Commit | Line | Data |
---|---|---|
4ed46869 KH |
1 | ;;; mule.el --- basic commands for mulitilingual environment |
2 | ||
4ed46869 | 3 | ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN. |
58cfed09 | 4 | ;; Licensed to the Free Software Foundation. |
4ed46869 KH |
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*")))) | |
850101ed | 52 | (enable-multibyte-characters t) |
4ed46869 KH |
53 | (load-in-progress t)) |
54 | (or nomessage (message "Loading %s..." file)) | |
55 | (unwind-protect | |
56 | (progn | |
57 | (save-excursion | |
58 | (set-buffer buffer) | |
59 | (insert-file-contents fullname) | |
60 | ;; We must set `buffer-file-name' for `eval-buffer' and | |
61 | ;; `load-history'. | |
62 | (setq buffer-file-name file) | |
63 | ;; Make `kill-buffer' quiet. | |
64 | (set-buffer-modified-p nil)) | |
65 | ;; Eval in the original buffer. | |
66 | (eval-buffer buffer)) | |
cfc70cdf RS |
67 | (let (kill-buffer-hook kill-buffer-query-functions) |
68 | (kill-buffer buffer))) | |
4ed46869 KH |
69 | (let ((hook (assoc file after-load-alist))) |
70 | (if hook | |
71 | (mapcar (function eval) (cdr hook)))) | |
72 | (or nomessage noninteractive | |
73 | (message "Loading %s...done" file)) | |
74 | t))) | |
75 | ||
76 | ;; API (Application Program Interface) for charsets. | |
77 | ||
78 | ;; Return t if OBJ is a quoted symbol. | |
79 | (defsubst quoted-symbol-p (obj) | |
80 | (and (listp obj) (eq (car obj) 'quote))) | |
81 | ||
82 | (defsubst charsetp (object) | |
83 | "T is OBJECT is a charset." | |
84 | (and (symbolp object) (vectorp (get object 'charset)))) | |
85 | ||
86 | (defsubst charset-info (charset) | |
87 | "Return a vector of information of CHARSET. | |
88 | The elements of the vector are: | |
89 | CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION, | |
90 | LEADING-CODE-BASE, LEADING-CODE-EXT, | |
91 | ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE, | |
92 | REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION, | |
93 | PLIST, | |
94 | where | |
95 | CHARSET-ID (integer) is the identification number of the charset. | |
96 | DIMENSION (integer) is the number of bytes to represent a character of | |
97 | the charset: 1 or 2. | |
98 | CHARS (integer) is the number of characters in a dimension: 94 or 96. | |
99 | BYTE (integer) is the length of multi-byte form of a character in | |
100 | the charset: one of 1, 2, 3, and 4. | |
101 | WIDTH (integer) is the number of columns a character in the charset | |
102 | occupies on the screen: one of 0, 1, and 2. | |
103 | DIRECTION (integer) is the rendering direction of characters in the | |
104 | charset when rendering. If 0, render from right to left, else | |
105 | render from left to right. | |
106 | LEADING-CODE-BASE (integer) is the base leading-code for the | |
107 | charset. | |
108 | LEADING-CODE-EXT (integer) is the extended leading-code for the | |
109 | charset. All charsets of less than 0xA0 has the value 0. | |
110 | ISO-FINAL-CHAR (character) is the final character of the | |
111 | corresponding ISO 2022 charset. | |
112 | ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked | |
113 | while encoding to variants of ISO 2022 coding system, one of the | |
114 | following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR). | |
115 | REVERSE-CHARSET (integer) is the charset which differs only in | |
116 | LEFT-TO-RIGHT value from the charset. If there's no such a | |
117 | charset, the value is -1. | |
118 | SHORT-NAME (string) is the short name to refer to the charset. | |
119 | LONG-NAME (string) is the long name to refer to the charset | |
120 | DESCRIPTION (string) is the description string of the charset. | |
121 | PLIST (property list) may contain any type of information a user | |
122 | want to put and get by functions `put-charset-property' and | |
123 | `get-charset-property' respectively." | |
124 | (get charset 'charset)) | |
125 | ||
126 | (defmacro charset-id (charset) | |
127 | "Return charset identification number of CHARSET." | |
128 | (if (and (listp charset) (eq (car charset) 'quote)) | |
129 | (aref (charset-info (nth 1 charset)) 0) | |
130 | `(aref (charset-info ,charset) 0))) | |
131 | ||
132 | (defmacro charset-bytes (charset) | |
900dc6e3 KH |
133 | "Return bytes of CHARSET. |
134 | See the function `charset-info' for more detail." | |
4ed46869 KH |
135 | (if (quoted-symbol-p charset) |
136 | (aref (charset-info (nth 1 charset)) 1) | |
137 | `(aref (charset-info ,charset) 1))) | |
138 | ||
139 | (defmacro charset-dimension (charset) | |
900dc6e3 KH |
140 | "Return dimension of CHARSET. |
141 | See the function `charset-info' for more detail." | |
4ed46869 KH |
142 | (if (quoted-symbol-p charset) |
143 | (aref (charset-info (nth 1 charset)) 2) | |
144 | `(aref (charset-info ,charset) 2))) | |
145 | ||
146 | (defmacro charset-chars (charset) | |
900dc6e3 KH |
147 | "Return character numbers contained in a dimension of CHARSET. |
148 | See the function `charset-info' for more detail." | |
4ed46869 KH |
149 | (if (quoted-symbol-p charset) |
150 | (aref (charset-info (nth 1 charset)) 3) | |
151 | `(aref (charset-info ,charset) 3))) | |
152 | ||
153 | (defmacro charset-width (charset) | |
900dc6e3 KH |
154 | "Return width (how many column occupied on a screen) of CHARSET. |
155 | See the function `charset-info' for more detail." | |
4ed46869 KH |
156 | (if (quoted-symbol-p charset) |
157 | (aref (charset-info (nth 1 charset)) 4) | |
158 | `(aref (charset-info ,charset) 4))) | |
159 | ||
160 | (defmacro charset-direction (charset) | |
900dc6e3 KH |
161 | "Return direction of CHARSET. |
162 | See the function `charset-info' for more detail." | |
4ed46869 KH |
163 | (if (quoted-symbol-p charset) |
164 | (aref (charset-info (nth 1 charset)) 5) | |
165 | `(aref (charset-info ,charset) 5))) | |
166 | ||
167 | (defmacro charset-iso-final-char (charset) | |
900dc6e3 KH |
168 | "Return final char of CHARSET. |
169 | See the function `charset-info' for more detail." | |
4ed46869 KH |
170 | (if (quoted-symbol-p charset) |
171 | (aref (charset-info (nth 1 charset)) 8) | |
172 | `(aref (charset-info ,charset) 8))) | |
173 | ||
174 | (defmacro charset-iso-graphic-plane (charset) | |
900dc6e3 KH |
175 | "Return graphic plane of CHARSET. |
176 | See the function `charset-info' for more detail." | |
4ed46869 KH |
177 | (if (quoted-symbol-p charset) |
178 | (aref (charset-info (nth 1 charset)) 9) | |
179 | `(aref (charset-info ,charset) 9))) | |
180 | ||
181 | (defmacro charset-reverse-charset (charset) | |
900dc6e3 KH |
182 | "Return reverse charset of CHARSET. |
183 | See the function `charset-info' for more detail." | |
4ed46869 KH |
184 | (if (quoted-symbol-p charset) |
185 | (aref (charset-info (nth 1 charset)) 10) | |
186 | `(aref (charset-info ,charset) 10))) | |
187 | ||
188 | (defmacro charset-short-name (charset) | |
900dc6e3 KH |
189 | "Return short name of CHARSET. |
190 | See the function `charset-info' for more detail." | |
4ed46869 KH |
191 | (if (quoted-symbol-p charset) |
192 | (aref (charset-info (nth 1 charset)) 11) | |
193 | `(aref (charset-info ,charset) 11))) | |
194 | ||
195 | (defmacro charset-long-name (charset) | |
900dc6e3 KH |
196 | "Return long name of CHARSET. |
197 | See the function `charset-info' for more detail." | |
4ed46869 KH |
198 | (if (quoted-symbol-p charset) |
199 | (aref (charset-info (nth 1 charset)) 12) | |
200 | `(aref (charset-info ,charset) 12))) | |
201 | ||
202 | (defmacro charset-description (charset) | |
900dc6e3 KH |
203 | "Return descriptoin of CHARSET. |
204 | See the function `charset-info' for more detail." | |
4ed46869 KH |
205 | (if (quoted-symbol-p charset) |
206 | (aref (charset-info (nth 1 charset)) 13) | |
207 | `(aref (charset-info ,charset) 13))) | |
208 | ||
209 | (defmacro charset-plist (charset) | |
900dc6e3 KH |
210 | "Return list charset property of CHARSET. |
211 | See the function `charset-info' for more detail." | |
4ed46869 | 212 | (if (quoted-symbol-p charset) |
f98e2797 | 213 | `(aref ,(charset-info (nth 1 charset)) 14) |
4ed46869 KH |
214 | `(aref (charset-info ,charset) 14))) |
215 | ||
216 | (defun set-charset-plist (charset plist) | |
900dc6e3 | 217 | "Set CHARSET's property list to PLIST, and retrun PLIST." |
4ed46869 KH |
218 | (aset (charset-info charset) 14 plist)) |
219 | ||
220 | (defmacro make-char (charset &optional c1 c2) | |
f98e2797 KH |
221 | "Return a character of CHARSET and position-codes CODE1 and CODE2. |
222 | CODE1 and CODE2 are optional, but if you don't supply | |
900dc6e3 | 223 | sufficient position-codes, return a generic character which stands for |
f98e2797 | 224 | all characters or group of characters in the character sets. |
a73a8c89 | 225 | A generic character can be used to index a char table (e.g. syntax-table)." |
4ed46869 KH |
226 | (if (quoted-symbol-p charset) |
227 | `(make-char-internal ,(charset-id (nth 1 charset)) ,c1 ,c2) | |
228 | `(make-char-internal (charset-id ,charset) ,c1 ,c2))) | |
229 | ||
69eba008 | 230 | (defmacro charset-list () |
900dc6e3 KH |
231 | "Return list of charsets ever defined. |
232 | ||
233 | This macro is provided for backward compatibility. | |
234 | Now we have the variable `charset-list'." | |
13d5617d KH |
235 | 'charset-list) |
236 | ||
237 | (defsubst generic-char-p (char) | |
238 | "Return t if and only if CHAR is a generic character. | |
239 | See also the documentation of make-char." | |
240 | (let ((l (split-char char))) | |
241 | (and (or (= (nth 1 l) 0) (eq (nth 2 l) 0)) | |
242 | (not (eq (car l) 'composition))))) | |
69eba008 | 243 | |
8057896b | 244 | ;; Coding system staffs |
4ed46869 | 245 | |
8057896b | 246 | ;; Coding system is a symbol that has the property `coding-system'. |
4ed46869 | 247 | ;; |
8057896b KH |
248 | ;; The value of the property `coding-system' is a vector of the |
249 | ;; following format: | |
250 | ;; [TYPE MNEMONIC DOC-STRING NOT-USED-NOW FLAGS] | |
251 | ;; We call this vector as coding-spec. See comments in src/coding.c | |
252 | ;; for more detail. The property value may be another coding system, | |
253 | ;; in which case, the coding-spec should be taken from that | |
254 | ;; coding-system. The 4th element NOT-USED-NOW is kept just for | |
255 | ;; backward compatibility with old version of Mule. | |
256 | ||
257 | (defconst coding-spec-type-idx 0) | |
258 | (defconst coding-spec-mnemonic-idx 1) | |
259 | (defconst coding-spec-doc-string-idx 2) | |
260 | (defconst coding-spec-flags-idx 4) | |
261 | ||
262 | ;; Coding system may have proerpty `eol-type'. The value of the | |
263 | ;; property `eol-type' is integer 0..2 or a vector of three coding | |
264 | ;; systems. The integer value 0, 1, and 2 indicate the format of | |
4ed46869 KH |
265 | ;; end-of-line LF, CRLF, and CR respectively. The vector value |
266 | ;; indicates that the format of end-of-line should be detected | |
8057896b KH |
267 | ;; automatically. Nth element of the vector is the subsidiary coding |
268 | ;; system whose `eol-type' property is N. | |
4ed46869 | 269 | ;; |
8057896b KH |
270 | ;; Coding system may also have properties `post-read-conversion' and |
271 | ;; `pre-write-conversion. Values of these properties are functions. | |
4ed46869 KH |
272 | ;; |
273 | ;; The function in `post-read-conversion' is called after some text is | |
8057896b | 274 | ;; inserted and decoded along the coding system and before any |
4ed46869 KH |
275 | ;; functions in `after-insert-functions' are called. The arguments to |
276 | ;; this function is the same as those of a function in | |
277 | ;; `after-insert-functions', i.e. LENGTH of a text while putting point | |
278 | ;; at the head of the text to be decoded | |
279 | ;; | |
280 | ;; The function in `pre-write-conversion' is called after all | |
281 | ;; functions in `write-region-annotate-functions' and | |
282 | ;; `buffer-file-format' are called, and before the text is encoded by | |
8057896b | 283 | ;; the coding system. The arguments to this function is the same as |
4ed46869 KH |
284 | ;; those of a function in `write-region-annotate-functions', i.e. FROM |
285 | ;; and TO specifying region of a text. | |
286 | ||
8057896b KH |
287 | ;; Return Nth element of coding-spec of CODING-SYSTEM. |
288 | (defun coding-system-spec-ref (coding-system n) | |
289 | (check-coding-system coding-system) | |
290 | (let ((vec (coding-system-spec coding-system))) | |
291 | (and vec (aref vec n)))) | |
4ed46869 | 292 | |
4ed46869 | 293 | (defun coding-system-type (coding-system) |
6e9722b0 | 294 | "Return TYPE element in coding-spec of CODING-SYSTEM." |
8057896b | 295 | (coding-system-spec-ref coding-system coding-spec-type-idx)) |
4ed46869 | 296 | |
4ed46869 | 297 | (defun coding-system-mnemonic (coding-system) |
8057896b KH |
298 | "Return MNEMONIC element in coding-spec of CODING-SYSTEM." |
299 | (or (coding-system-spec-ref coding-system coding-spec-mnemonic-idx) | |
300 | ?-)) | |
4ed46869 | 301 | |
8057896b KH |
302 | (defun coding-system-doc-string (coding-system) |
303 | "Return DOC-STRING element in coding-spec of CODING-SYSTEM." | |
304 | (coding-system-spec-ref coding-system coding-spec-doc-string-idx)) | |
4ed46869 | 305 | |
4ed46869 | 306 | (defun coding-system-flags (coding-system) |
8057896b KH |
307 | "Return FLAGS element in coding-spec of CODING-SYSTEM." |
308 | (coding-system-spec-ref coding-system coding-spec-flags-idx)) | |
4ed46869 | 309 | |
8057896b KH |
310 | (defun coding-system-eol-type (coding-system) |
311 | "Return eol-type property of CODING-SYSTEM." | |
69eba008 | 312 | (check-coding-system coding-system) |
4ed46869 KH |
313 | (and coding-system |
314 | (or (get coding-system 'eol-type) | |
8057896b | 315 | (coding-system-eol-type (get coding-system 'coding-system))))) |
4ed46869 | 316 | |
6e9722b0 KH |
317 | (defun coding-system-category (coding-system) |
318 | "Return coding category of CODING-SYSTEM." | |
319 | (and coding-system | |
320 | (symbolp coding-system) | |
321 | (or (get coding-system 'coding-category) | |
322 | (coding-system-category (get coding-system 'coding-system))))) | |
323 | ||
bd882697 KH |
324 | (defun coding-system-parent (coding-system) |
325 | "Return parent of CODING-SYSTEM." | |
326 | (let ((parent (get coding-system 'parent-coding-system))) | |
327 | (and parent | |
328 | (or (coding-system-parent parent) | |
329 | parent)))) | |
330 | ||
6e9722b0 KH |
331 | ;; Make subsidiary coding systems (eol-type variants) of CODING-SYSTEM. |
332 | (defun make-subsidiary-coding-system (coding-system) | |
8057896b KH |
333 | (let ((subsidiaries (vector (intern (format "%s-unix" coding-system)) |
334 | (intern (format "%s-dos" coding-system)) | |
335 | (intern (format "%s-mac" coding-system)))) | |
336 | (i 0)) | |
337 | (while (< i 3) | |
6e9722b0 | 338 | (put (aref subsidiaries i) 'coding-system coding-system) |
8057896b KH |
339 | (put (aref subsidiaries i) 'eol-type i) |
340 | (put (aref subsidiaries i) 'eol-variant t) | |
341 | (setq i (1+ i))) | |
342 | subsidiaries)) | |
4ed46869 | 343 | |
8057896b KH |
344 | (defun make-coding-system (coding-system type mnemonic doc-string |
345 | &optional flags) | |
4ed46869 | 346 | "Define a new CODING-SYSTEM (symbol). |
8057896b KH |
347 | Remaining arguments are TYPE, MNEMONIC, DOC-STRING, and FLAGS (optional) which |
348 | construct a coding-spec of CODING-SYSTEM in the following format: | |
349 | [TYPE MNEMONIC DOC-STRING nil FLAGS] | |
4ed46869 KH |
350 | TYPE is an integer value indicating the type of coding-system as follows: |
351 | 0: Emacs internal format, | |
352 | 1: Shift-JIS (or MS-Kanji) used mainly on Japanese PC, | |
353 | 2: ISO-2022 including many variants, | |
354 | 3: Big5 used mainly on Chinese PC, | |
355 | 4: private, CCL programs provide encoding/decoding algorithm. | |
356 | MNEMONIC is a character to be displayed on mode line for the coding-system. | |
8057896b | 357 | DOC-STRING is a documentation string for the coding-system. |
4ed46869 | 358 | FLAGS specifies more precise information of each TYPE. |
d6d6d592 | 359 | |
4ed46869 KH |
360 | If TYPE is 2 (ISO-2022), FLAGS should be a list of: |
361 | CHARSET0, CHARSET1, CHARSET2, CHARSET3, SHORT-FORM, | |
362 | ASCII-EOL, ASCII-CNTL, SEVEN, LOCKING-SHIFT, SINGLE-SHIFT, | |
d6d6d592 | 363 | USE-ROMAN, USE-OLDJIS, NO-ISO6429, INIT-BOL, DESIGNATION-BOL, |
850101ed | 364 | SAFE, ACCEPT-LATIN-EXTRA-CODE. |
4ed46869 KH |
365 | CHARSETn are character sets initially designated to Gn graphic registers. |
366 | If CHARSETn is nil, Gn is never used. | |
367 | If CHARSETn is t, Gn can be used but nothing designated initially. | |
368 | If CHARSETn is a list of character sets, those character sets are | |
369 | designated to Gn on output, but nothing designated to Gn initially. | |
370 | SHORT-FORM non-nil means use short designation sequence on output. | |
371 | ASCII-EOL non-nil means designate ASCII to g0 at end of line on output. | |
372 | ASCII-CNTL non-nil means designate ASCII to g0 before control codes and | |
373 | SPACE on output. | |
374 | SEVEN non-nil means use 7-bit code only on output. | |
375 | LOCKING-SHIFT non-nil means use locking-shift. | |
376 | SINGLE-SHIFT non-nil means use single-shift. | |
377 | USE-ROMAN non-nil means designate JIS0201-1976-Roman instead of ASCII. | |
378 | USE-OLDJIS non-nil means designate JIS0208-1976 instead of JIS0208-1983. | |
379 | NO-ISO6429 non-nil means not use ISO6429's direction specification. | |
69eba008 KH |
380 | INIT-BOL non-nil means any designation state is assumed to be reset |
381 | to initial at each beginning of line on output. | |
382 | DESIGNATION-BOL non-nil means designation sequences should be placed | |
383 | at beginning of line on output. | |
d6d6d592 KH |
384 | SAFE non-nil means convert unexpected characters to `?' on output. |
385 | Unexpected characters are what not specified in CHARSETn directly. | |
850101ed RS |
386 | ACCEPT-LATIN-EXTRA-CODE non-nil means code-detection routine accepts |
387 | a code specified in `latin-extra-code-table' (which see) as a valid | |
388 | code of the coding system. | |
d6d6d592 | 389 | |
4ed46869 KH |
390 | If TYPE is 4 (private), FLAGS should be a cons of CCL programs, |
391 | for encoding and decoding. See the documentation of CCL for more detail." | |
392 | ||
393 | ;; At first, set a value of `coding-system' property. | |
6e9722b0 KH |
394 | (let ((coding-spec (make-vector 5 nil)) |
395 | coding-category) | |
8057896b KH |
396 | (if (or (not (integerp type)) (< type 0) (> type 4)) |
397 | (error "TYPE argument must be 0..4")) | |
398 | (if (or (not (integerp mnemonic)) (<= mnemonic ? ) (> mnemonic 127)) | |
399 | (error "MNEMONIC arguemnt must be a printable character.")) | |
400 | (aset coding-spec 0 type) | |
401 | (aset coding-spec 1 mnemonic) | |
402 | (aset coding-spec 2 (if (stringp doc-string) doc-string "")) | |
403 | (aset coding-spec 3 nil) ; obsolete element | |
6e9722b0 KH |
404 | (cond ((= type 0) |
405 | (setq coding-category 'coding-category-emacs-mule)) | |
406 | ((= type 1) | |
407 | (setq coding-category 'coding-category-sjis)) | |
408 | ((= type 2) ; ISO2022 | |
4ed46869 | 409 | (let ((i 0) |
6e9722b0 KH |
410 | (vec (make-vector 32 nil)) |
411 | (no-initial-designation t) | |
412 | (g1-designation nil)) | |
4ed46869 KH |
413 | (while (< i 4) |
414 | (let ((charset (car flags))) | |
6e9722b0 KH |
415 | (if (and no-initial-designation |
416 | (> i 0) | |
417 | (or (charsetp charset) | |
418 | (and (consp charset) | |
419 | (charsetp (car charset))))) | |
420 | (setq no-initial-designation nil)) | |
421 | (if (charsetp charset) | |
422 | (if (= i 1) (setq g1-designation charset)) | |
423 | (if (consp charset) | |
424 | (let ((tail charset) | |
425 | elt) | |
426 | (while tail | |
427 | (setq elt (car tail)) | |
69eba008 KH |
428 | (or (not elt) (eq elt t) (charsetp elt) |
429 | (error "Invalid charset: %s" elt)) | |
6e9722b0 KH |
430 | (setq tail (cdr tail))) |
431 | (setq g1-designation (car charset))) | |
432 | (if (and charset (not (eq charset t))) | |
433 | (error "Invalid charset: %s" charset)))) | |
4ed46869 KH |
434 | (aset vec i charset)) |
435 | (setq flags (cdr flags) i (1+ i))) | |
436 | (while (and (< i 32) flags) | |
437 | (aset vec i (car flags)) | |
438 | (setq flags (cdr flags) i (1+ i))) | |
6e9722b0 KH |
439 | (aset coding-spec 4 vec) |
440 | (if no-initial-designation | |
441 | (put coding-system 'no-initial-designation t)) | |
442 | (setq coding-category | |
443 | (if (aref vec 8) ; Use locking-shift. | |
dc64cd19 KH |
444 | (or (and (aref vec 7) 'coding-category-iso-7-else) |
445 | 'coding-category-iso-8-else) | |
6e9722b0 KH |
446 | (if (aref vec 7) ; 7-bit only. |
447 | (if (aref vec 9) ; Use single-shift. | |
dc64cd19 | 448 | 'coding-category-iso-7-else |
6e9722b0 KH |
449 | 'coding-category-iso-7) |
450 | (if no-initial-designation | |
dc64cd19 | 451 | 'coding-category-iso-8-else |
6e9722b0 KH |
452 | (if (and (charsetp g1-designation) |
453 | (= (charset-dimension g1-designation) 2)) | |
454 | 'coding-category-iso-8-2 | |
455 | 'coding-category-iso-8-1))))))) | |
456 | ((= type 3) | |
457 | (setq coding-category 'coding-category-big5)) | |
458 | ((= type 4) ; private | |
459 | (setq coding-category 'coding-category-binary) | |
4ed46869 KH |
460 | (if (and (consp flags) |
461 | (vectorp (car flags)) | |
462 | (vectorp (cdr flags))) | |
8057896b | 463 | (aset coding-spec 4 flags) |
6e9722b0 KH |
464 | (error "Invalid FLAGS argument for TYPE 4 (CCL)")))) |
465 | (put coding-system 'coding-system coding-spec) | |
466 | (put coding-system 'coding-category coding-category) | |
467 | (put coding-category 'coding-systems | |
468 | (cons coding-system (get coding-category 'coding-systems)))) | |
4ed46869 KH |
469 | |
470 | ;; Next, set a value of `eol-type' property. The value is a vector | |
6e9722b0 | 471 | ;; of subsidiary coding systems, each corresponds to a coding system |
4ed46869 | 472 | ;; for the detected end-of-line format. |
8057896b KH |
473 | (put coding-system 'eol-type |
474 | (if (<= type 3) | |
6e9722b0 | 475 | (make-subsidiary-coding-system coding-system) |
8057896b KH |
476 | 0))) |
477 | ||
a42763dc | 478 | (defun define-coding-system-alias (alias coding-system) |
358d28fb | 479 | "Define ALIAS as an alias for coding system CODING-SYSTEM." |
8057896b | 480 | (check-coding-system coding-system) |
6e9722b0 KH |
481 | (let ((parent (coding-system-parent coding-system))) |
482 | (if parent | |
483 | (setq coding-system parent))) | |
8057896b | 484 | (put alias 'coding-system coding-system) |
6e9722b0 KH |
485 | (put alias 'parent-coding-system coding-system) |
486 | (put coding-system 'alias-coding-systems | |
487 | (cons alias (get coding-system 'alias-coding-systems))) | |
488 | (let ((eol-variants (coding-system-eol-type coding-system)) | |
489 | subsidiaries) | |
490 | (if (vectorp eol-variants) | |
491 | (let ((i 0)) | |
492 | (setq subsidiaries (make-subsidiary-coding-system alias)) | |
493 | (while (< i 3) | |
494 | (put (aref subsidiaries i) 'parent-coding-system | |
495 | (aref eol-variants i)) | |
496 | (put (aref eol-variants i) 'alias-coding-systems | |
497 | (cons (aref subsidiaries i) (get (aref eol-variants i) | |
498 | 'alias-coding-systems))) | |
499 | (setq i (1+ i))))))) | |
4ed46869 KH |
500 | |
501 | (defun set-buffer-file-coding-system (coding-system &optional force) | |
358d28fb RS |
502 | "Set the file coding-system of the current buffer to CODING-SYSTEM. |
503 | This means that when you save the buffer, it will be converted | |
504 | according to CODING-SYSTEM. For a list of possible values of CODING-SYSTEM, | |
505 | use \\[list-coding-systems]. | |
506 | ||
507 | If the buffer's previous file coding-system value specifies end-of-line | |
508 | conversion, and CODING-SYSTEM does not specify one, CODING-SYSTEM is | |
509 | merged with the already-specified end-of-line conversion. | |
510 | However, if the optional prefix argument FORCE is non-nil, | |
511 | them CODING-SYSTEM is used exactly as specified." | |
4ed46869 KH |
512 | (interactive "zBuffer-file-coding-system: \nP") |
513 | (check-coding-system coding-system) | |
514 | (if (null force) | |
8057896b KH |
515 | (let ((x (coding-system-eol-type buffer-file-coding-system)) |
516 | (y (coding-system-eol-type coding-system))) | |
4ed46869 KH |
517 | (if (and (numberp x) (>= x 0) (<= x 2) (vectorp y)) |
518 | (setq coding-system (aref y x))))) | |
519 | (setq buffer-file-coding-system coding-system) | |
520 | (set-buffer-modified-p t) | |
521 | (force-mode-line-update)) | |
522 | ||
358d28fb RS |
523 | (defvar default-terminal-coding-system nil |
524 | "Default value for the terminal coding system. | |
525 | This is normally set according to the selected language environment. | |
526 | See also the command `set-terminal-coding-system'.") | |
527 | ||
df100398 KH |
528 | (defun set-terminal-coding-system (coding-system) |
529 | "Set coding system of your terminal to CODING-SYSTEM. | |
358d28fb RS |
530 | All text output to the terminal will be encoded |
531 | with the specified coding system. | |
532 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. | |
533 | The default is determined by the selected language environment | |
534 | or by the previous use of this command." | |
535 | (interactive | |
536 | (list (read-coding-system | |
537 | (format "Coding system for terminal display (default, %s): " | |
538 | (if (and (not (terminal-coding-system)) | |
539 | default-terminal-coding-system) | |
540 | default-terminal-coding-system))))) | |
541 | (if (and (not coding-system) | |
542 | (not (terminal-coding-system))) | |
543 | (setq coding-system default-terminal-coding-system)) | |
544 | (if coding-system | |
545 | (setq default-terminal-coding-system coding-system)) | |
df100398 KH |
546 | (set-terminal-coding-system-internal coding-system) |
547 | (redraw-frame (selected-frame))) | |
548 | ||
358d28fb RS |
549 | (defvar default-keyboard-coding-system nil |
550 | "Default value of the keyboard coding system. | |
551 | This is normally set according to the selected language environment. | |
552 | See also the command `set-keyboard-coding-system'.") | |
553 | ||
df100398 | 554 | (defun set-keyboard-coding-system (coding-system) |
358d28fb RS |
555 | "Set coding system for keyboard input to CODING-SYSTEM. |
556 | In addition, this command enables Encoded-kbd minor mode. | |
557 | \(If CODING-SYSTEM is nil, Encoded-bkd mode is turned off.) | |
558 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. | |
559 | The default is determined by the selected language environment | |
560 | or by the previous use of this command." | |
561 | (interactive | |
562 | (list (read-coding-system | |
563 | (format "Coding system for keyboard input (default, %s): " | |
564 | (if (and (not (keyboard-coding-system)) | |
565 | default-keyboard-coding-system) | |
566 | default-keyboard-coding-system))))) | |
567 | (if (and (not coding-system) | |
568 | (not (keyboard-coding-system))) | |
569 | (setq coding-system default-keyboard-coding-system)) | |
570 | (if coding-system | |
571 | (setq default-keyboard-coding-system coding-system)) | |
df100398 KH |
572 | (set-keyboard-coding-system-internal coding-system) |
573 | (encoded-kbd-mode (if coding-system 1 0))) | |
574 | ||
575 | (defun set-buffer-process-coding-system (decoding encoding) | |
358d28fb | 576 | "Set coding systems for the process associated with the current buffer. |
df100398 | 577 | DECODING is the coding system to be used to decode input from the process, |
358d28fb RS |
578 | ENCODING is the coding system to be used to encode output to the process. |
579 | ||
580 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]." | |
4ed46869 KH |
581 | (interactive |
582 | "zCoding-system for process input: \nzCoding-system for process output: ") | |
583 | (let ((proc (get-buffer-process (current-buffer)))) | |
584 | (if (null proc) | |
585 | (error "no process") | |
df100398 KH |
586 | (check-coding-system decoding) |
587 | (check-coding-system encoding) | |
588 | (set-process-coding-system proc decoding encoding))) | |
4ed46869 KH |
589 | (force-mode-line-update)) |
590 | ||
4ed46869 KH |
591 | (defun set-coding-priority (arg) |
592 | "Set priority of coding-category according to LIST. | |
593 | LIST is a list of coding-categories ordered by priority." | |
594 | (let (l) | |
595 | ;; Put coding-categories listed in ARG to L while checking the | |
596 | ;; validity. We assume that `coding-category-list' contains whole | |
597 | ;; coding-categories. | |
598 | (while arg | |
599 | (if (null (memq (car arg) coding-category-list)) | |
600 | (error "Invalid element in argument: %s" (car arg))) | |
601 | (setq l (cons (car arg) l)) | |
602 | (setq arg (cdr arg))) | |
603 | ;; Put coding-category not listed in ARG to L. | |
604 | (while coding-category-list | |
605 | (if (null (memq (car coding-category-list) l)) | |
606 | (setq l (cons (car coding-category-list) l))) | |
607 | (setq coding-category-list (cdr coding-category-list))) | |
608 | ;; Update `coding-category-list' and return it. | |
609 | (setq coding-category-list (nreverse l)))) | |
610 | ||
611 | ;;; FILE I/O | |
612 | ||
613 | ;; Set buffer-file-coding-system of the current buffer after some text | |
614 | ;; is inserted. | |
615 | (defun after-insert-file-set-buffer-file-coding-system (inserted) | |
616 | (if last-coding-system-used | |
617 | (let ((coding-system | |
618 | (find-new-buffer-file-coding-system last-coding-system-used)) | |
619 | (modified-p (buffer-modified-p))) | |
620 | (if coding-system | |
621 | (set-buffer-file-coding-system coding-system)) | |
622 | (set-buffer-modified-p modified-p))) | |
623 | nil) | |
624 | ||
625 | (setq after-insert-file-functions | |
626 | (cons 'after-insert-file-set-buffer-file-coding-system | |
627 | after-insert-file-functions)) | |
628 | ||
8057896b | 629 | ;; The coding-spec and eol-type of coding-system returned is decided |
4ed46869 KH |
630 | ;; independently in the following order. |
631 | ;; 1. That of buffer-file-coding-system locally bound. | |
632 | ;; 2. That of CODING. | |
633 | ||
634 | (defun find-new-buffer-file-coding-system (coding) | |
635 | "Return a coding system for a buffer when a file of CODING is inserted. | |
a73a8c89 KH |
636 | The local variable `buffer-file-coding-system' of the current buffer |
637 | is set to the returned value. | |
df100398 | 638 | Return nil if there's no need of setting new buffer-file-coding-system." |
4ed46869 KH |
639 | (let (local-coding local-eol |
640 | found-eol | |
641 | new-coding new-eol) | |
642 | (if (null coding) | |
643 | ;; Nothing found about coding. | |
644 | nil | |
645 | ||
646 | ;; Get information of the current local value of | |
647 | ;; `buffer-file-coding-system' in LOCAL-EOL and LOCAL-CODING. | |
648 | (if (local-variable-p 'buffer-file-coding-system) | |
649 | ;; Something already set locally. | |
650 | (progn | |
8057896b | 651 | (setq local-eol (coding-system-eol-type buffer-file-coding-system)) |
4ed46869 KH |
652 | (if (null (numberp local-eol)) |
653 | ;; But eol-type is not yet set. | |
654 | (setq local-eol nil)) | |
655 | (if (null (eq (coding-system-type buffer-file-coding-system) t)) | |
13d5617d | 656 | ;; This is not `undecided'. |
4ed46869 KH |
657 | (progn |
658 | (setq local-coding buffer-file-coding-system) | |
659 | (while (symbolp (get local-coding 'coding-system)) | |
660 | (setq local-coding (get local-coding 'coding-system)))) | |
661 | ))) | |
662 | ||
663 | (if (and local-eol local-coding) | |
664 | ;; The current buffer has already set full coding-system, we | |
665 | ;; had better not change it. | |
666 | nil | |
667 | ||
8057896b | 668 | (setq found-eol (coding-system-eol-type coding)) |
4ed46869 KH |
669 | (if (null (numberp found-eol)) |
670 | ;; But eol-type is not found. | |
671 | (setq found-eol nil)) | |
4ed46869 KH |
672 | |
673 | ;; The local setting takes precedence over the found one. | |
674 | (setq new-coding (or local-coding coding)) | |
675 | (setq new-eol (or local-eol found-eol)) | |
676 | (if (and (numberp new-eol) | |
8057896b | 677 | (vectorp (coding-system-eol-type new-coding))) |
4ed46869 | 678 | (setq new-coding |
8057896b | 679 | (aref (coding-system-eol-type new-coding) new-eol))) |
4ed46869 KH |
680 | new-coding)))) |
681 | ||
fe831d33 GV |
682 | (defun modify-coding-system-alist (target-type regexp coding-system) |
683 | "Modify one of look up tables for finding a coding system on I/O operation. | |
8c453b46 RS |
684 | There are three of such tables, `file-coding-system-alist', |
685 | `process-coding-system-alist', and `network-coding-system-alist'. | |
fe831d33 GV |
686 | |
687 | TARGET-TYPE specifies which of them to modify. | |
8c453b46 RS |
688 | If it is `file', it affects `file-coding-system-alist' (which see). |
689 | If it is `process', it affects `process-coding-system-alist' (which see). | |
690 | If it is `network', it affects `network-codign-system-alist' (which see). | |
fe831d33 GV |
691 | |
692 | REGEXP is a regular expression matching a target of I/O operation. | |
693 | The target is a file name if TARGET-TYPE is `file', a program name if | |
694 | TARGET-TYPE is `process', or a network service name or a port number | |
695 | to connect to if TARGET-TYPE is `network'. | |
696 | ||
697 | CODING-SYSTEM is a coding system to perform code conversion on the I/O | |
8c453b46 RS |
698 | operation, or a cons cell (DECODING . ENCODING) specifying the coding systems |
699 | for decoding and encoding respectively, | |
700 | or a function symbol which, when called, returns such a cons cell." | |
fe831d33 GV |
701 | (or (memq target-type '(file process network)) |
702 | (error "Invalid target type: %s" target-type)) | |
703 | (or (stringp regexp) | |
704 | (and (eq target-type 'network) (integerp regexp)) | |
705 | (error "Invalid regular expression: %s" regexp)) | |
706 | (if (symbolp coding-system) | |
707 | (if (not (fboundp coding-system)) | |
708 | (progn | |
709 | (check-coding-system coding-system) | |
710 | (setq coding-system (cons coding-system coding-system)))) | |
711 | (check-coding-system (car coding-system)) | |
712 | (check-coding-system (cdr coding-system))) | |
713 | (cond ((eq target-type 'file) | |
714 | (let ((slot (assoc regexp file-coding-system-alist))) | |
715 | (if slot | |
716 | (setcdr slot coding-system) | |
717 | (setq file-coding-system-alist | |
718 | (cons (cons regexp coding-system) | |
719 | file-coding-system-alist))))) | |
720 | ((eq target-type 'process) | |
721 | (let ((slot (assoc regexp process-coding-system-alist))) | |
722 | (if slot | |
723 | (setcdr slot coding-system) | |
724 | (setq process-coding-system-alist | |
725 | (cons (cons regexp coding-system) | |
726 | process-coding-system-alist))))) | |
727 | (t | |
728 | (let ((slot (assoc regexp network-coding-system-alist))) | |
729 | (if slot | |
730 | (setcdr slot coding-system) | |
731 | (setq network-coding-system-alist | |
732 | (cons (cons regexp coding-system) | |
733 | network-coding-system-alist))))))) | |
734 | ||
a73a8c89 KH |
735 | (defun make-unification-table (&rest args) |
736 | "Make a unification table (char table) from arguments. | |
13d5617d KH |
737 | Each argument is a list of the form (FROM . TO), |
738 | where FROM is a character to be unified to TO. | |
739 | ||
740 | FROM can be a generic character (see make-char). In this case, TO is | |
741 | a generic character containing the same number of charcters or a | |
742 | oridinal character. If FROM and TO are both generic characters, all | |
743 | characters belonging to FROM are unified to characters belonging to TO | |
744 | without changing their position code(s)." | |
a73a8c89 KH |
745 | (let ((table (make-char-table 'character-unification-table)) |
746 | revlist) | |
747 | (while args | |
748 | (let ((elts (car args))) | |
749 | (while elts | |
13d5617d KH |
750 | (let* ((from (car (car elts))) |
751 | (from-i 0) ; degree of freedom of FROM | |
752 | (from-rev (nreverse (split-char from))) | |
753 | (to (cdr (car elts))) | |
754 | (to-i 0) ; degree of freedom of TO | |
755 | (to-rev (nreverse (split-char to)))) | |
756 | ;; Check numbers of heading 0s in FROM-REV and TO-REV. | |
757 | (while (eq (car from-rev) 0) | |
758 | (setq from-i (1+ from-i) from-rev (cdr from-rev))) | |
759 | (while (eq (car to-rev) 0) | |
760 | (setq to-i (1+ to-i) to-rev (cdr to-rev))) | |
761 | (if (and (/= from-i to-i) (/= to-i 0)) | |
762 | (error "Invalid character pair (%d . %d)" from to)) | |
763 | ;; If we have already unified TO to TO-ALT, FROM should | |
764 | ;; also be unified to TO-ALT. But, this is only if TO is | |
765 | ;; a generic character or TO-ALT is not a generic | |
766 | ;; character. | |
767 | (let ((to-alt (aref table to))) | |
768 | (if (and to-alt | |
769 | (or (> to-i 0) (not (generic-char-p to-alt)))) | |
770 | (setq to to-alt))) | |
771 | (if (> from-i 0) | |
772 | (set-char-table-default table from to) | |
773 | (aset table from to)) | |
a73a8c89 KH |
774 | ;; If we have already unified some chars to FROM, they |
775 | ;; should also be unified to TO. | |
776 | (let ((l (assq from revlist))) | |
777 | (if l | |
778 | (let ((ch (car l))) | |
779 | (setcar l to) | |
780 | (setq l (cdr l)) | |
781 | (while l | |
782 | (aset table ch to) | |
783 | (setq l (cdr l)) )))) | |
784 | ;; Now update REVLIST. | |
785 | (let ((l (assq to revlist))) | |
786 | (if l | |
787 | (setcdr l (cons from (cdr l))) | |
788 | (setq revlist (cons (list to from) revlist))))) | |
789 | (setq elts (cdr elts)))) | |
790 | (setq args (cdr args))) | |
791 | ;; Return TABLE just created. | |
792 | table)) | |
793 | ||
69eba008 KH |
794 | ;;; Initialize some variables. |
795 | ||
796 | (put 'use-default-ascent 'char-table-extra-slots 0) | |
797 | (setq use-default-ascent (make-char-table 'use-default-ascent)) | |
d6d6d592 KH |
798 | (put 'ignore-relative-composition 'char-table-extra-slots 0) |
799 | (setq ignore-relative-composition | |
800 | (make-char-table 'ignore-relative-composition)) | |
69eba008 KH |
801 | |
802 | ;;; | |
4ed46869 KH |
803 | (provide 'mule) |
804 | ||
805 | ;;; mule.el ends here |