Return a coding system for a buffer when a file of CODING is inserted. | |
a73a8c89 KH |
502 | The local variable `buffer-file-coding-system' of the current buffer |
503 | is set to the returned value. | |
504 | Return nil if there's no need of setting new buffer-file-coding-system." | |
4ed46869 KH |
505 | (let (local-coding local-eol |
506 | found-eol | |
507 | new-coding new-eol) | |
508 | (if (null coding) | |
509 | ;; Nothing found about coding. | |
510 | nil | |
511 | ||
512 | ;; Get information of the current local value of | |
513 | ;; `buffer-file-coding-system' in LOCAL-EOL and LOCAL-CODING. | |
514 | (if (local-variable-p 'buffer-file-coding-system) | |
515 | ;; Something already set locally. | |
516 | (progn | |
517 | (setq local-eol (coding-system-eoltype buffer-file-coding-system)) | |
518 | (if (null (numberp local-eol)) | |
519 | ;; But eol-type is not yet set. | |
520 | (setq local-eol nil)) | |
521 | (if (null (eq (coding-system-type buffer-file-coding-system) t)) | |
13d5617d | 522 | ;; This is not `undecided'. |
4ed46869 KH |
523 | (progn |
524 | (setq local-coding buffer-file-coding-system) | |
525 | (while (symbolp (get local-coding 'coding-system)) | |
526 | (setq local-coding (get local-coding 'coding-system)))) | |
527 | ))) | |
528 | ||
529 | (if (and local-eol local-coding) | |
530 | ;; The current buffer has already set full coding-system, we | |
531 | ;; had better not change it. | |
532 | nil | |
533 | ||
534 | (setq found-eol (coding-system-eoltype coding)) | |
535 | (if (null (numberp found-eol)) | |
536 | ;; But eol-type is not found. | |
537 | (setq found-eol nil)) | |
538 | (if (eq (coding-system-type coding) t) | |
13d5617d KH |
539 | ;; This is `undecided', which means nothing found except |
540 | ;; for eol-type. | |
4ed46869 KH |
541 | (setq coding nil)) |
542 | ||
543 | ;; The local setting takes precedence over the found one. | |
544 | (setq new-coding (or local-coding coding)) | |
545 | (setq new-eol (or local-eol found-eol)) | |
546 | (if (and (numberp new-eol) | |
547 | (vectorp (coding-system-eoltype new-coding))) | |
548 | (setq new-coding | |
549 | (aref (coding-system-eoltype new-coding) new-eol))) | |
550 | new-coding)))) | |
551 | ||
a73a8c89 KH |
552 | (defun make-unification-table (&rest args) |
553 | "Make a unification table (char table) from arguments. | |
13d5617d KH |
554 | Each argument is a list of the form (FROM . TO), |
555 | where FROM is a character to be unified to TO. | |
556 | ||
557 | FROM can be a generic character (see make-char). In this case, TO is | |
558 | a generic character containing the same number of charcters or a | |
559 | oridinal character. If FROM and TO are both generic characters, all | |
560 | characters belonging to FROM are unified to characters belonging to TO | |
561 | without changing their position code(s)." | |
a73a8c89 KH |
562 | (let ((table (make-char-table 'character-unification-table)) |
563 | revlist) | |
564 | (while args | |
565 | (let ((elts (car args))) | |
566 | (while elts | |
13d5617d KH |
567 | (let* ((from (car (car elts))) |
568 | (from-i 0) ; degree of freedom of FROM | |
569 | (from-rev (nreverse (split-char from))) | |
570 | (to (cdr (car elts))) | |
571 | (to-i 0) ; degree of freedom of TO | |
572 | (to-rev (nreverse (split-char to)))) | |
573 | ;; Check numbers of heading 0s in FROM-REV and TO-REV. | |
574 | (while (eq (car from-rev) 0) | |
575 | (setq from-i (1+ from-i) from-rev (cdr from-rev))) | |
576 | (while (eq (car to-rev) 0) | |
577 | (setq to-i (1+ to-i) to-rev (cdr to-rev))) | |
578 | (if (and (/= from-i to-i) (/= to-i 0)) | |
579 | (error "Invalid character pair (%d . %d)" from to)) | |
580 | ;; If we have already unified TO to TO-ALT, FROM should | |
581 | ;; also be unified to TO-ALT. But, this is only if TO is | |
582 | ;; a generic character or TO-ALT is not a generic | |
583 | ;; character. | |
584 | (let ((to-alt (aref table to))) | |
585 | (if (and to-alt | |
586 | (or (> to-i 0) (not (generic-char-p to-alt)))) | |
587 | (setq to to-alt))) | |
588 | (if (> from-i 0) | |
589 | (set-char-table-default table from to) | |
590 | (aset table from to)) | |
a73a8c89 KH |
591 | ;; If we have already unified some chars to FROM, they |
592 | ;; should also be unified to TO. | |
593 | (let ((l (assq from revlist))) | |
594 | (if l | |
595 | (let ((ch (car l))) | |
596 | (setcar l to) | |
597 | (setq l (cdr l)) | |
598 | (while l | |
599 | (aset table ch to) | |
600 | (setq l (cdr l)) )))) | |
601 | ;; Now update REVLIST. | |
602 | (let ((l (assq to revlist))) | |
603 | (if l | |
604 | (setcdr l (cons from (cdr l))) | |
605 | (setq revlist (cons (list to from) revlist))))) | |
606 | (setq elts (cdr elts)))) | |
607 | (setq args (cdr args))) | |
608 | ;; Return TABLE just created. | |
609 | table)) | |
610 | ||
69eba008 KH |
611 | ;;; Initialize some variables. |
612 | ||
613 | (put 'use-default-ascent 'char-table-extra-slots 0) | |
614 | (setq use-default-ascent (make-char-table 'use-default-ascent)) | |
615 | ||
616 | ;;; | |
4ed46869 KH |
617 | (provide 'mule) |
618 | ||
619 | ;;; mule.el ends here |