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 | ||
d2542b42 | 27 | (defconst mule-version "5.0 (SAKAKI)" "\ |
4ed46869 KH |
28 | Version number and name of this version of MULE (multilingual environment).") |
29 | ||
35554641 | 30 | (defconst mule-version-date "1999.12.7" "\ |
4ed46869 KH |
31 | Distribution date of this version of MULE (multilingual environment).") |
32 | ||
33 | (defun load-with-code-conversion (fullname file &optional noerror nomessage) | |
0f69cb38 KH |
34 | "Execute a file of Lisp code named FILE whose absolute name is FULLNAME. |
35 | The file contents are decoded before evaluation if necessary. | |
4ed46869 KH |
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. | |
88162676 RS |
48 | ;; So that we don't get completely screwed if the |
49 | ;; file is encoded in some complicated character set, | |
50 | ;; read it with real decoding, as a multibyte buffer, | |
51 | ;; even if this is a --unibyte Emacs session. | |
52 | (let ((default-major-mode 'fundamental-mode) | |
53 | (default-enable-multibyte-characters t)) | |
4ed46869 KH |
54 | ;; We can't use `generate-new-buffer' because files.el |
55 | ;; is not yet loaded. | |
56 | (get-buffer-create (generate-new-buffer-name " *load*")))) | |
db5cae4b SM |
57 | (load-in-progress t) |
58 | (source (save-match-data (string-match "\\.el\\'" fullname)))) | |
59 | (unless nomessage | |
60 | (if source | |
61 | (message "Loading %s (source)..." file) | |
62 | (message "Loading %s..." file))) | |
63 | (when purify-flag | |
64 | (setq preloaded-file-list (cons file preloaded-file-list))) | |
4ed46869 | 65 | (unwind-protect |
a6acd8a2 | 66 | (let ((load-file-name fullname) |
1c4cc63a | 67 | (set-auto-coding-for-load t) |
a6acd8a2 | 68 | (inhibit-file-name-operation nil)) |
4ed46869 KH |
69 | (save-excursion |
70 | (set-buffer buffer) | |
71 | (insert-file-contents fullname) | |
7d276780 EZ |
72 | ;; If the loaded file was inserted with no-conversion or |
73 | ;; raw-text coding system, make the buffer unibyte. | |
74 | ;; Otherwise, eval-buffer might try to interpret random | |
75 | ;; binary junk as multibyte characters. | |
76 | (if (and enable-multibyte-characters | |
77 | (or (eq (coding-system-type last-coding-system-used) 5) | |
78 | (eq last-coding-system-used 'no-conversion))) | |
79 | (set-buffer-multibyte nil)) | |
4ed46869 KH |
80 | ;; Make `kill-buffer' quiet. |
81 | (set-buffer-modified-p nil)) | |
0f69cb38 | 82 | ;; Have the original buffer current while we eval. |
88162676 RS |
83 | (eval-buffer buffer nil file |
84 | ;; If this Emacs is running with --unibyte, | |
85 | ;; convert multibyte strings to unibyte | |
86 | ;; after reading them. | |
ba74e833 | 87 | ;; (not default-enable-multibyte-characters) |
8dd08b5b | 88 | nil t |
ba74e833 | 89 | )) |
cfc70cdf RS |
90 | (let (kill-buffer-hook kill-buffer-query-functions) |
91 | (kill-buffer buffer))) | |
4ed46869 | 92 | (let ((hook (assoc file after-load-alist))) |
db5cae4b SM |
93 | (when hook |
94 | (mapcar (function eval) (cdr hook)))) | |
95 | (unless (or nomessage noninteractive) | |
96 | (if source | |
97 | (message "Loading %s (source)...done" file) | |
98 | (message "Loading %s...done" file))) | |
4ed46869 KH |
99 | t))) |
100 | ||
101 | ;; API (Application Program Interface) for charsets. | |
102 | ||
2e02a76f RS |
103 | ;; Return t if OBJ is a quoted symbol |
104 | ;; and the symbol is the name of a standard charset. | |
105 | (defsubst charset-quoted-standard-p (obj) | |
106 | (and (listp obj) (eq (car obj) 'quote) | |
107 | (symbolp (car-safe (cdr obj))) | |
108 | (let ((vector (get (car-safe (cdr obj)) 'charset))) | |
109 | (and (vectorp vector) | |
110 | (< (aref vector 0) 160))))) | |
4ed46869 KH |
111 | |
112 | (defsubst charsetp (object) | |
e4a7fdfc | 113 | "T if OBJECT is a charset." |
4ed46869 KH |
114 | (and (symbolp object) (vectorp (get object 'charset)))) |
115 | ||
116 | (defsubst charset-info (charset) | |
117 | "Return a vector of information of CHARSET. | |
118 | The elements of the vector are: | |
119 | CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION, | |
120 | LEADING-CODE-BASE, LEADING-CODE-EXT, | |
121 | ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE, | |
122 | REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION, | |
123 | PLIST, | |
124 | where | |
125 | CHARSET-ID (integer) is the identification number of the charset. | |
aa8bb645 KH |
126 | BYTES (integer) is the length of multi-byte form of a character in |
127 | the charset: one of 1, 2, 3, and 4. | |
4ed46869 KH |
128 | DIMENSION (integer) is the number of bytes to represent a character of |
129 | the charset: 1 or 2. | |
130 | CHARS (integer) is the number of characters in a dimension: 94 or 96. | |
4ed46869 KH |
131 | WIDTH (integer) is the number of columns a character in the charset |
132 | occupies on the screen: one of 0, 1, and 2. | |
133 | DIRECTION (integer) is the rendering direction of characters in the | |
08b4ace5 KH |
134 | charset when rendering. If 0, render from left to right, else |
135 | render from right to left. | |
4ed46869 KH |
136 | LEADING-CODE-BASE (integer) is the base leading-code for the |
137 | charset. | |
138 | LEADING-CODE-EXT (integer) is the extended leading-code for the | |
139 | charset. All charsets of less than 0xA0 has the value 0. | |
140 | ISO-FINAL-CHAR (character) is the final character of the | |
141 | corresponding ISO 2022 charset. | |
142 | ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked | |
143 | while encoding to variants of ISO 2022 coding system, one of the | |
144 | following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR). | |
145 | REVERSE-CHARSET (integer) is the charset which differs only in | |
146 | LEFT-TO-RIGHT value from the charset. If there's no such a | |
147 | charset, the value is -1. | |
148 | SHORT-NAME (string) is the short name to refer to the charset. | |
149 | LONG-NAME (string) is the long name to refer to the charset | |
150 | DESCRIPTION (string) is the description string of the charset. | |
151 | PLIST (property list) may contain any type of information a user | |
152 | want to put and get by functions `put-charset-property' and | |
153 | `get-charset-property' respectively." | |
154 | (get charset 'charset)) | |
155 | ||
40c81f74 PE |
156 | ;; It is better not to use backquote in this file, |
157 | ;; because that makes a bootstrapping problem | |
158 | ;; if you need to recompile all the Lisp files using interpreted code. | |
159 | ||
4ed46869 KH |
160 | (defmacro charset-id (charset) |
161 | "Return charset identification number of CHARSET." | |
2e02a76f | 162 | (if (charset-quoted-standard-p charset) |
4ed46869 | 163 | (aref (charset-info (nth 1 charset)) 0) |
40c81f74 | 164 | (list 'aref (list 'charset-info charset) 0))) |
4ed46869 KH |
165 | |
166 | (defmacro charset-bytes (charset) | |
900dc6e3 KH |
167 | "Return bytes of CHARSET. |
168 | See the function `charset-info' for more detail." | |
2e02a76f | 169 | (if (charset-quoted-standard-p charset) |
4ed46869 | 170 | (aref (charset-info (nth 1 charset)) 1) |
40c81f74 | 171 | (list 'aref (list 'charset-info charset) 1))) |
4ed46869 KH |
172 | |
173 | (defmacro charset-dimension (charset) | |
900dc6e3 KH |
174 | "Return dimension of CHARSET. |
175 | See the function `charset-info' for more detail." | |
2e02a76f | 176 | (if (charset-quoted-standard-p charset) |
4ed46869 | 177 | (aref (charset-info (nth 1 charset)) 2) |
40c81f74 | 178 | (list 'aref (list 'charset-info charset) 2))) |
4ed46869 KH |
179 | |
180 | (defmacro charset-chars (charset) | |
900dc6e3 KH |
181 | "Return character numbers contained in a dimension of CHARSET. |
182 | See the function `charset-info' for more detail." | |
2e02a76f | 183 | (if (charset-quoted-standard-p charset) |
4ed46869 | 184 | (aref (charset-info (nth 1 charset)) 3) |
40c81f74 | 185 | (list 'aref (list 'charset-info charset) 3))) |
4ed46869 KH |
186 | |
187 | (defmacro charset-width (charset) | |
900dc6e3 KH |
188 | "Return width (how many column occupied on a screen) of CHARSET. |
189 | See the function `charset-info' for more detail." | |
2e02a76f | 190 | (if (charset-quoted-standard-p charset) |
4ed46869 | 191 | (aref (charset-info (nth 1 charset)) 4) |
40c81f74 | 192 | (list 'aref (list 'charset-info charset) 4))) |
4ed46869 KH |
193 | |
194 | (defmacro charset-direction (charset) | |
900dc6e3 KH |
195 | "Return direction of CHARSET. |
196 | See the function `charset-info' for more detail." | |
2e02a76f | 197 | (if (charset-quoted-standard-p charset) |
4ed46869 | 198 | (aref (charset-info (nth 1 charset)) 5) |
40c81f74 | 199 | (list 'aref (list 'charset-info charset) 5))) |
4ed46869 KH |
200 | |
201 | (defmacro charset-iso-final-char (charset) | |
900dc6e3 KH |
202 | "Return final char of CHARSET. |
203 | See the function `charset-info' for more detail." | |
2e02a76f | 204 | (if (charset-quoted-standard-p charset) |
4ed46869 | 205 | (aref (charset-info (nth 1 charset)) 8) |
40c81f74 | 206 | (list 'aref (list 'charset-info charset) 8))) |
4ed46869 KH |
207 | |
208 | (defmacro charset-iso-graphic-plane (charset) | |
900dc6e3 KH |
209 | "Return graphic plane of CHARSET. |
210 | See the function `charset-info' for more detail." | |
2e02a76f | 211 | (if (charset-quoted-standard-p charset) |
4ed46869 | 212 | (aref (charset-info (nth 1 charset)) 9) |
40c81f74 | 213 | (list 'aref (list 'charset-info charset) 9))) |
4ed46869 KH |
214 | |
215 | (defmacro charset-reverse-charset (charset) | |
900dc6e3 KH |
216 | "Return reverse charset of CHARSET. |
217 | See the function `charset-info' for more detail." | |
2e02a76f | 218 | (if (charset-quoted-standard-p charset) |
4ed46869 | 219 | (aref (charset-info (nth 1 charset)) 10) |
40c81f74 | 220 | (list 'aref (list 'charset-info charset) 10))) |
4ed46869 KH |
221 | |
222 | (defmacro charset-short-name (charset) | |
900dc6e3 KH |
223 | "Return short name of CHARSET. |
224 | See the function `charset-info' for more detail." | |
2e02a76f | 225 | (if (charset-quoted-standard-p charset) |
4ed46869 | 226 | (aref (charset-info (nth 1 charset)) 11) |
40c81f74 | 227 | (list 'aref (list 'charset-info charset) 11))) |
4ed46869 KH |
228 | |
229 | (defmacro charset-long-name (charset) | |
900dc6e3 KH |
230 | "Return long name of CHARSET. |
231 | See the function `charset-info' for more detail." | |
2e02a76f | 232 | (if (charset-quoted-standard-p charset) |
4ed46869 | 233 | (aref (charset-info (nth 1 charset)) 12) |
40c81f74 | 234 | (list 'aref (list 'charset-info charset) 12))) |
4ed46869 KH |
235 | |
236 | (defmacro charset-description (charset) | |
0611934b | 237 | "Return description of CHARSET. |
900dc6e3 | 238 | See the function `charset-info' for more detail." |
2e02a76f | 239 | (if (charset-quoted-standard-p charset) |
4ed46869 | 240 | (aref (charset-info (nth 1 charset)) 13) |
40c81f74 | 241 | (list 'aref (list 'charset-info charset) 13))) |
4ed46869 KH |
242 | |
243 | (defmacro charset-plist (charset) | |
900dc6e3 KH |
244 | "Return list charset property of CHARSET. |
245 | See the function `charset-info' for more detail." | |
40c81f74 PE |
246 | (list 'aref |
247 | (if (charset-quoted-standard-p charset) | |
248 | (charset-info (nth 1 charset)) | |
249 | (list 'charset-info charset)) | |
250 | 14)) | |
4ed46869 KH |
251 | |
252 | (defun set-charset-plist (charset plist) | |
e8dd0160 | 253 | "Set CHARSET's property list to PLIST, and return PLIST." |
4ed46869 KH |
254 | (aset (charset-info charset) 14 plist)) |
255 | ||
d3e7e7cf EZ |
256 | (defun make-char (charset &optional code1 code2) |
257 | "Return a character of CHARSET whose position codes are CODE1 and CODE2. | |
f98e2797 | 258 | CODE1 and CODE2 are optional, but if you don't supply |
a9fb0b58 KH |
259 | sufficient position codes, return a generic character which stands for |
260 | all characters or group of characters in the character set. | |
d3e7e7cf EZ |
261 | A generic character can be used to index a char table (e.g. syntax-table). |
262 | If CODE1 or CODE2 are invalid (out of range), this function signals an error." | |
636799f2 | 263 | (make-char-internal (charset-id charset) code1 code2)) |
a6acd8a2 KH |
264 | |
265 | (put 'make-char 'byte-compile | |
266 | (function | |
267 | (lambda (form) | |
268 | (let ((charset (nth 1 form))) | |
269 | (if (charset-quoted-standard-p charset) | |
270 | (byte-compile-normal-call | |
271 | (cons 'make-char-internal | |
272 | (cons (charset-id (nth 1 charset)) (nthcdr 2 form)))) | |
273 | (byte-compile-normal-call | |
274 | (cons 'make-char-internal | |
275 | (cons (list 'charset-id charset) (nthcdr 2 form))))))))) | |
4ed46869 | 276 | |
d3675a42 | 277 | (defun charset-list () |
900dc6e3 KH |
278 | "Return list of charsets ever defined. |
279 | ||
d3675a42 | 280 | This function is provided for backward compatibility. |
900dc6e3 | 281 | Now we have the variable `charset-list'." |
d3675a42 KH |
282 | charset-list) |
283 | ||
13d5617d KH |
284 | (defsubst generic-char-p (char) |
285 | "Return t if and only if CHAR is a generic character. | |
286 | See also the documentation of make-char." | |
1d935669 KH |
287 | (and (>= char 0400) |
288 | (let ((l (split-char char))) | |
289 | (and (or (= (nth 1 l) 0) (eq (nth 2 l) 0)) | |
290 | (not (eq (car l) 'composition)))))) | |
69eba008 | 291 | |
cc57cc54 | 292 | (defun decode-char (ccs code-point &optional restriction) |
7f341749 | 293 | "Return character specified by coded character set CCS and CODE-POINT in it. |
cc57cc54 | 294 | Return nil if such a character is not supported. |
e76938e7 DL |
295 | Currently the only supported coded character set is `ucs' (ISO/IEC |
296 | 10646: Universal Multi-Octet Coded Character Set). | |
cc57cc54 KH |
297 | |
298 | Optional argument RESTRICTION specifies a way to map the pair of CCS | |
299 | and CODE-POINT to a chracter. Currently not supported and just ignored." | |
300 | (cond ((eq ccs 'ucs) | |
301 | (cond ((< code-point 128) | |
302 | code-point) | |
303 | ((< code-point 256) | |
304 | (make-char 'latin-iso8859-1 code-point)) | |
305 | ((< code-point #x2500) | |
306 | (setq code-point (- code-point #x0100)) | |
307 | (make-char 'mule-unicode-0100-24ff | |
308 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
309 | ((< code-point #x33ff) | |
310 | (setq code-point (- code-point #x2500)) | |
311 | (make-char 'mule-unicode-2500-33ff | |
312 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
313 | ((and (>= code-point #xe000) (< code-point #x10000)) | |
314 | (setq code-point (- code-point #xe000)) | |
315 | (make-char 'mule-unicode-e000-ffff | |
316 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
317 | )))) | |
318 | ||
319 | (defun encode-char (char ccs &optional restriction) | |
7f341749 | 320 | "Return code-point in coded character set CCS that corresponds to CHAR. |
cc57cc54 | 321 | Return nil if CHAR is not included in CCS. |
e76938e7 DL |
322 | Currently the only supported coded character set is `ucs' (ISO/IEC |
323 | 10646: Universal Multi-Octet Coded Character Set). | |
7f341749 KH |
324 | |
325 | CHAR should be in one of these charsets: | |
cc57cc54 KH |
326 | ascii, latin-iso8859-1, mule-unicode-0100-24ff, mule-unicode-2500-33ff, |
327 | mule-unicode-e000-ffff | |
328 | Otherwise, return nil. | |
329 | ||
330 | Optional argument RESTRICTION specifies a way to map CHAR to a | |
331 | code-point in CCS. Currently not supported and just ignored." | |
332 | (let* ((split (split-char char)) | |
333 | (charset (car split))) | |
334 | (cond ((eq ccs 'ucs) | |
335 | (cond ((eq charset 'ascii) | |
336 | char) | |
337 | ((eq charset 'latin-iso8859-1) | |
338 | (+ (nth 1 split) 128)) | |
339 | ((eq charset 'mule-unicode-0100-24ff) | |
340 | (+ #x0100 (+ (* (- (nth 1 split) 32) 96) | |
341 | (- (nth 2 split) 32)))) | |
342 | ((eq charset 'mule-unicode-2500-33ff) | |
343 | (+ #x2500 (+ (* (- (nth 1 split) 32) 96) | |
344 | (- (nth 2 split) 32)))) | |
345 | ((eq charset 'mule-unicode-e000-ffff) | |
346 | (+ #xe000 (+ (* (- (nth 1 split) 32) 96) | |
347 | (- (nth 2 split) 32))))))))) | |
348 | ||
0269ddfb | 349 | \f |
e76938e7 | 350 | ;; Coding system stuff |
4ed46869 | 351 | |
8057896b | 352 | ;; Coding system is a symbol that has the property `coding-system'. |
4ed46869 | 353 | ;; |
8057896b KH |
354 | ;; The value of the property `coding-system' is a vector of the |
355 | ;; following format: | |
d3675a42 | 356 | ;; [TYPE MNEMONIC DOC-STRING PLIST FLAGS] |
8057896b | 357 | ;; We call this vector as coding-spec. See comments in src/coding.c |
d3675a42 | 358 | ;; for more detail. |
8057896b KH |
359 | |
360 | (defconst coding-spec-type-idx 0) | |
361 | (defconst coding-spec-mnemonic-idx 1) | |
362 | (defconst coding-spec-doc-string-idx 2) | |
0269ddfb | 363 | (defconst coding-spec-plist-idx 3) |
8057896b KH |
364 | (defconst coding-spec-flags-idx 4) |
365 | ||
a6acd8a2 KH |
366 | ;; PLIST is a property list of a coding system. To share PLIST among |
367 | ;; alias coding systems, a coding system has PLIST in coding-spec | |
368 | ;; instead of having it in normal property list of Lisp symbol. | |
369 | ;; Here's a list of coding system properties currently being used. | |
0269ddfb KH |
370 | ;; |
371 | ;; o coding-category | |
372 | ;; | |
373 | ;; The value is a coding category the coding system belongs to. The | |
1b46a680 KH |
374 | ;; function `make-coding-system' sets this value automatically |
375 | ;; unless its argument PROPERTIES specifies this property. | |
4ed46869 | 376 | ;; |
0269ddfb | 377 | ;; o alias-coding-systems |
4ed46869 | 378 | ;; |
0269ddfb KH |
379 | ;; The value is a list of coding systems of the same alias group. The |
380 | ;; first element is the coding system made at first, which we call as | |
1b46a680 KH |
381 | ;; `base coding system'. The function `make-coding-system' sets this |
382 | ;; value automatically and `define-coding-system-alias' updates it. | |
0269ddfb KH |
383 | ;; |
384 | ;; o post-read-conversion | |
385 | ;; | |
386 | ;; The value is a function to call after some text is inserted and | |
387 | ;; decoded by the coding system itself and before any functions in | |
388 | ;; `after-insert-functions' are called. The arguments to this | |
389 | ;; function is the same as those of a function in | |
4ed46869 KH |
390 | ;; `after-insert-functions', i.e. LENGTH of a text while putting point |
391 | ;; at the head of the text to be decoded | |
392 | ;; | |
0269ddfb KH |
393 | ;; o pre-write-conversion |
394 | ;; | |
395 | ;; The value is a function to call after all functions in | |
396 | ;; `write-region-annotate-functions' and `buffer-file-format' are | |
397 | ;; called, and before the text is encoded by the coding system itself. | |
398 | ;; The arguments to this function is the same as those of a function | |
399 | ;; in `write-region-annotate-functions', i.e. FROM and TO specifying | |
400 | ;; region of a text. | |
401 | ;; | |
f967223b | 402 | ;; o translation-table-for-decode |
0269ddfb | 403 | ;; |
f967223b KH |
404 | ;; The value is a translation table to be applied on decoding. See |
405 | ;; the function `make-translation-table' for the format of translation | |
406 | ;; table. | |
0269ddfb | 407 | ;; |
f967223b | 408 | ;; o translation-table-for-encode |
0269ddfb | 409 | ;; |
f967223b | 410 | ;; The value is a translation table to be applied on encoding. |
a6acd8a2 | 411 | ;; |
c11a8f77 KH |
412 | ;; o safe-chars |
413 | ;; | |
414 | ;; The value is a char table. If a character has non-nil value in it, | |
415 | ;; the character is safely supported by the coding system. This | |
416 | ;; overrides the specification of safe-charsets. | |
417 | ||
a6acd8a2 KH |
418 | ;; o safe-charsets |
419 | ;; | |
420 | ;; The value is a list of charsets safely supported by the coding | |
421 | ;; system. The value t means that all charsets Emacs handles are | |
422 | ;; supported. Even if some charset is not in this list, it doesn't | |
423 | ;; mean that the charset can't be encoded in the coding system, | |
424 | ;; instead, it just means that some other receiver of a text encoded | |
425 | ;; in the coding system won't be able to handle that charset. | |
b25eef20 KH |
426 | ;; |
427 | ;; o mime-charset | |
428 | ;; | |
429 | ;; The value is a symbol of which name is `MIME-charset' parameter of | |
430 | ;; the coding system. | |
c76b5c99 | 431 | ;; |
1322c8ca KH |
432 | ;; o charset-origin-alist |
433 | ;; | |
434 | ;; The value is a list of this form: | |
435 | ;; (CHARSET EXTERNAL-CHARSET-NAME ENCODING-FUNCTION). | |
436 | ;; ENCODING-FUNCTION is a function to encode a character in CHARSET | |
437 | ;; to the code in EXTERNAL-CHARSET-NAME. The command what-cursor-position | |
438 | ;; uses this information of the buffer-file-coding-system. | |
caa85ad9 KH |
439 | ;; ENCODING-FUNCTION may be a translation table or a symbol whose |
440 | ;; property `translation-table' is a translation table. In these case, | |
441 | ;; the translation table is used to encode the character. | |
1322c8ca | 442 | ;; |
c76b5c99 KH |
443 | ;; o valid-codes (meaningful only for a coding system based on CCL) |
444 | ;; | |
445 | ;; The value is a list to indicate valid byte ranges of the encoded | |
446 | ;; file. Each element of the list is an integer or a cons of integer. | |
447 | ;; In the former case, the integer value is a valid byte code. In the | |
e8dd0160 | 448 | ;; latter case, the integers specifies the range of valid byte codes. |
c76b5c99 | 449 | |
0269ddfb KH |
450 | |
451 | ;; Return coding-spec of CODING-SYSTEM | |
452 | (defsubst coding-system-spec (coding-system) | |
453 | (get (check-coding-system coding-system) 'coding-system)) | |
4ed46869 | 454 | |
4ed46869 | 455 | (defun coding-system-type (coding-system) |
0269ddfb KH |
456 | "Return the coding type of CODING-SYSTEM. |
457 | A coding type is an integer value indicating the encoding method | |
458 | of CODING-SYSTEM. See the function `make-coding-system' for more detail." | |
459 | (aref (coding-system-spec coding-system) coding-spec-type-idx)) | |
4ed46869 | 460 | |
4ed46869 | 461 | (defun coding-system-mnemonic (coding-system) |
0269ddfb | 462 | "Return the mnemonic character of CODING-SYSTEM. |
6e2c8840 KH |
463 | The mnemonic character of a coding system is used in mode line |
464 | to indicate the coding system. If the arg is nil, return ?-." | |
465 | (let ((spec (coding-system-spec coding-system))) | |
466 | (if spec (aref spec coding-spec-mnemonic-idx) ?-))) | |
4ed46869 | 467 | |
8057896b | 468 | (defun coding-system-doc-string (coding-system) |
0269ddfb KH |
469 | "Return the documentation string for CODING-SYSTEM." |
470 | (aref (coding-system-spec coding-system) coding-spec-doc-string-idx)) | |
4ed46869 | 471 | |
d3675a42 | 472 | (defun coding-system-plist (coding-system) |
0269ddfb KH |
473 | "Return the property list of CODING-SYSTEM." |
474 | (aref (coding-system-spec coding-system) coding-spec-plist-idx)) | |
d3675a42 | 475 | |
4ed46869 | 476 | (defun coding-system-flags (coding-system) |
0269ddfb KH |
477 | "Return `flags' of CODING-SYSTEM. |
478 | A `flags' of a coding system is a vector of length 32 indicating detailed | |
479 | information of a coding system. See the function `make-coding-system' | |
480 | for more detail." | |
481 | (aref (coding-system-spec coding-system) coding-spec-flags-idx)) | |
482 | ||
483 | (defun coding-system-get (coding-system prop) | |
484 | "Extract a value from CODING-SYSTEM's property list for property PROP." | |
485 | (plist-get (coding-system-plist coding-system) prop)) | |
486 | ||
487 | (defun coding-system-put (coding-system prop val) | |
488 | "Change value in CODING-SYSTEM's property list PROP to VAL." | |
489 | (let ((plist (coding-system-plist coding-system))) | |
490 | (if plist | |
491 | (plist-put plist prop val) | |
492 | (aset (coding-system-spec coding-system) coding-spec-plist-idx | |
493 | (list prop val))))) | |
494 | ||
495 | (defun coding-system-category (coding-system) | |
496 | "Return the coding category of CODING-SYSTEM." | |
497 | (coding-system-get coding-system 'coding-category)) | |
498 | ||
499 | (defun coding-system-base (coding-system) | |
500 | "Return the base coding system of CODING-SYSTEM. | |
d9e3229d KH |
501 | A base coding system is what made by `make-coding-system'. |
502 | Any alias nor subsidiary coding systems are not base coding system." | |
0269ddfb KH |
503 | (car (coding-system-get coding-system 'alias-coding-systems))) |
504 | ||
505 | (defalias 'coding-system-parent 'coding-system-base) | |
2598a293 | 506 | (make-obsolete 'coding-system-parent 'coding-system-base "20.3") |
0269ddfb KH |
507 | |
508 | ;; Coding system also has a property `eol-type'. | |
509 | ;; | |
510 | ;; This property indicates how the coding system handles end-of-line | |
511 | ;; format. The value is integer 0, 1, 2, or a vector of three coding | |
512 | ;; systems. Each integer value 0, 1, and 2 indicates the format of | |
513 | ;; end-of-line LF, CRLF, and CR respectively. A vector value | |
514 | ;; indicates that the format of end-of-line should be detected | |
515 | ;; automatically. Nth element of the vector is the subsidiary coding | |
516 | ;; system whose `eol-type' property is N. | |
4ed46869 | 517 | |
8057896b | 518 | (defun coding-system-eol-type (coding-system) |
0269ddfb KH |
519 | "Return eol-type of CODING-SYSTEM. |
520 | An eol-type is integer 0, 1, 2, or a vector of coding systems. | |
4ed46869 | 521 | |
0269ddfb KH |
522 | Integer values 0, 1, and 2 indicate a format of end-of-line; LF, |
523 | CRLF, and CR respectively. | |
524 | ||
525 | A vector value indicates that a format of end-of-line should be | |
526 | detected automatically. Nth element of the vector is the subsidiary | |
527 | coding system whose eol-type is N." | |
528 | (get coding-system 'eol-type)) | |
bd882697 | 529 | |
857ea15c AS |
530 | (defun coding-system-lessp (x y) |
531 | (cond ((eq x 'no-conversion) t) | |
532 | ((eq y 'no-conversion) nil) | |
533 | ((eq x 'emacs-mule) t) | |
534 | ((eq y 'emacs-mule) nil) | |
535 | ((eq x 'undecided) t) | |
536 | ((eq y 'undecided) nil) | |
537 | (t (let ((c1 (coding-system-mnemonic x)) | |
538 | (c2 (coding-system-mnemonic y))) | |
539 | (or (< (downcase c1) (downcase c2)) | |
540 | (and (not (> (downcase c1) (downcase c2))) | |
541 | (< c1 c2))))))) | |
542 | ||
543 | ;; Add CODING-SYSTEM to coding-system-list while keeping it sorted. | |
544 | (defun add-to-coding-system-list (coding-system) | |
545 | (if (or (null coding-system-list) | |
546 | (coding-system-lessp coding-system (car coding-system-list))) | |
547 | (setq coding-system-list (cons coding-system coding-system-list)) | |
548 | (let ((len (length coding-system-list)) | |
549 | mid (tem coding-system-list)) | |
550 | (while (> len 1) | |
551 | (setq mid (nthcdr (/ len 2) tem)) | |
552 | (if (coding-system-lessp (car mid) coding-system) | |
553 | (setq tem mid | |
554 | len (- len (/ len 2))) | |
555 | (setq len (/ len 2)))) | |
556 | (setcdr tem (cons coding-system (cdr tem)))))) | |
557 | ||
80a7463d | 558 | (defun coding-system-list (&optional base-only) |
c11a8f77 KH |
559 | "Return a list of all existing non-subsidiary coding systems. |
560 | If optional arg BASE-ONLY is non-nil, only base coding systems are listed. | |
561 | The value doesn't include subsidiary coding systems which are what | |
562 | made from bases and aliases automatically for various end-of-line | |
563 | formats (e.g. iso-latin-1-unix, koi8-r-dos)." | |
80a7463d KH |
564 | (let* ((codings (copy-sequence coding-system-list)) |
565 | (tail (cons nil codings))) | |
566 | ;; Remove subsidiary coding systems (eol variants) and alias | |
567 | ;; coding systems (if necessary). | |
568 | (while (cdr tail) | |
569 | (let* ((coding (car (cdr tail))) | |
570 | (aliases (coding-system-get coding 'alias-coding-systems))) | |
571 | (if (or | |
572 | ;; CODING is an eol variant if not in ALIASES. | |
573 | (not (memq coding aliases)) | |
574 | ;; CODING is an alias if it is not car of ALIASES. | |
575 | (and base-only (not (eq coding (car aliases))))) | |
576 | (setcdr tail (cdr (cdr tail))) | |
577 | (setq tail (cdr tail))))) | |
578 | codings)) | |
579 | ||
c11a8f77 KH |
580 | (defun register-char-codings (coding-system safe-chars) |
581 | (let ((general (char-table-extra-slot char-coding-system-table 0))) | |
582 | (if (eq safe-chars t) | |
583 | (or (memq coding-system general) | |
584 | (set-char-table-extra-slot char-coding-system-table 0 | |
585 | (cons coding-system general))) | |
586 | (map-char-table | |
587 | (function | |
588 | (lambda (key val) | |
589 | (if (and (>= key 128) val) | |
590 | (let ((codings (aref char-coding-system-table key))) | |
591 | (or (memq coding-system codings) | |
592 | (aset char-coding-system-table key | |
593 | (cons coding-system codings))))))) | |
594 | safe-chars)))) | |
595 | ||
596 | ||
6e9722b0 KH |
597 | ;; Make subsidiary coding systems (eol-type variants) of CODING-SYSTEM. |
598 | (defun make-subsidiary-coding-system (coding-system) | |
0269ddfb KH |
599 | (let ((coding-spec (coding-system-spec coding-system)) |
600 | (subsidiaries (vector (intern (format "%s-unix" coding-system)) | |
8057896b KH |
601 | (intern (format "%s-dos" coding-system)) |
602 | (intern (format "%s-mac" coding-system)))) | |
0269ddfb KH |
603 | (i 0) |
604 | temp) | |
8057896b | 605 | (while (< i 3) |
0269ddfb | 606 | (put (aref subsidiaries i) 'coding-system coding-spec) |
8057896b | 607 | (put (aref subsidiaries i) 'eol-type i) |
857ea15c | 608 | (add-to-coding-system-list (aref subsidiaries i)) |
0269ddfb KH |
609 | (setq coding-system-alist |
610 | (cons (list (symbol-name (aref subsidiaries i))) | |
611 | coding-system-alist)) | |
8057896b KH |
612 | (setq i (1+ i))) |
613 | subsidiaries)) | |
4ed46869 | 614 | |
8057896b | 615 | (defun make-coding-system (coding-system type mnemonic doc-string |
1b46a680 KH |
616 | &optional |
617 | flags | |
618 | properties | |
619 | eol-type) | |
3bb1accb | 620 | "Define a new coding system CODING-SYSTEM (symbol). |
a6acd8a2 | 621 | Remaining arguments are TYPE, MNEMONIC, DOC-STRING, FLAGS (optional), |
d9e3229d | 622 | and PROPERTIES (optional) which construct a coding-spec of CODING-SYSTEM |
a6acd8a2 | 623 | in the following format: |
0269ddfb | 624 | [TYPE MNEMONIC DOC-STRING PLIST FLAGS] |
3bb1accb KH |
625 | |
626 | TYPE is an integer value indicating the type of the coding system as follows: | |
4ed46869 KH |
627 | 0: Emacs internal format, |
628 | 1: Shift-JIS (or MS-Kanji) used mainly on Japanese PC, | |
629 | 2: ISO-2022 including many variants, | |
630 | 3: Big5 used mainly on Chinese PC, | |
cd9c3177 KH |
631 | 4: private, CCL programs provide encoding/decoding algorithm, |
632 | 5: Raw-text, which means that text contains random 8-bit codes. | |
0269ddfb | 633 | |
3bb1accb | 634 | MNEMONIC is a character to be displayed on mode line for the coding system. |
0269ddfb | 635 | |
3bb1accb | 636 | DOC-STRING is a documentation string for the coding system. |
0269ddfb | 637 | |
3bb1accb | 638 | FLAGS specifies more detailed information of the coding system as follows: |
d6d6d592 | 639 | |
3bb1accb | 640 | If TYPE is 2 (ISO-2022), FLAGS is a list of these elements: |
4ed46869 KH |
641 | CHARSET0, CHARSET1, CHARSET2, CHARSET3, SHORT-FORM, |
642 | ASCII-EOL, ASCII-CNTL, SEVEN, LOCKING-SHIFT, SINGLE-SHIFT, | |
d6d6d592 | 643 | USE-ROMAN, USE-OLDJIS, NO-ISO6429, INIT-BOL, DESIGNATION-BOL, |
850101ed | 644 | SAFE, ACCEPT-LATIN-EXTRA-CODE. |
4ed46869 KH |
645 | CHARSETn are character sets initially designated to Gn graphic registers. |
646 | If CHARSETn is nil, Gn is never used. | |
647 | If CHARSETn is t, Gn can be used but nothing designated initially. | |
648 | If CHARSETn is a list of character sets, those character sets are | |
649 | designated to Gn on output, but nothing designated to Gn initially. | |
421e3b4e | 650 | But, character set `ascii' can be designated only to G0. |
4ed46869 KH |
651 | SHORT-FORM non-nil means use short designation sequence on output. |
652 | ASCII-EOL non-nil means designate ASCII to g0 at end of line on output. | |
653 | ASCII-CNTL non-nil means designate ASCII to g0 before control codes and | |
654 | SPACE on output. | |
655 | SEVEN non-nil means use 7-bit code only on output. | |
656 | LOCKING-SHIFT non-nil means use locking-shift. | |
657 | SINGLE-SHIFT non-nil means use single-shift. | |
658 | USE-ROMAN non-nil means designate JIS0201-1976-Roman instead of ASCII. | |
659 | USE-OLDJIS non-nil means designate JIS0208-1976 instead of JIS0208-1983. | |
660 | NO-ISO6429 non-nil means not use ISO6429's direction specification. | |
69eba008 KH |
661 | INIT-BOL non-nil means any designation state is assumed to be reset |
662 | to initial at each beginning of line on output. | |
663 | DESIGNATION-BOL non-nil means designation sequences should be placed | |
664 | at beginning of line on output. | |
a6acd8a2 | 665 | SAFE non-nil means convert unsafe characters to `?' on output. |
c11a8f77 KH |
666 | Characters not specified in the property `safe-charsets' nor |
667 | `safe-chars' are unsafe. | |
850101ed RS |
668 | ACCEPT-LATIN-EXTRA-CODE non-nil means code-detection routine accepts |
669 | a code specified in `latin-extra-code-table' (which see) as a valid | |
670 | code of the coding system. | |
d6d6d592 | 671 | |
3bb1accb KH |
672 | If TYPE is 4 (private), FLAGS should be a cons of CCL programs, for |
673 | decoding and encoding. CCL programs should be specified by their | |
674 | symbols. | |
d9e3229d KH |
675 | |
676 | PROPERTIES is an alist of properties vs the corresponding values. | |
677 | These properties are set in PLIST, a property list. This function | |
678 | also sets properties `coding-category' and `alias-coding-systems' | |
679 | automatically. | |
4ed46869 | 680 | |
1b46a680 KH |
681 | EOL-TYPE specifies the EOL type of the coding-system in one of the |
682 | following formats: | |
683 | ||
684 | o symbol (unix, dos, or mac) | |
685 | ||
686 | The symbol `unix' means Unix-like EOL (LF), `dos' means | |
687 | DOS-like EOL (CRLF), and `mac' means MAC-like EOL (CR). | |
688 | ||
689 | o number (0, 1, or 2) | |
690 | ||
691 | The number 0, 1, and 2 mean UNIX, DOS, and MAC-like EOL | |
692 | respectively. | |
693 | ||
694 | o vector of coding-systems of length 3 | |
695 | ||
696 | The EOL type is detected automatically for the coding system. | |
697 | And, according to the detected EOL type, one of the coding | |
698 | systems in the vector is selected. Elements of the vector | |
4d3a9228 | 699 | corresponds to Unix-like EOL, DOS-like EOL, and Mac-like EOL |
1b46a680 KH |
700 | in this order. |
701 | ||
3bb1accb KH |
702 | Kludgy features for backward compatibility: |
703 | ||
704 | 1. If TYPE is 4 and car or cdr of FLAGS is a vector, the vector is | |
705 | treated as a compiled CCL code. | |
706 | ||
707 | 2. If PROPERTIES is just a list of character sets, the list is set as | |
708 | a value of `safe-charsets' in PLIST." | |
0269ddfb | 709 | ;; Set a value of `coding-system' property. |
6e9722b0 | 710 | (let ((coding-spec (make-vector 5 nil)) |
d9e3229d KH |
711 | (no-initial-designation t) |
712 | (no-alternative-designation t) | |
6e9722b0 | 713 | coding-category) |
cd9c3177 | 714 | (if (or (not (integerp type)) (< type 0) (> type 5)) |
0269ddfb | 715 | (error "TYPE argument must be 0..5")) |
8057896b | 716 | (if (or (not (integerp mnemonic)) (<= mnemonic ? ) (> mnemonic 127)) |
e8dd0160 | 717 | (error "MNEMONIC argument must be an ASCII printable character.")) |
0269ddfb KH |
718 | (aset coding-spec coding-spec-type-idx type) |
719 | (aset coding-spec coding-spec-mnemonic-idx mnemonic) | |
720 | (aset coding-spec coding-spec-doc-string-idx | |
e127a722 | 721 | (purecopy (if (stringp doc-string) doc-string ""))) |
6e9722b0 KH |
722 | (cond ((= type 0) |
723 | (setq coding-category 'coding-category-emacs-mule)) | |
724 | ((= type 1) | |
725 | (setq coding-category 'coding-category-sjis)) | |
726 | ((= type 2) ; ISO2022 | |
4ed46869 | 727 | (let ((i 0) |
6e9722b0 | 728 | (vec (make-vector 32 nil)) |
05134257 KH |
729 | (g1-designation nil) |
730 | (fl flags)) | |
4ed46869 | 731 | (while (< i 4) |
05134257 | 732 | (let ((charset (car fl))) |
6e9722b0 KH |
733 | (if (and no-initial-designation |
734 | (> i 0) | |
735 | (or (charsetp charset) | |
736 | (and (consp charset) | |
737 | (charsetp (car charset))))) | |
738 | (setq no-initial-designation nil)) | |
739 | (if (charsetp charset) | |
740 | (if (= i 1) (setq g1-designation charset)) | |
741 | (if (consp charset) | |
742 | (let ((tail charset) | |
743 | elt) | |
744 | (while tail | |
745 | (setq elt (car tail)) | |
d9e3229d KH |
746 | (if (eq elt t) |
747 | (setq no-alternative-designation nil) | |
748 | (if (and elt (not (charsetp elt))) | |
749 | (error "Invalid charset: %s" elt))) | |
6e9722b0 KH |
750 | (setq tail (cdr tail))) |
751 | (setq g1-designation (car charset))) | |
d9e3229d KH |
752 | (if charset |
753 | (if (eq charset t) | |
754 | (setq no-alternative-designation nil) | |
755 | (error "Invalid charset: %s" charset))))) | |
4ed46869 | 756 | (aset vec i charset)) |
05134257 KH |
757 | (setq fl (cdr fl) i (1+ i))) |
758 | (while (and (< i 32) fl) | |
759 | (aset vec i (car fl)) | |
760 | (setq fl (cdr fl) i (1+ i))) | |
6e9722b0 | 761 | (aset coding-spec 4 vec) |
6e9722b0 KH |
762 | (setq coding-category |
763 | (if (aref vec 8) ; Use locking-shift. | |
dc64cd19 KH |
764 | (or (and (aref vec 7) 'coding-category-iso-7-else) |
765 | 'coding-category-iso-8-else) | |
6e9722b0 KH |
766 | (if (aref vec 7) ; 7-bit only. |
767 | (if (aref vec 9) ; Use single-shift. | |
dc64cd19 | 768 | 'coding-category-iso-7-else |
d9e3229d KH |
769 | (if no-alternative-designation |
770 | 'coding-category-iso-7-tight | |
771 | 'coding-category-iso-7)) | |
3ad911d8 KH |
772 | (if (or no-initial-designation |
773 | (not no-alternative-designation)) | |
dc64cd19 | 774 | 'coding-category-iso-8-else |
6e9722b0 KH |
775 | (if (and (charsetp g1-designation) |
776 | (= (charset-dimension g1-designation) 2)) | |
777 | 'coding-category-iso-8-2 | |
778 | 'coding-category-iso-8-1))))))) | |
779 | ((= type 3) | |
780 | (setq coding-category 'coding-category-big5)) | |
781 | ((= type 4) ; private | |
c76b5c99 | 782 | (setq coding-category 'coding-category-ccl) |
3bb1accb KH |
783 | (if (not (consp flags)) |
784 | (error "Invalid FLAGS argument for TYPE 4 (CCL)") | |
785 | (let ((decoder (check-ccl-program | |
786 | (car flags) | |
787 | (intern (format "%s-decoder" coding-system)))) | |
788 | (encoder (check-ccl-program | |
789 | (cdr flags) | |
790 | (intern (format "%s-encoder" coding-system))))) | |
791 | (if (and decoder encoder) | |
792 | (aset coding-spec 4 (cons decoder encoder)) | |
793 | (error "Invalid FLAGS argument for TYPE 4 (CCL)"))))) | |
cd9c3177 KH |
794 | (t ; i.e. (= type 5) |
795 | (setq coding-category 'coding-category-raw-text))) | |
0269ddfb KH |
796 | |
797 | (let ((plist (list 'coding-category coding-category | |
d9e3229d | 798 | 'alias-coding-systems (list coding-system)))) |
0269ddfb | 799 | (if no-initial-designation |
d9e3229d KH |
800 | (plist-put plist 'no-initial-designation t)) |
801 | (if (and properties | |
802 | (or (eq properties t) | |
803 | (not (consp (car properties))))) | |
804 | ;; In the old version, the arg PROPERTIES is a list to be | |
805 | ;; set in PLIST as a value of property `safe-charsets'. | |
c11a8f77 KH |
806 | (setq properties (list (cons 'safe-charsets properties)))) |
807 | ;; In the current version PROPERTIES is a property list. | |
808 | ;; Reflect it into PLIST one by one while handling safe-chars | |
809 | ;; specially. | |
810 | (let ((safe-charsets (cdr (assq 'safe-charsets properties))) | |
811 | (safe-chars (cdr (assq 'safe-chars properties))) | |
812 | (l properties) | |
813 | prop val) | |
814 | ;; If only safe-charsets is specified, make a char-table from | |
815 | ;; it, and store that char-table as the value of `safe-chars'. | |
816 | (if (and (not safe-chars) safe-charsets) | |
817 | (let (charset) | |
818 | (if (eq safe-charsets t) | |
819 | (setq safe-chars t) | |
820 | (setq safe-chars (make-char-table 'safe-chars)) | |
821 | (while safe-charsets | |
822 | (setq charset (car safe-charsets) | |
823 | safe-charsets (cdr safe-charsets)) | |
824 | (cond ((eq charset 'ascii)) ; just ignore | |
825 | ((eq charset 'eight-bit-control) | |
826 | (let ((i 128)) | |
827 | (while (< i 160) | |
828 | (aset safe-chars i t) | |
829 | (setq i (1+ i))))) | |
830 | ((eq charset 'eight-bit-graphic) | |
831 | (let ((i 160)) | |
832 | (while (< i 256) | |
833 | (aset safe-chars i t) | |
834 | (setq i (1+ i))))) | |
835 | (t | |
836 | (aset safe-chars (make-char charset) t))))) | |
837 | (setq l (cons (cons 'safe-chars safe-chars) l)))) | |
838 | (while l | |
839 | (setq prop (car (car l)) val (cdr (car l)) l (cdr l)) | |
840 | (if (eq prop 'safe-chars) | |
841 | (progn | |
cc57cc54 KH |
842 | (if (and (symbolp val) |
843 | (get val 'translation-table)) | |
844 | (setq safe-chars (get val 'translation-table))) | |
845 | (register-char-codings coding-system safe-chars) | |
846 | (setq val safe-chars))) | |
c11a8f77 | 847 | (plist-put plist prop val))) |
1b46a680 KH |
848 | ;; The property `coding-category' may have been set differently |
849 | ;; through PROPERTIES. | |
850 | (setq coding-category (plist-get plist 'coding-category)) | |
0269ddfb | 851 | (aset coding-spec coding-spec-plist-idx plist)) |
6e9722b0 | 852 | (put coding-system 'coding-system coding-spec) |
6e9722b0 KH |
853 | (put coding-category 'coding-systems |
854 | (cons coding-system (get coding-category 'coding-systems)))) | |
4ed46869 | 855 | |
1b46a680 | 856 | ;; Next, set a value of `eol-type' property. |
af873cd1 | 857 | (if (not eol-type) |
1b46a680 KH |
858 | ;; If EOL-TYPE is nil, set a vector of subsidiary coding |
859 | ;; systems, each corresponds to a coding system for the detected | |
860 | ;; EOL format. | |
861 | (setq eol-type (make-subsidiary-coding-system coding-system))) | |
862 | (setq eol-type | |
863 | (cond ((or (eq eol-type 'unix) (null eol-type)) | |
864 | 0) | |
865 | ((eq eol-type 'dos) | |
866 | 1) | |
867 | ((eq eol-type 'mac) | |
868 | 2) | |
869 | ((or (and (vectorp eol-type) | |
870 | (= (length eol-type) 3)) | |
871 | (and (numberp eol-type) | |
872 | (and (>= eol-type 0) | |
873 | (<= eol-type 2)))) | |
874 | eol-type) | |
875 | (t | |
876 | (error "Invalid EOL-TYPE spec:%S" eol-type)))) | |
877 | (put coding-system 'eol-type eol-type) | |
0269ddfb KH |
878 | |
879 | ;; At last, register CODING-SYSTEM in `coding-system-list' and | |
880 | ;; `coding-system-alist'. | |
857ea15c | 881 | (add-to-coding-system-list coding-system) |
0269ddfb | 882 | (setq coding-system-alist (cons (list (symbol-name coding-system)) |
d9e3229d | 883 | coding-system-alist)) |
05134257 KH |
884 | |
885 | ;; For a coding system of cateogory iso-8-1 and iso-8-2, create | |
886 | ;; XXX-with-esc variants. | |
887 | (let ((coding-category (coding-system-category coding-system))) | |
888 | (if (or (eq coding-category 'coding-category-iso-8-1) | |
889 | (eq coding-category 'coding-category-iso-8-2)) | |
890 | (let ((esc (intern (concat (symbol-name coding-system) "-with-esc"))) | |
c11a8f77 KH |
891 | (doc (format "Same as %s but can handle any charsets by ISO's escape sequences." coding-system)) |
892 | (safe-charsets (assq 'safe-charsets properties)) | |
893 | (mime-charset (assq 'mime-charset properties))) | |
894 | (if safe-charsets | |
895 | (setcdr safe-charsets t) | |
896 | (setq properties (cons (cons 'safe-charsets t) properties))) | |
897 | (if mime-charset | |
898 | (setcdr mime-charset nil)) | |
05134257 KH |
899 | (make-coding-system esc type mnemonic doc |
900 | (if (listp (car flags)) | |
901 | (cons (append (car flags) '(t)) (cdr flags)) | |
902 | (cons (list (car flags) t) (cdr flags))) | |
c11a8f77 | 903 | properties)))) |
05134257 | 904 | |
d9e3229d | 905 | coding-system) |
8057896b | 906 | |
a42763dc | 907 | (defun define-coding-system-alias (alias coding-system) |
358d28fb | 908 | "Define ALIAS as an alias for coding system CODING-SYSTEM." |
0269ddfb KH |
909 | (put alias 'coding-system (coding-system-spec coding-system)) |
910 | (nconc (coding-system-get alias 'alias-coding-systems) (list alias)) | |
857ea15c | 911 | (add-to-coding-system-list alias) |
0269ddfb KH |
912 | (setq coding-system-alist (cons (list (symbol-name alias)) |
913 | coding-system-alist)) | |
914 | (let ((eol-type (coding-system-eol-type coding-system))) | |
915 | (if (vectorp eol-type) | |
916 | (put alias 'eol-type (make-subsidiary-coding-system alias)) | |
917 | (put alias 'eol-type eol-type)))) | |
4ed46869 KH |
918 | |
919 | (defun set-buffer-file-coding-system (coding-system &optional force) | |
358d28fb RS |
920 | "Set the file coding-system of the current buffer to CODING-SYSTEM. |
921 | This means that when you save the buffer, it will be converted | |
922 | according to CODING-SYSTEM. For a list of possible values of CODING-SYSTEM, | |
923 | use \\[list-coding-systems]. | |
924 | ||
925 | If the buffer's previous file coding-system value specifies end-of-line | |
926 | conversion, and CODING-SYSTEM does not specify one, CODING-SYSTEM is | |
927 | merged with the already-specified end-of-line conversion. | |
b839fdcc KH |
928 | |
929 | If the buffer's previous file coding-system value specifies text | |
930 | conversion, and CODING-SYSTEM does not specify one, CODING-SYSTEM is | |
931 | merged with the already-specified text conversion. | |
932 | ||
933 | However, if the optional prefix argument FORCE is non-nil, then | |
934 | CODING-SYSTEM is used exactly as specified. | |
aeef8f07 KH |
935 | |
936 | This marks the buffer modified so that the succeeding \\[save-buffer] | |
937 | surely saves the buffer with CODING-SYSTEM. From a program, if you | |
938 | don't want to mark the buffer modified, just set the variable | |
939 | `buffer-file-coding-system' directly." | |
d9e3229d | 940 | (interactive "zCoding system for visited file (default, nil): \nP") |
4ed46869 | 941 | (check-coding-system coding-system) |
36d455c4 | 942 | (if (and coding-system buffer-file-coding-system (null force)) |
8dd735c1 KH |
943 | (let ((base (coding-system-base buffer-file-coding-system)) |
944 | (eol (coding-system-eol-type buffer-file-coding-system))) | |
945 | ;; If CODING-SYSTEM doesn't specify text conversion, merge | |
946 | ;; with that of buffer-file-coding-system. | |
947 | (if (eq (coding-system-base coding-system) 'undecided) | |
948 | (setq coding-system (coding-system-change-text-conversion | |
949 | coding-system base))) | |
950 | ;; If CODING-SYSTEM doesn't specify eol conversion, merge with | |
951 | ;; that of buffer-file-coding-system. | |
952 | (if (and (vectorp (coding-system-eol-type coding-system)) | |
953 | (numberp eol) (>= eol 0) (<= eol 2)) | |
954 | (setq coding-system (coding-system-change-eol-conversion | |
955 | coding-system eol))))) | |
4ed46869 KH |
956 | (setq buffer-file-coding-system coding-system) |
957 | (set-buffer-modified-p t) | |
958 | (force-mode-line-update)) | |
959 | ||
358d28fb RS |
960 | (defvar default-terminal-coding-system nil |
961 | "Default value for the terminal coding system. | |
962 | This is normally set according to the selected language environment. | |
963 | See also the command `set-terminal-coding-system'.") | |
964 | ||
df100398 KH |
965 | (defun set-terminal-coding-system (coding-system) |
966 | "Set coding system of your terminal to CODING-SYSTEM. | |
358d28fb RS |
967 | All text output to the terminal will be encoded |
968 | with the specified coding system. | |
969 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. | |
970 | The default is determined by the selected language environment | |
971 | or by the previous use of this command." | |
972 | (interactive | |
2e02a76f RS |
973 | (list (let ((default (if (and (not (terminal-coding-system)) |
974 | default-terminal-coding-system) | |
975 | default-terminal-coding-system))) | |
976 | (read-coding-system | |
977 | (format "Coding system for terminal display (default, %s): " | |
978 | default) | |
979 | default)))) | |
358d28fb RS |
980 | (if (and (not coding-system) |
981 | (not (terminal-coding-system))) | |
982 | (setq coding-system default-terminal-coding-system)) | |
983 | (if coding-system | |
984 | (setq default-terminal-coding-system coding-system)) | |
df100398 KH |
985 | (set-terminal-coding-system-internal coding-system) |
986 | (redraw-frame (selected-frame))) | |
987 | ||
358d28fb RS |
988 | (defvar default-keyboard-coding-system nil |
989 | "Default value of the keyboard coding system. | |
990 | This is normally set according to the selected language environment. | |
991 | See also the command `set-keyboard-coding-system'.") | |
992 | ||
df100398 | 993 | (defun set-keyboard-coding-system (coding-system) |
358d28fb RS |
994 | "Set coding system for keyboard input to CODING-SYSTEM. |
995 | In addition, this command enables Encoded-kbd minor mode. | |
6d34f495 DL |
996 | \(If CODING-SYSTEM is nil, Encoded-kbd mode is turned off -- see |
997 | `encoded-kbd-mode'.) | |
358d28fb RS |
998 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. |
999 | The default is determined by the selected language environment | |
1000 | or by the previous use of this command." | |
1001 | (interactive | |
2e02a76f RS |
1002 | (list (let ((default (if (and (not (keyboard-coding-system)) |
1003 | default-keyboard-coding-system) | |
1004 | default-keyboard-coding-system))) | |
1005 | (read-coding-system | |
1006 | (format "Coding system for keyboard input (default, %s): " | |
1007 | default) | |
1008 | default)))) | |
358d28fb RS |
1009 | (if (and (not coding-system) |
1010 | (not (keyboard-coding-system))) | |
1011 | (setq coding-system default-keyboard-coding-system)) | |
1012 | (if coding-system | |
1013 | (setq default-keyboard-coding-system coding-system)) | |
df100398 KH |
1014 | (set-keyboard-coding-system-internal coding-system) |
1015 | (encoded-kbd-mode (if coding-system 1 0))) | |
1016 | ||
6d34f495 DL |
1017 | (defcustom keyboard-coding-system nil |
1018 | "Specify coding system for keyboard input. | |
1019 | If you set this on a terminal which can't distinguish Meta keys from | |
1020 | 8-bit characters, you will have to use ESC to type Meta characters. | |
1021 | See Info node `Specify Coding' and Info node `Single-Byte Character Support'. | |
1022 | ||
1023 | Setting this variable directly does not take effect; | |
1024 | use either M-x customize or \\[set-keyboard-coding-system]." | |
1025 | :type '(coding-system :tag "Coding system") | |
1026 | :link '(info-link "(emacs)Specify Coding") | |
1027 | :link '(info-link "(emacs)Single-Byte Character Support") | |
1028 | :set (lambda (symbol value) | |
1029 | ;; Don't load encoded-kbd-mode unnecessarily. | |
1030 | (if (or value (boundp 'encoded-kbd-mode)) | |
1031 | (set-keyboard-coding-system value) | |
1032 | (set-default 'keyboard-coding-system nil))) ; must initialize | |
1033 | :version "21.1" | |
1034 | :group 'keyboard | |
1035 | :group 'mule) | |
1036 | ||
df100398 | 1037 | (defun set-buffer-process-coding-system (decoding encoding) |
358d28fb | 1038 | "Set coding systems for the process associated with the current buffer. |
df100398 | 1039 | DECODING is the coding system to be used to decode input from the process, |
358d28fb RS |
1040 | ENCODING is the coding system to be used to encode output to the process. |
1041 | ||
1042 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]." | |
4ed46869 | 1043 | (interactive |
83911021 | 1044 | "zCoding-system for output from the process: \nzCoding-system for input to the process: ") |
4ed46869 KH |
1045 | (let ((proc (get-buffer-process (current-buffer)))) |
1046 | (if (null proc) | |
1047 | (error "no process") | |
df100398 KH |
1048 | (check-coding-system decoding) |
1049 | (check-coding-system encoding) | |
1050 | (set-process-coding-system proc decoding encoding))) | |
4ed46869 KH |
1051 | (force-mode-line-update)) |
1052 | ||
d0b99881 RS |
1053 | (defalias 'set-clipboard-coding-system 'set-selection-coding-system) |
1054 | ||
14915c37 | 1055 | (defun set-selection-coding-system (coding-system) |
b25eef20 KH |
1056 | "Make CODING-SYSTEM used for communicating with other X clients . |
1057 | When sending or receiving text via cut_buffer, selection, and clipboard, | |
1058 | the text is encoded or decoded by CODING-SYSTEM." | |
a03b3ce1 | 1059 | (interactive "zCoding system for X selection: ") |
b25eef20 | 1060 | (check-coding-system coding-system) |
14915c37 | 1061 | (setq selection-coding-system coding-system)) |
b25eef20 | 1062 | |
e8dd0160 | 1063 | ;; Coding system lastly specified by the command |
a03b3ce1 KH |
1064 | ;; set-next-selection-coding-system. |
1065 | (defvar last-next-selection-coding-system nil) | |
1066 | ||
1067 | (defun set-next-selection-coding-system (coding-system) | |
1068 | "Make CODING-SYSTEM used for the next communication with other X clients. | |
1069 | This setting is effective for the next communication only." | |
1070 | (interactive | |
1071 | (list (read-coding-system | |
1072 | (if last-next-selection-coding-system | |
1073 | (format "Coding system for the next X selection (default, %S): " | |
1074 | last-next-selection-coding-system) | |
1075 | "Coding system for the next X selection: ") | |
1076 | last-next-selection-coding-system))) | |
1077 | (if coding-system | |
1078 | (setq last-next-selection-coding-system coding-system) | |
1079 | (setq coding-system last-next-selection-coding-system)) | |
1080 | (check-coding-system coding-system) | |
1081 | ||
1082 | (setq next-selection-coding-system coding-system)) | |
1083 | ||
4ed46869 | 1084 | (defun set-coding-priority (arg) |
d9e3229d KH |
1085 | "Set priority of coding categories according to LIST. |
1086 | LIST is a list of coding categories ordered by priority." | |
1087 | (let ((l arg) | |
1088 | (current-list (copy-sequence coding-category-list))) | |
e8dd0160 | 1089 | ;; Check the validity of ARG while deleting coding categories in |
d9e3229d KH |
1090 | ;; ARG from CURRENT-LIST. We assume that CODING-CATEGORY-LIST |
1091 | ;; contains all coding categories. | |
1092 | (while l | |
1093 | (if (or (null (get (car l) 'coding-category-index)) | |
1094 | (null (memq (car l) current-list))) | |
1095 | (error "Invalid or duplicated element in argument: %s" arg)) | |
1096 | (setq current-list (delq (car l) current-list)) | |
1097 | (setq l (cdr l))) | |
4ed46869 | 1098 | ;; Update `coding-category-list' and return it. |
2feaf204 KH |
1099 | (setq coding-category-list (append arg current-list)) |
1100 | (set-coding-priority-internal))) | |
4ed46869 KH |
1101 | |
1102 | ;;; FILE I/O | |
1103 | ||
e76938e7 | 1104 | (defcustom auto-coding-alist |
0735296c DL |
1105 | '(("\\.\\(arc\\|zip\\|lzh\\|zoo\\|jar\\|tar\\|tgz\\)\\'" . no-conversion) |
1106 | ("\\.\\(ARC\\|ZIP\\|LZH\\|ZOO\\|JAR\\|TAR\\|TGZ\\)\\'" . no-conversion)) | |
835f49b8 KH |
1107 | "Alist of filename patterns vs corresponding coding systems. |
1108 | Each element looks like (REGEXP . CODING-SYSTEM). | |
558b0c86 | 1109 | A file whose name matches REGEXP is decoded by CODING-SYSTEM on reading. |
835f49b8 | 1110 | |
7fed493a RS |
1111 | The settings in this alist take priority over `coding:' tags |
1112 | in the file (see the function `set-auto-coding') | |
e76938e7 DL |
1113 | and the contents of `file-coding-system-alist'." |
1114 | :group 'files | |
1115 | :group 'mule | |
1116 | :type '(repeat (cons (regexp :tag "File name regexp") | |
1117 | (symbol :tag "Coding system")))) | |
835f49b8 | 1118 | |
1c4cc63a KH |
1119 | (defvar set-auto-coding-for-load nil |
1120 | "Non-nil means look for `load-coding' property instead of `coding'. | |
1121 | This is used for loading and byte-compiling Emacs Lisp files.") | |
1122 | ||
8a592131 RS |
1123 | (defun auto-coding-alist-lookup (filename) |
1124 | "Return the coding system specified by `auto-coding-alist' for FILENAME." | |
1125 | (let ((alist auto-coding-alist) | |
ca128d75 | 1126 | (case-fold-search (memq system-type '(vax-vms windows-nt ms-dos))) |
8a592131 RS |
1127 | coding-system) |
1128 | (while (and alist (not coding-system)) | |
1129 | (if (string-match (car (car alist)) filename) | |
1130 | (setq coding-system (cdr (car alist))) | |
1131 | (setq alist (cdr alist)))) | |
1132 | coding-system)) | |
1133 | ||
835f49b8 KH |
1134 | (defun set-auto-coding (filename size) |
1135 | "Return coding system for a file FILENAME of which SIZE bytes follow point. | |
1c4cc63a KH |
1136 | These bytes should include at least the first 1k of the file |
1137 | and the last 3k of the file, but the middle may be omitted. | |
63561304 | 1138 | |
835f49b8 KH |
1139 | It checks FILENAME against the variable `auto-coding-alist'. |
1140 | If FILENAME doesn't match any entries in the variable, | |
1141 | it checks for a `coding:' tag in the first one or two lines following | |
e8dd0160 | 1142 | point. If no `coding:' tag is found, it checks for local variables |
1c4cc63a | 1143 | list in the last 3K bytes out of the SIZE bytes. |
63561304 KH |
1144 | |
1145 | The return value is the specified coding system, | |
1146 | or nil if nothing specified. | |
87aba788 | 1147 | |
ba74e833 | 1148 | The variable `set-auto-coding-function' (which see) is set to this |
87aba788 | 1149 | function by default." |
8a592131 | 1150 | (let ((coding-system (auto-coding-alist-lookup filename))) |
835f49b8 KH |
1151 | |
1152 | (or coding-system | |
1153 | (let* ((case-fold-search t) | |
1154 | (head-start (point)) | |
1155 | (head-end (+ head-start (min size 1024))) | |
1156 | (tail-start (+ head-start (max (- size 3072) 0))) | |
1157 | (tail-end (+ head-start size)) | |
1158 | coding-system head-found tail-found pos) | |
1159 | ;; Try a short cut by searching for the string "coding:" | |
e8dd0160 | 1160 | ;; and for "unibyte:" at the head and tail of SIZE bytes. |
835f49b8 KH |
1161 | (setq head-found (or (search-forward "coding:" head-end t) |
1162 | (search-forward "unibyte:" head-end t))) | |
1163 | (if (and head-found (> head-found tail-start)) | |
1164 | ;; Head and tail are overlapped. | |
1165 | (setq tail-found head-found) | |
1166 | (goto-char tail-start) | |
1167 | (setq tail-found (or (search-forward "coding:" tail-end t) | |
1168 | (search-forward "unibyte:" tail-end t)))) | |
1169 | ||
1170 | ;; At first check the head. | |
1171 | (when head-found | |
1172 | (goto-char head-start) | |
1173 | (setq pos (re-search-forward "[\n\r]" head-end t)) | |
1174 | (if (and pos | |
1175 | (= (char-after head-start) ?#) | |
1176 | (= (char-after (1+ head-start)) ?!)) | |
1177 | ;; If the file begins with "#!" (exec interpreter magic), | |
1178 | ;; look for coding frobs in the first two lines. You cannot | |
1179 | ;; necessarily put them in the first line of such a file | |
1180 | ;; without screwing up the interpreter invocation. | |
1181 | (setq pos (search-forward "\n" head-end t))) | |
1182 | (if pos (setq head-end pos)) | |
1183 | (when (< head-found head-end) | |
1184 | (goto-char head-start) | |
1185 | (when (and set-auto-coding-for-load | |
1186 | (re-search-forward | |
1187 | "-\\*-\\(.*;\\)?[ \t]*unibyte:[ \t]*\\([^ ;]+\\)" | |
1188 | head-end t)) | |
1189 | (setq coding-system 'raw-text)) | |
1190 | (when (and (not coding-system) | |
1191 | (re-search-forward | |
1192 | "-\\*-\\(.*;\\)?[ \t]*coding:[ \t]*\\([^ ;]+\\)" | |
1193 | head-end t)) | |
1194 | (setq coding-system (intern (match-string 2))) | |
1195 | (or (coding-system-p coding-system) | |
1196 | (setq coding-system nil))))) | |
1197 | ||
1198 | ;; If no coding: tag in the head, check the tail. | |
1199 | (when (and tail-found (not coding-system)) | |
1200 | (goto-char tail-start) | |
1201 | (search-forward "\n\^L" nil t) | |
1202 | (if (re-search-forward | |
1203 | "^\\(.*\\)[ \t]*Local Variables:[ \t]*\\(.*\\)$" tail-end t) | |
1204 | ;; The prefix is what comes before "local variables:" in its | |
1205 | ;; line. The suffix is what comes after "local variables:" | |
1206 | ;; in its line. | |
1207 | (let* ((prefix (regexp-quote (match-string 1))) | |
1208 | (suffix (regexp-quote (match-string 2))) | |
d318fcf4 EZ |
1209 | (re-coding |
1210 | (concat | |
1211 | "^" prefix | |
1212 | "[ \t]*coding[ \t]*:[ \t]*\\([^ \t]+\\)[ \t]*" | |
1213 | suffix "$")) | |
1214 | (re-unibyte | |
1215 | (concat | |
1216 | "^" prefix | |
1217 | "[ \t]*unibyte[ \t]*:[ \t]*\\([^ \t]+\\)[ \t]*" | |
1218 | suffix "$")) | |
1219 | (re-end | |
1220 | (concat "^" prefix "[ \t]*end *:[ \t]*" suffix "$")) | |
835f49b8 KH |
1221 | (pos (point))) |
1222 | (re-search-forward re-end tail-end 'move) | |
1223 | (setq tail-end (point)) | |
1224 | (goto-char pos) | |
1225 | (when (and set-auto-coding-for-load | |
1226 | (re-search-forward re-unibyte tail-end t)) | |
1227 | (setq coding-system 'raw-text)) | |
1228 | (when (and (not coding-system) | |
1229 | (re-search-forward re-coding tail-end t)) | |
1230 | (setq coding-system (intern (match-string 1))) | |
1231 | (or (coding-system-p coding-system) | |
1232 | (setq coding-system nil)))))) | |
1233 | coding-system)))) | |
63561304 KH |
1234 | |
1235 | (setq set-auto-coding-function 'set-auto-coding) | |
87aba788 | 1236 | |
4ed46869 KH |
1237 | ;; Set buffer-file-coding-system of the current buffer after some text |
1238 | ;; is inserted. | |
1239 | (defun after-insert-file-set-buffer-file-coding-system (inserted) | |
1240 | (if last-coding-system-used | |
1241 | (let ((coding-system | |
1242 | (find-new-buffer-file-coding-system last-coding-system-used)) | |
1243 | (modified-p (buffer-modified-p))) | |
0269ddfb | 1244 | (when coding-system |
71983219 | 1245 | (set-buffer-file-coding-system coding-system t) |
d0c26c63 KH |
1246 | (if (and enable-multibyte-characters |
1247 | (or (eq coding-system 'no-conversion) | |
ccb77b4e | 1248 | (eq (coding-system-type coding-system) 5)) |
136e48e4 KH |
1249 | ;; If buffer was unmodified and the size is the |
1250 | ;; same as INSERTED, we must be visiting it. | |
1251 | (not modified-p) | |
1252 | (= (buffer-size) inserted)) | |
ccb77b4e RS |
1253 | ;; For coding systems no-conversion and raw-text..., |
1254 | ;; edit the buffer as unibyte. | |
d0c26c63 KH |
1255 | (let ((pos-byte (position-bytes (+ (point) inserted)))) |
1256 | (set-buffer-multibyte nil) | |
1257 | (setq inserted (- pos-byte (position-bytes (point)))))) | |
0269ddfb | 1258 | (set-buffer-modified-p modified-p)))) |
d0c26c63 | 1259 | inserted) |
4ed46869 | 1260 | |
84c9d215 KH |
1261 | (add-hook 'after-insert-file-functions |
1262 | 'after-insert-file-set-buffer-file-coding-system) | |
4ed46869 | 1263 | |
8057896b | 1264 | ;; The coding-spec and eol-type of coding-system returned is decided |
4ed46869 KH |
1265 | ;; independently in the following order. |
1266 | ;; 1. That of buffer-file-coding-system locally bound. | |
1267 | ;; 2. That of CODING. | |
1268 | ||
1269 | (defun find-new-buffer-file-coding-system (coding) | |
1270 | "Return a coding system for a buffer when a file of CODING is inserted. | |
a73a8c89 KH |
1271 | The local variable `buffer-file-coding-system' of the current buffer |
1272 | is set to the returned value. | |
509064c5 | 1273 | Return nil if there's no need to set `buffer-file-coding-system'." |
4ed46869 | 1274 | (let (local-coding local-eol |
b685f8d6 | 1275 | found-coding found-eol |
4ed46869 KH |
1276 | new-coding new-eol) |
1277 | (if (null coding) | |
1278 | ;; Nothing found about coding. | |
1279 | nil | |
1280 | ||
b685f8d6 RS |
1281 | ;; Get information of `buffer-file-coding-system' in LOCAL-EOL |
1282 | ;; and LOCAL-CODING. | |
1283 | (setq local-eol (coding-system-eol-type buffer-file-coding-system)) | |
1284 | (if (null (numberp local-eol)) | |
1285 | ;; But eol-type is not yet set. | |
1286 | (setq local-eol nil)) | |
0269ddfb KH |
1287 | (if (and buffer-file-coding-system |
1288 | (not (eq (coding-system-type buffer-file-coding-system) t))) | |
1289 | ;; This is not `undecided'. | |
1290 | (setq local-coding (coding-system-base buffer-file-coding-system))) | |
b685f8d6 RS |
1291 | |
1292 | (if (and (local-variable-p 'buffer-file-coding-system) | |
1293 | local-eol local-coding) | |
4ed46869 KH |
1294 | ;; The current buffer has already set full coding-system, we |
1295 | ;; had better not change it. | |
1296 | nil | |
1297 | ||
8057896b | 1298 | (setq found-eol (coding-system-eol-type coding)) |
4ed46869 | 1299 | (if (null (numberp found-eol)) |
be02cd54 EZ |
1300 | ;; But eol-type is not found. |
1301 | ;; If EOL conversions are inhibited, force unix eol-type. | |
1302 | (setq found-eol (if inhibit-eol-conversion 0))) | |
c76b5c99 KH |
1303 | (if (eq (coding-system-type coding) t) |
1304 | (setq found-coding 'undecided) | |
1305 | (setq found-coding (coding-system-base coding))) | |
1306 | ||
1307 | (if (and (not found-eol) (eq found-coding 'undecided)) | |
1308 | ;; No valid coding information found. | |
1309 | nil | |
1310 | ||
1311 | ;; Some coding information (eol or text) found. | |
1312 | ||
1313 | ;; The local setting takes precedence over the found one. | |
1314 | (setq new-coding (if (local-variable-p 'buffer-file-coding-system) | |
1315 | (or local-coding found-coding) | |
1316 | (or found-coding local-coding))) | |
1317 | (setq new-eol (if (local-variable-p 'buffer-file-coding-system) | |
1318 | (or local-eol found-eol) | |
1319 | (or found-eol local-eol))) | |
1320 | ||
1321 | (let ((eol-type (coding-system-eol-type new-coding))) | |
1322 | (if (and (numberp new-eol) (vectorp eol-type)) | |
1323 | (aref eol-type new-eol) | |
1324 | new-coding))))))) | |
4ed46869 | 1325 | |
fe831d33 GV |
1326 | (defun modify-coding-system-alist (target-type regexp coding-system) |
1327 | "Modify one of look up tables for finding a coding system on I/O operation. | |
8c453b46 RS |
1328 | There are three of such tables, `file-coding-system-alist', |
1329 | `process-coding-system-alist', and `network-coding-system-alist'. | |
fe831d33 GV |
1330 | |
1331 | TARGET-TYPE specifies which of them to modify. | |
8c453b46 RS |
1332 | If it is `file', it affects `file-coding-system-alist' (which see). |
1333 | If it is `process', it affects `process-coding-system-alist' (which see). | |
e8dd0160 | 1334 | If it is `network', it affects `network-coding-system-alist' (which see). |
fe831d33 GV |
1335 | |
1336 | REGEXP is a regular expression matching a target of I/O operation. | |
1337 | The target is a file name if TARGET-TYPE is `file', a program name if | |
1338 | TARGET-TYPE is `process', or a network service name or a port number | |
1339 | to connect to if TARGET-TYPE is `network'. | |
1340 | ||
1341 | CODING-SYSTEM is a coding system to perform code conversion on the I/O | |
8c453b46 RS |
1342 | operation, or a cons cell (DECODING . ENCODING) specifying the coding systems |
1343 | for decoding and encoding respectively, | |
1344 | or a function symbol which, when called, returns such a cons cell." | |
fe831d33 GV |
1345 | (or (memq target-type '(file process network)) |
1346 | (error "Invalid target type: %s" target-type)) | |
1347 | (or (stringp regexp) | |
1348 | (and (eq target-type 'network) (integerp regexp)) | |
1349 | (error "Invalid regular expression: %s" regexp)) | |
1350 | (if (symbolp coding-system) | |
1351 | (if (not (fboundp coding-system)) | |
1352 | (progn | |
1353 | (check-coding-system coding-system) | |
1354 | (setq coding-system (cons coding-system coding-system)))) | |
1355 | (check-coding-system (car coding-system)) | |
1356 | (check-coding-system (cdr coding-system))) | |
1357 | (cond ((eq target-type 'file) | |
1358 | (let ((slot (assoc regexp file-coding-system-alist))) | |
1359 | (if slot | |
1360 | (setcdr slot coding-system) | |
1361 | (setq file-coding-system-alist | |
1362 | (cons (cons regexp coding-system) | |
1363 | file-coding-system-alist))))) | |
1364 | ((eq target-type 'process) | |
1365 | (let ((slot (assoc regexp process-coding-system-alist))) | |
1366 | (if slot | |
1367 | (setcdr slot coding-system) | |
1368 | (setq process-coding-system-alist | |
1369 | (cons (cons regexp coding-system) | |
1370 | process-coding-system-alist))))) | |
1371 | (t | |
1372 | (let ((slot (assoc regexp network-coding-system-alist))) | |
1373 | (if slot | |
1374 | (setcdr slot coding-system) | |
1375 | (setq network-coding-system-alist | |
1376 | (cons (cons regexp coding-system) | |
1377 | network-coding-system-alist))))))) | |
1378 | ||
b25eef20 | 1379 | (defun make-translation-table (&rest args) |
f967223b | 1380 | "Make a translation table (char table) from arguments. |
13d5617d | 1381 | Each argument is a list of the form (FROM . TO), |
b25eef20 | 1382 | where FROM is a character to be translated to TO. |
13d5617d | 1383 | |
452fdb31 RS |
1384 | FROM can be a generic character (see `make-char'). In this case, TO is |
1385 | a generic character containing the same number of characters, or a | |
1386 | ordinary character. If FROM and TO are both generic characters, all | |
b25eef20 | 1387 | characters belonging to FROM are translated to characters belonging to TO |
13d5617d | 1388 | without changing their position code(s)." |
f967223b | 1389 | (let ((table (make-char-table 'translation-table)) |
a73a8c89 KH |
1390 | revlist) |
1391 | (while args | |
1392 | (let ((elts (car args))) | |
1393 | (while elts | |
13d5617d KH |
1394 | (let* ((from (car (car elts))) |
1395 | (from-i 0) ; degree of freedom of FROM | |
1396 | (from-rev (nreverse (split-char from))) | |
1397 | (to (cdr (car elts))) | |
1398 | (to-i 0) ; degree of freedom of TO | |
1399 | (to-rev (nreverse (split-char to)))) | |
1400 | ;; Check numbers of heading 0s in FROM-REV and TO-REV. | |
1401 | (while (eq (car from-rev) 0) | |
1402 | (setq from-i (1+ from-i) from-rev (cdr from-rev))) | |
1403 | (while (eq (car to-rev) 0) | |
1404 | (setq to-i (1+ to-i) to-rev (cdr to-rev))) | |
1405 | (if (and (/= from-i to-i) (/= to-i 0)) | |
1406 | (error "Invalid character pair (%d . %d)" from to)) | |
b25eef20 KH |
1407 | ;; If we have already translated TO to TO-ALT, FROM should |
1408 | ;; also be translated to TO-ALT. But, this is only if TO | |
1409 | ;; is a generic character or TO-ALT is not a generic | |
13d5617d KH |
1410 | ;; character. |
1411 | (let ((to-alt (aref table to))) | |
1412 | (if (and to-alt | |
1413 | (or (> to-i 0) (not (generic-char-p to-alt)))) | |
1414 | (setq to to-alt))) | |
1415 | (if (> from-i 0) | |
1416 | (set-char-table-default table from to) | |
1417 | (aset table from to)) | |
b25eef20 KH |
1418 | ;; If we have already translated some chars to FROM, they |
1419 | ;; should also be translated to TO. | |
a73a8c89 KH |
1420 | (let ((l (assq from revlist))) |
1421 | (if l | |
1422 | (let ((ch (car l))) | |
1423 | (setcar l to) | |
1424 | (setq l (cdr l)) | |
1425 | (while l | |
1426 | (aset table ch to) | |
1427 | (setq l (cdr l)) )))) | |
1428 | ;; Now update REVLIST. | |
1429 | (let ((l (assq to revlist))) | |
1430 | (if l | |
1431 | (setcdr l (cons from (cdr l))) | |
1432 | (setq revlist (cons (list to from) revlist))))) | |
1433 | (setq elts (cdr elts)))) | |
1434 | (setq args (cdr args))) | |
1435 | ;; Return TABLE just created. | |
1436 | table)) | |
1437 | ||
c76b5c99 KH |
1438 | (defun make-translation-table-from-vector (vec) |
1439 | "Make translation table from decoding vector VEC. | |
1440 | VEC is an array of 256 elements to map unibyte codes to multibyte characters. | |
1441 | See also the variable `nonascii-translation-table'." | |
1442 | (let ((table (make-char-table 'translation-table)) | |
1443 | (rev-table (make-char-table 'translation-table)) | |
1444 | (i 0) | |
1445 | ch) | |
1446 | (while (< i 256) | |
1447 | (setq ch (aref vec i)) | |
1448 | (aset table i ch) | |
1449 | (if (>= ch 256) | |
1450 | (aset rev-table ch i)) | |
1451 | (setq i (1+ i))) | |
1452 | (set-char-table-extra-slot table 0 rev-table) | |
1453 | table)) | |
1454 | ||
f967223b | 1455 | (defun define-translation-table (symbol &rest args) |
e8dd0160 | 1456 | "Define SYMBOL as a name of translation table made by ARGS. |
b25eef20 | 1457 | |
007c79c8 KH |
1458 | If the first element of ARGS is a char-table of which purpose is |
1459 | translation-table, just define SYMBOL as the name of it. | |
1460 | ||
1461 | In the other case, ARGS are the same as arguments to the function | |
1462 | `make-translation-table' (which see). | |
b25eef20 | 1463 | |
452fdb31 RS |
1464 | This function sets properties `translation-table' and |
1465 | `translation-table-id' of SYMBOL to the created table itself and | |
f967223b | 1466 | identification number of the table respectively." |
007c79c8 KH |
1467 | (let ((table (if (and (char-table-p (car args)) |
1468 | (eq (char-table-subtype (car args)) | |
1469 | 'translation-table)) | |
1470 | (car args) | |
1471 | (apply 'make-translation-table args))) | |
f967223b | 1472 | (len (length translation-table-vector)) |
d9e3229d | 1473 | (id 0) |
b25eef20 | 1474 | (done nil)) |
f967223b | 1475 | (put symbol 'translation-table table) |
b25eef20 KH |
1476 | (while (not done) |
1477 | (if (>= id len) | |
f967223b KH |
1478 | (setq translation-table-vector |
1479 | (vconcat translation-table-vector (make-vector len nil)))) | |
1480 | (let ((slot (aref translation-table-vector id))) | |
b25eef20 KH |
1481 | (if (or (not slot) |
1482 | (eq (car slot) symbol)) | |
1483 | (progn | |
f967223b | 1484 | (aset translation-table-vector id (cons symbol table)) |
007c79c8 KH |
1485 | (setq done t)) |
1486 | (setq id (1+ id))))) | |
f967223b | 1487 | (put symbol 'translation-table-id id) |
d9e3229d KH |
1488 | id)) |
1489 | ||
35554641 KH |
1490 | (put 'with-category-table 'lisp-indent-function 1) |
1491 | ||
1492 | (defmacro with-category-table (category-table &rest body) | |
1493 | `(let ((current-category-table (category-table))) | |
1494 | (set-category-table ,category-table) | |
1495 | (unwind-protect | |
1496 | (progn ,@body) | |
1497 | (set-category-table current-category-table)))) | |
1498 | ||
69eba008 KH |
1499 | ;;; Initialize some variables. |
1500 | ||
1501 | (put 'use-default-ascent 'char-table-extra-slots 0) | |
1502 | (setq use-default-ascent (make-char-table 'use-default-ascent)) | |
d6d6d592 KH |
1503 | (put 'ignore-relative-composition 'char-table-extra-slots 0) |
1504 | (setq ignore-relative-composition | |
1505 | (make-char-table 'ignore-relative-composition)) | |
69eba008 KH |
1506 | |
1507 | ;;; | |
4ed46869 KH |
1508 | (provide 'mule) |
1509 | ||
1510 | ;;; mule.el ends here |