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. |
32db08f1 | 5 | ;; Copyright (C) 2001, 2002 Free Software Foundation, Inc. |
4ed46869 KH |
6 | |
7 | ;; Keywords: mule, multilingual, character set, coding system | |
8 | ||
9 | ;; This file is part of GNU Emacs. | |
10 | ||
11 | ;; GNU Emacs is free software; you can redistribute it and/or modify | |
12 | ;; it under the terms of the GNU General Public License as published by | |
13 | ;; the Free Software Foundation; either version 2, or (at your option) | |
14 | ;; any later version. | |
15 | ||
16 | ;; GNU Emacs is distributed in the hope that it will be useful, | |
17 | ;; but WITHOUT ANY WARRANTY; without even the implied warranty of | |
18 | ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
19 | ;; GNU General Public License for more details. | |
20 | ||
21 | ;; You should have received a copy of the GNU General Public License | |
369314dc KH |
22 | ;; along with GNU Emacs; see the file COPYING. If not, write to the |
23 | ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
24 | ;; Boston, MA 02111-1307, USA. | |
4ed46869 | 25 | |
60370d40 PJ |
26 | ;;; Commentary: |
27 | ||
4ed46869 KH |
28 | ;;; Code: |
29 | ||
d2542b42 | 30 | (defconst mule-version "5.0 (SAKAKI)" "\ |
4ed46869 KH |
31 | Version number and name of this version of MULE (multilingual environment).") |
32 | ||
35554641 | 33 | (defconst mule-version-date "1999.12.7" "\ |
4ed46869 KH |
34 | Distribution date of this version of MULE (multilingual environment).") |
35 | ||
36 | (defun load-with-code-conversion (fullname file &optional noerror nomessage) | |
0f69cb38 KH |
37 | "Execute a file of Lisp code named FILE whose absolute name is FULLNAME. |
38 | The file contents are decoded before evaluation if necessary. | |
4ed46869 KH |
39 | If optional second arg NOERROR is non-nil, |
40 | report no error if FILE doesn't exist. | |
41 | Print messages at start and end of loading unless | |
42 | optional third arg NOMESSAGE is non-nil. | |
43 | Return t if file exists." | |
44 | (if (null (file-readable-p fullname)) | |
45 | (and (null noerror) | |
46 | (signal 'file-error (list "Cannot open load file" file))) | |
47 | ;; Read file with code conversion, and then eval. | |
48 | (let* ((buffer | |
49 | ;; To avoid any autoloading, set default-major-mode to | |
50 | ;; fundamental-mode. | |
88162676 RS |
51 | ;; So that we don't get completely screwed if the |
52 | ;; file is encoded in some complicated character set, | |
53 | ;; read it with real decoding, as a multibyte buffer, | |
54 | ;; even if this is a --unibyte Emacs session. | |
55 | (let ((default-major-mode 'fundamental-mode) | |
56 | (default-enable-multibyte-characters t)) | |
4ed46869 KH |
57 | ;; We can't use `generate-new-buffer' because files.el |
58 | ;; is not yet loaded. | |
59 | (get-buffer-create (generate-new-buffer-name " *load*")))) | |
db5cae4b SM |
60 | (load-in-progress t) |
61 | (source (save-match-data (string-match "\\.el\\'" fullname)))) | |
62 | (unless nomessage | |
63 | (if source | |
64 | (message "Loading %s (source)..." file) | |
65 | (message "Loading %s..." file))) | |
66 | (when purify-flag | |
67 | (setq preloaded-file-list (cons file preloaded-file-list))) | |
4ed46869 | 68 | (unwind-protect |
a6acd8a2 | 69 | (let ((load-file-name fullname) |
1c4cc63a | 70 | (set-auto-coding-for-load t) |
a6acd8a2 | 71 | (inhibit-file-name-operation nil)) |
4ed46869 KH |
72 | (save-excursion |
73 | (set-buffer buffer) | |
74 | (insert-file-contents fullname) | |
7d276780 EZ |
75 | ;; If the loaded file was inserted with no-conversion or |
76 | ;; raw-text coding system, make the buffer unibyte. | |
77 | ;; Otherwise, eval-buffer might try to interpret random | |
78 | ;; binary junk as multibyte characters. | |
79 | (if (and enable-multibyte-characters | |
80 | (or (eq (coding-system-type last-coding-system-used) 5) | |
81 | (eq last-coding-system-used 'no-conversion))) | |
82 | (set-buffer-multibyte nil)) | |
4ed46869 KH |
83 | ;; Make `kill-buffer' quiet. |
84 | (set-buffer-modified-p nil)) | |
0f69cb38 | 85 | ;; Have the original buffer current while we eval. |
88162676 RS |
86 | (eval-buffer buffer nil file |
87 | ;; If this Emacs is running with --unibyte, | |
88 | ;; convert multibyte strings to unibyte | |
89 | ;; after reading them. | |
ba74e833 | 90 | ;; (not default-enable-multibyte-characters) |
8dd08b5b | 91 | nil t |
ba74e833 | 92 | )) |
cfc70cdf RS |
93 | (let (kill-buffer-hook kill-buffer-query-functions) |
94 | (kill-buffer buffer))) | |
4ed46869 | 95 | (let ((hook (assoc file after-load-alist))) |
db5cae4b SM |
96 | (when hook |
97 | (mapcar (function eval) (cdr hook)))) | |
98 | (unless (or nomessage noninteractive) | |
99 | (if source | |
100 | (message "Loading %s (source)...done" file) | |
101 | (message "Loading %s...done" file))) | |
4ed46869 KH |
102 | t))) |
103 | ||
104 | ;; API (Application Program Interface) for charsets. | |
105 | ||
2e02a76f | 106 | (defsubst charset-quoted-standard-p (obj) |
521d4010 | 107 | "Return t if OBJ is a quoted symbol, and is the name of a standard charset." |
2e02a76f RS |
108 | (and (listp obj) (eq (car obj) 'quote) |
109 | (symbolp (car-safe (cdr obj))) | |
110 | (let ((vector (get (car-safe (cdr obj)) 'charset))) | |
111 | (and (vectorp vector) | |
112 | (< (aref vector 0) 160))))) | |
4ed46869 KH |
113 | |
114 | (defsubst charsetp (object) | |
e4a7fdfc | 115 | "T if OBJECT is a charset." |
4ed46869 KH |
116 | (and (symbolp object) (vectorp (get object 'charset)))) |
117 | ||
118 | (defsubst charset-info (charset) | |
119 | "Return a vector of information of CHARSET. | |
120 | The elements of the vector are: | |
121 | CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION, | |
122 | LEADING-CODE-BASE, LEADING-CODE-EXT, | |
123 | ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE, | |
124 | REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION, | |
125 | PLIST, | |
126 | where | |
127 | CHARSET-ID (integer) is the identification number of the charset. | |
aa8bb645 KH |
128 | BYTES (integer) is the length of multi-byte form of a character in |
129 | the charset: one of 1, 2, 3, and 4. | |
4ed46869 KH |
130 | DIMENSION (integer) is the number of bytes to represent a character of |
131 | the charset: 1 or 2. | |
132 | CHARS (integer) is the number of characters in a dimension: 94 or 96. | |
4ed46869 KH |
133 | WIDTH (integer) is the number of columns a character in the charset |
134 | occupies on the screen: one of 0, 1, and 2. | |
135 | DIRECTION (integer) is the rendering direction of characters in the | |
08b4ace5 KH |
136 | charset when rendering. If 0, render from left to right, else |
137 | render from right to left. | |
4ed46869 KH |
138 | LEADING-CODE-BASE (integer) is the base leading-code for the |
139 | charset. | |
140 | LEADING-CODE-EXT (integer) is the extended leading-code for the | |
141 | charset. All charsets of less than 0xA0 has the value 0. | |
142 | ISO-FINAL-CHAR (character) is the final character of the | |
7dd4c92d KH |
143 | corresponding ISO 2022 charset. If the charset is not assigned |
144 | any final character, the value is -1. | |
4ed46869 KH |
145 | ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked |
146 | while encoding to variants of ISO 2022 coding system, one of the | |
147 | following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR). | |
7dd4c92d | 148 | If the charset is not assigned any final character, the value is -1. |
4ed46869 KH |
149 | REVERSE-CHARSET (integer) is the charset which differs only in |
150 | LEFT-TO-RIGHT value from the charset. If there's no such a | |
151 | charset, the value is -1. | |
152 | SHORT-NAME (string) is the short name to refer to the charset. | |
153 | LONG-NAME (string) is the long name to refer to the charset | |
154 | DESCRIPTION (string) is the description string of the charset. | |
155 | PLIST (property list) may contain any type of information a user | |
156 | want to put and get by functions `put-charset-property' and | |
157 | `get-charset-property' respectively." | |
158 | (get charset 'charset)) | |
159 | ||
40c81f74 PE |
160 | ;; It is better not to use backquote in this file, |
161 | ;; because that makes a bootstrapping problem | |
162 | ;; if you need to recompile all the Lisp files using interpreted code. | |
163 | ||
4ed46869 KH |
164 | (defmacro charset-id (charset) |
165 | "Return charset identification number of CHARSET." | |
2e02a76f | 166 | (if (charset-quoted-standard-p charset) |
4ed46869 | 167 | (aref (charset-info (nth 1 charset)) 0) |
40c81f74 | 168 | (list 'aref (list 'charset-info charset) 0))) |
4ed46869 KH |
169 | |
170 | (defmacro charset-bytes (charset) | |
900dc6e3 KH |
171 | "Return bytes of CHARSET. |
172 | See the function `charset-info' for more detail." | |
2e02a76f | 173 | (if (charset-quoted-standard-p charset) |
4ed46869 | 174 | (aref (charset-info (nth 1 charset)) 1) |
40c81f74 | 175 | (list 'aref (list 'charset-info charset) 1))) |
4ed46869 KH |
176 | |
177 | (defmacro charset-dimension (charset) | |
900dc6e3 KH |
178 | "Return dimension of CHARSET. |
179 | See the function `charset-info' for more detail." | |
2e02a76f | 180 | (if (charset-quoted-standard-p charset) |
4ed46869 | 181 | (aref (charset-info (nth 1 charset)) 2) |
40c81f74 | 182 | (list 'aref (list 'charset-info charset) 2))) |
4ed46869 KH |
183 | |
184 | (defmacro charset-chars (charset) | |
900dc6e3 KH |
185 | "Return character numbers contained in a dimension of CHARSET. |
186 | See the function `charset-info' for more detail." | |
2e02a76f | 187 | (if (charset-quoted-standard-p charset) |
4ed46869 | 188 | (aref (charset-info (nth 1 charset)) 3) |
40c81f74 | 189 | (list 'aref (list 'charset-info charset) 3))) |
4ed46869 KH |
190 | |
191 | (defmacro charset-width (charset) | |
900dc6e3 KH |
192 | "Return width (how many column occupied on a screen) of CHARSET. |
193 | See the function `charset-info' for more detail." | |
2e02a76f | 194 | (if (charset-quoted-standard-p charset) |
4ed46869 | 195 | (aref (charset-info (nth 1 charset)) 4) |
40c81f74 | 196 | (list 'aref (list 'charset-info charset) 4))) |
4ed46869 KH |
197 | |
198 | (defmacro charset-direction (charset) | |
900dc6e3 KH |
199 | "Return direction of CHARSET. |
200 | See the function `charset-info' for more detail." | |
2e02a76f | 201 | (if (charset-quoted-standard-p charset) |
4ed46869 | 202 | (aref (charset-info (nth 1 charset)) 5) |
40c81f74 | 203 | (list 'aref (list 'charset-info charset) 5))) |
4ed46869 KH |
204 | |
205 | (defmacro charset-iso-final-char (charset) | |
900dc6e3 KH |
206 | "Return final char of CHARSET. |
207 | See the function `charset-info' for more detail." | |
2e02a76f | 208 | (if (charset-quoted-standard-p charset) |
4ed46869 | 209 | (aref (charset-info (nth 1 charset)) 8) |
40c81f74 | 210 | (list 'aref (list 'charset-info charset) 8))) |
4ed46869 KH |
211 | |
212 | (defmacro charset-iso-graphic-plane (charset) | |
900dc6e3 KH |
213 | "Return graphic plane of CHARSET. |
214 | See the function `charset-info' for more detail." | |
2e02a76f | 215 | (if (charset-quoted-standard-p charset) |
4ed46869 | 216 | (aref (charset-info (nth 1 charset)) 9) |
40c81f74 | 217 | (list 'aref (list 'charset-info charset) 9))) |
4ed46869 KH |
218 | |
219 | (defmacro charset-reverse-charset (charset) | |
900dc6e3 KH |
220 | "Return reverse charset of CHARSET. |
221 | See the function `charset-info' for more detail." | |
2e02a76f | 222 | (if (charset-quoted-standard-p charset) |
4ed46869 | 223 | (aref (charset-info (nth 1 charset)) 10) |
40c81f74 | 224 | (list 'aref (list 'charset-info charset) 10))) |
4ed46869 KH |
225 | |
226 | (defmacro charset-short-name (charset) | |
900dc6e3 KH |
227 | "Return short name of CHARSET. |
228 | See the function `charset-info' for more detail." | |
2e02a76f | 229 | (if (charset-quoted-standard-p charset) |
4ed46869 | 230 | (aref (charset-info (nth 1 charset)) 11) |
40c81f74 | 231 | (list 'aref (list 'charset-info charset) 11))) |
4ed46869 KH |
232 | |
233 | (defmacro charset-long-name (charset) | |
900dc6e3 KH |
234 | "Return long name of CHARSET. |
235 | See the function `charset-info' for more detail." | |
2e02a76f | 236 | (if (charset-quoted-standard-p charset) |
4ed46869 | 237 | (aref (charset-info (nth 1 charset)) 12) |
40c81f74 | 238 | (list 'aref (list 'charset-info charset) 12))) |
4ed46869 KH |
239 | |
240 | (defmacro charset-description (charset) | |
0611934b | 241 | "Return description of CHARSET. |
900dc6e3 | 242 | See the function `charset-info' for more detail." |
2e02a76f | 243 | (if (charset-quoted-standard-p charset) |
4ed46869 | 244 | (aref (charset-info (nth 1 charset)) 13) |
40c81f74 | 245 | (list 'aref (list 'charset-info charset) 13))) |
4ed46869 KH |
246 | |
247 | (defmacro charset-plist (charset) | |
900dc6e3 KH |
248 | "Return list charset property of CHARSET. |
249 | See the function `charset-info' for more detail." | |
40c81f74 PE |
250 | (list 'aref |
251 | (if (charset-quoted-standard-p charset) | |
252 | (charset-info (nth 1 charset)) | |
253 | (list 'charset-info charset)) | |
254 | 14)) | |
4ed46869 KH |
255 | |
256 | (defun set-charset-plist (charset plist) | |
e8dd0160 | 257 | "Set CHARSET's property list to PLIST, and return PLIST." |
4ed46869 KH |
258 | (aset (charset-info charset) 14 plist)) |
259 | ||
d3e7e7cf EZ |
260 | (defun make-char (charset &optional code1 code2) |
261 | "Return a character of CHARSET whose position codes are CODE1 and CODE2. | |
f98e2797 | 262 | CODE1 and CODE2 are optional, but if you don't supply |
a9fb0b58 KH |
263 | sufficient position codes, return a generic character which stands for |
264 | all characters or group of characters in the character set. | |
d3e7e7cf | 265 | A generic character can be used to index a char table (e.g. syntax-table). |
9ff05eae KH |
266 | |
267 | Such character sets as ascii, eight-bit-control, and eight-bit-graphic | |
268 | don't have corresponding generic characters. If CHARSET is one of | |
269 | them and you don't supply CODE1, return the character of the smallest | |
270 | code in CHARSET. | |
271 | ||
6fb7766c DL |
272 | If CODE1 or CODE2 are invalid (out of range), this function signals an |
273 | error. However, the eighth bit of both CODE1 and CODE2 is zeroed | |
274 | before they are used to index CHARSET. Thus you may use, say, the | |
275 | actual ISO 8859 character code rather than subtracting 128, as you | |
276 | would need to index the corresponding Emacs charset." | |
636799f2 | 277 | (make-char-internal (charset-id charset) code1 code2)) |
a6acd8a2 KH |
278 | |
279 | (put 'make-char 'byte-compile | |
521d4010 | 280 | (function |
a6acd8a2 KH |
281 | (lambda (form) |
282 | (let ((charset (nth 1 form))) | |
283 | (if (charset-quoted-standard-p charset) | |
284 | (byte-compile-normal-call | |
285 | (cons 'make-char-internal | |
286 | (cons (charset-id (nth 1 charset)) (nthcdr 2 form)))) | |
287 | (byte-compile-normal-call | |
288 | (cons 'make-char-internal | |
289 | (cons (list 'charset-id charset) (nthcdr 2 form))))))))) | |
4ed46869 | 290 | |
d3675a42 | 291 | (defun charset-list () |
900dc6e3 KH |
292 | "Return list of charsets ever defined. |
293 | ||
d3675a42 | 294 | This function is provided for backward compatibility. |
900dc6e3 | 295 | Now we have the variable `charset-list'." |
d3675a42 KH |
296 | charset-list) |
297 | ||
13d5617d KH |
298 | (defsubst generic-char-p (char) |
299 | "Return t if and only if CHAR is a generic character. | |
521d4010 | 300 | See also the documentation of `make-char'." |
1d935669 KH |
301 | (and (>= char 0400) |
302 | (let ((l (split-char char))) | |
303 | (and (or (= (nth 1 l) 0) (eq (nth 2 l) 0)) | |
304 | (not (eq (car l) 'composition)))))) | |
69eba008 | 305 | |
cc57cc54 | 306 | (defun decode-char (ccs code-point &optional restriction) |
7f341749 | 307 | "Return character specified by coded character set CCS and CODE-POINT in it. |
cc57cc54 | 308 | Return nil if such a character is not supported. |
e76938e7 | 309 | Currently the only supported coded character set is `ucs' (ISO/IEC |
9fb1b1a8 DL |
310 | 10646: Universal Multi-Octet Coded Character Set), and the result is |
311 | translated through the char table `utf-8-translation-table-for-decode'. | |
cc57cc54 KH |
312 | |
313 | Optional argument RESTRICTION specifies a way to map the pair of CCS | |
f7492bc7 | 314 | and CODE-POINT to a character. Currently not supported and just ignored." |
9fb1b1a8 DL |
315 | (cond |
316 | ((eq ccs 'ucs) | |
317 | (let ((c (cond | |
318 | ((< code-point 160) | |
319 | code-point) | |
320 | ((< code-point 256) | |
321 | (make-char 'latin-iso8859-1 code-point)) | |
322 | ((< code-point #x2500) | |
323 | (setq code-point (- code-point #x0100)) | |
324 | (make-char 'mule-unicode-0100-24ff | |
325 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
326 | ((< code-point #x3400) | |
327 | (setq code-point (- code-point #x2500)) | |
328 | (make-char 'mule-unicode-2500-33ff | |
329 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
330 | ((and (>= code-point #xe000) (< code-point #x10000)) | |
331 | (setq code-point (- code-point #xe000)) | |
332 | (make-char 'mule-unicode-e000-ffff | |
333 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32)))))) | |
334 | (if (and c (aref utf-8-translation-table-for-decode c)) | |
335 | (aref utf-8-translation-table-for-decode c) | |
336 | c))))) | |
cc57cc54 KH |
337 | |
338 | (defun encode-char (char ccs &optional restriction) | |
7f341749 | 339 | "Return code-point in coded character set CCS that corresponds to CHAR. |
cc57cc54 | 340 | Return nil if CHAR is not included in CCS. |
e76938e7 | 341 | Currently the only supported coded character set is `ucs' (ISO/IEC |
9fb1b1a8 DL |
342 | 10646: Universal Multi-Octet Coded Character Set), and CHAR is first |
343 | translated through the char-table `ucs-mule-to-mule-unicode'. | |
7f341749 KH |
344 | |
345 | CHAR should be in one of these charsets: | |
cc57cc54 | 346 | ascii, latin-iso8859-1, mule-unicode-0100-24ff, mule-unicode-2500-33ff, |
6ba9b20b | 347 | mule-unicode-e000-ffff, eight-bit-control |
cc57cc54 KH |
348 | Otherwise, return nil. |
349 | ||
350 | Optional argument RESTRICTION specifies a way to map CHAR to a | |
351 | code-point in CCS. Currently not supported and just ignored." | |
352 | (let* ((split (split-char char)) | |
9fb1b1a8 DL |
353 | (charset (car split)) |
354 | trans) | |
cc57cc54 | 355 | (cond ((eq ccs 'ucs) |
9fb1b1a8 DL |
356 | (setq trans (aref ucs-mule-to-mule-unicode char)) |
357 | (if trans | |
358 | (setq split (split-char trans) | |
359 | charset (car split))) | |
cc57cc54 KH |
360 | (cond ((eq charset 'ascii) |
361 | char) | |
362 | ((eq charset 'latin-iso8859-1) | |
363 | (+ (nth 1 split) 128)) | |
364 | ((eq charset 'mule-unicode-0100-24ff) | |
365 | (+ #x0100 (+ (* (- (nth 1 split) 32) 96) | |
366 | (- (nth 2 split) 32)))) | |
367 | ((eq charset 'mule-unicode-2500-33ff) | |
368 | (+ #x2500 (+ (* (- (nth 1 split) 32) 96) | |
369 | (- (nth 2 split) 32)))) | |
370 | ((eq charset 'mule-unicode-e000-ffff) | |
371 | (+ #xe000 (+ (* (- (nth 1 split) 32) 96) | |
6ba9b20b KH |
372 | (- (nth 2 split) 32)))) |
373 | ((eq charset 'eight-bit-control) | |
374 | char)))))) | |
cc57cc54 | 375 | |
0269ddfb | 376 | \f |
e76938e7 | 377 | ;; Coding system stuff |
4ed46869 | 378 | |
8057896b | 379 | ;; Coding system is a symbol that has the property `coding-system'. |
4ed46869 | 380 | ;; |
8057896b KH |
381 | ;; The value of the property `coding-system' is a vector of the |
382 | ;; following format: | |
d3675a42 | 383 | ;; [TYPE MNEMONIC DOC-STRING PLIST FLAGS] |
8057896b | 384 | ;; We call this vector as coding-spec. See comments in src/coding.c |
521d4010 | 385 | ;; for more detail. |
8057896b KH |
386 | |
387 | (defconst coding-spec-type-idx 0) | |
388 | (defconst coding-spec-mnemonic-idx 1) | |
389 | (defconst coding-spec-doc-string-idx 2) | |
0269ddfb | 390 | (defconst coding-spec-plist-idx 3) |
8057896b KH |
391 | (defconst coding-spec-flags-idx 4) |
392 | ||
a6acd8a2 KH |
393 | ;; PLIST is a property list of a coding system. To share PLIST among |
394 | ;; alias coding systems, a coding system has PLIST in coding-spec | |
395 | ;; instead of having it in normal property list of Lisp symbol. | |
396 | ;; Here's a list of coding system properties currently being used. | |
0269ddfb KH |
397 | ;; |
398 | ;; o coding-category | |
399 | ;; | |
400 | ;; The value is a coding category the coding system belongs to. The | |
521d4010 | 401 | ;; function `make-coding-system' sets this value automatically |
1b46a680 | 402 | ;; unless its argument PROPERTIES specifies this property. |
4ed46869 | 403 | ;; |
0269ddfb | 404 | ;; o alias-coding-systems |
4ed46869 | 405 | ;; |
0269ddfb KH |
406 | ;; The value is a list of coding systems of the same alias group. The |
407 | ;; first element is the coding system made at first, which we call as | |
1b46a680 KH |
408 | ;; `base coding system'. The function `make-coding-system' sets this |
409 | ;; value automatically and `define-coding-system-alias' updates it. | |
0269ddfb | 410 | ;; |
3c948dde KH |
411 | ;; See the documentation of make-coding-system for the meanings of the |
412 | ;; following properties. | |
4ed46869 | 413 | ;; |
3c948dde | 414 | ;; o post-read-conversion |
0269ddfb | 415 | ;; o pre-write-conversion |
f967223b | 416 | ;; o translation-table-for-decode |
f967223b | 417 | ;; o translation-table-for-encode |
c11a8f77 | 418 | ;; o safe-chars |
a6acd8a2 | 419 | ;; o safe-charsets |
b25eef20 | 420 | ;; o mime-charset |
c76b5c99 | 421 | ;; o valid-codes (meaningful only for a coding system based on CCL) |
c76b5c99 | 422 | |
0269ddfb | 423 | |
0269ddfb | 424 | (defsubst coding-system-spec (coding-system) |
521d4010 | 425 | "Return coding-spec of CODING-SYSTEM." |
0269ddfb | 426 | (get (check-coding-system coding-system) 'coding-system)) |
4ed46869 | 427 | |
4ed46869 | 428 | (defun coding-system-type (coding-system) |
0269ddfb KH |
429 | "Return the coding type of CODING-SYSTEM. |
430 | A coding type is an integer value indicating the encoding method | |
431 | of CODING-SYSTEM. See the function `make-coding-system' for more detail." | |
2653e2a7 | 432 | (aref (coding-system-spec coding-system) coding-spec-type-idx)) |
4ed46869 | 433 | |
4ed46869 | 434 | (defun coding-system-mnemonic (coding-system) |
0269ddfb | 435 | "Return the mnemonic character of CODING-SYSTEM. |
6e2c8840 KH |
436 | The mnemonic character of a coding system is used in mode line |
437 | to indicate the coding system. If the arg is nil, return ?-." | |
438 | (let ((spec (coding-system-spec coding-system))) | |
439 | (if spec (aref spec coding-spec-mnemonic-idx) ?-))) | |
4ed46869 | 440 | |
8057896b | 441 | (defun coding-system-doc-string (coding-system) |
0269ddfb | 442 | "Return the documentation string for CODING-SYSTEM." |
2653e2a7 | 443 | (aref (coding-system-spec coding-system) coding-spec-doc-string-idx)) |
4ed46869 | 444 | |
d3675a42 | 445 | (defun coding-system-plist (coding-system) |
0269ddfb | 446 | "Return the property list of CODING-SYSTEM." |
2653e2a7 | 447 | (aref (coding-system-spec coding-system) coding-spec-plist-idx)) |
d3675a42 | 448 | |
4ed46869 | 449 | (defun coding-system-flags (coding-system) |
0269ddfb KH |
450 | "Return `flags' of CODING-SYSTEM. |
451 | A `flags' of a coding system is a vector of length 32 indicating detailed | |
452 | information of a coding system. See the function `make-coding-system' | |
453 | for more detail." | |
2653e2a7 | 454 | (aref (coding-system-spec coding-system) coding-spec-flags-idx)) |
0269ddfb KH |
455 | |
456 | (defun coding-system-get (coding-system prop) | |
457 | "Extract a value from CODING-SYSTEM's property list for property PROP." | |
458 | (plist-get (coding-system-plist coding-system) prop)) | |
459 | ||
460 | (defun coding-system-put (coding-system prop val) | |
461 | "Change value in CODING-SYSTEM's property list PROP to VAL." | |
462 | (let ((plist (coding-system-plist coding-system))) | |
463 | (if plist | |
464 | (plist-put plist prop val) | |
2653e2a7 KH |
465 | (aset (coding-system-spec coding-system) coding-spec-plist-idx |
466 | (list prop val))))) | |
0269ddfb KH |
467 | |
468 | (defun coding-system-category (coding-system) | |
72f73748 DL |
469 | "Return the coding category of CODING-SYSTEM. |
470 | See also `coding-category-list'." | |
0269ddfb KH |
471 | (coding-system-get coding-system 'coding-category)) |
472 | ||
473 | (defun coding-system-base (coding-system) | |
474 | "Return the base coding system of CODING-SYSTEM. | |
d9e3229d KH |
475 | A base coding system is what made by `make-coding-system'. |
476 | Any alias nor subsidiary coding systems are not base coding system." | |
0269ddfb KH |
477 | (car (coding-system-get coding-system 'alias-coding-systems))) |
478 | ||
479 | (defalias 'coding-system-parent 'coding-system-base) | |
2598a293 | 480 | (make-obsolete 'coding-system-parent 'coding-system-base "20.3") |
0269ddfb KH |
481 | |
482 | ;; Coding system also has a property `eol-type'. | |
483 | ;; | |
484 | ;; This property indicates how the coding system handles end-of-line | |
485 | ;; format. The value is integer 0, 1, 2, or a vector of three coding | |
486 | ;; systems. Each integer value 0, 1, and 2 indicates the format of | |
487 | ;; end-of-line LF, CRLF, and CR respectively. A vector value | |
488 | ;; indicates that the format of end-of-line should be detected | |
489 | ;; automatically. Nth element of the vector is the subsidiary coding | |
490 | ;; system whose `eol-type' property is N. | |
4ed46869 | 491 | |
8057896b | 492 | (defun coding-system-eol-type (coding-system) |
0269ddfb KH |
493 | "Return eol-type of CODING-SYSTEM. |
494 | An eol-type is integer 0, 1, 2, or a vector of coding systems. | |
4ed46869 | 495 | |
0269ddfb KH |
496 | Integer values 0, 1, and 2 indicate a format of end-of-line; LF, |
497 | CRLF, and CR respectively. | |
498 | ||
499 | A vector value indicates that a format of end-of-line should be | |
500 | detected automatically. Nth element of the vector is the subsidiary | |
501 | coding system whose eol-type is N." | |
502 | (get coding-system 'eol-type)) | |
bd882697 | 503 | |
857ea15c AS |
504 | (defun coding-system-lessp (x y) |
505 | (cond ((eq x 'no-conversion) t) | |
506 | ((eq y 'no-conversion) nil) | |
507 | ((eq x 'emacs-mule) t) | |
508 | ((eq y 'emacs-mule) nil) | |
509 | ((eq x 'undecided) t) | |
510 | ((eq y 'undecided) nil) | |
511 | (t (let ((c1 (coding-system-mnemonic x)) | |
512 | (c2 (coding-system-mnemonic y))) | |
513 | (or (< (downcase c1) (downcase c2)) | |
514 | (and (not (> (downcase c1) (downcase c2))) | |
515 | (< c1 c2))))))) | |
516 | ||
857ea15c | 517 | (defun add-to-coding-system-list (coding-system) |
521d4010 | 518 | "Add CODING-SYSTEM to `coding-system-list' while keeping it sorted." |
857ea15c AS |
519 | (if (or (null coding-system-list) |
520 | (coding-system-lessp coding-system (car coding-system-list))) | |
521 | (setq coding-system-list (cons coding-system coding-system-list)) | |
522 | (let ((len (length coding-system-list)) | |
523 | mid (tem coding-system-list)) | |
524 | (while (> len 1) | |
525 | (setq mid (nthcdr (/ len 2) tem)) | |
526 | (if (coding-system-lessp (car mid) coding-system) | |
527 | (setq tem mid | |
528 | len (- len (/ len 2))) | |
529 | (setq len (/ len 2)))) | |
530 | (setcdr tem (cons coding-system (cdr tem)))))) | |
531 | ||
80a7463d | 532 | (defun coding-system-list (&optional base-only) |
c11a8f77 KH |
533 | "Return a list of all existing non-subsidiary coding systems. |
534 | If optional arg BASE-ONLY is non-nil, only base coding systems are listed. | |
535 | The value doesn't include subsidiary coding systems which are what | |
536 | made from bases and aliases automatically for various end-of-line | |
537 | formats (e.g. iso-latin-1-unix, koi8-r-dos)." | |
80a7463d KH |
538 | (let* ((codings (copy-sequence coding-system-list)) |
539 | (tail (cons nil codings))) | |
540 | ;; Remove subsidiary coding systems (eol variants) and alias | |
541 | ;; coding systems (if necessary). | |
542 | (while (cdr tail) | |
543 | (let* ((coding (car (cdr tail))) | |
544 | (aliases (coding-system-get coding 'alias-coding-systems))) | |
545 | (if (or | |
546 | ;; CODING is an eol variant if not in ALIASES. | |
547 | (not (memq coding aliases)) | |
548 | ;; CODING is an alias if it is not car of ALIASES. | |
549 | (and base-only (not (eq coding (car aliases))))) | |
550 | (setcdr tail (cdr (cdr tail))) | |
551 | (setq tail (cdr tail))))) | |
552 | codings)) | |
553 | ||
6fb7766c DL |
554 | (defun map-charset-chars (func charset) |
555 | "Use FUNC to map over all characters in CHARSET for side effects. | |
556 | FUNC is a function of two args, the start and end (inclusive) of a | |
557 | character code range. Thus FUNC should iterate over [START, END]." | |
558 | (let* ((dim (charset-dimension charset)) | |
559 | (chars (charset-chars charset)) | |
560 | (start (if (= chars 94) | |
561 | 33 | |
562 | 32))) | |
563 | (if (= dim 1) | |
564 | (funcall func | |
565 | (make-char charset start) | |
566 | (make-char charset (+ start chars -1))) | |
567 | (dotimes (i chars) | |
568 | (funcall func | |
569 | (make-char charset (+ i start) start) | |
570 | (make-char charset (+ i start) (+ start chars -1))))))) | |
571 | ||
c11a8f77 | 572 | (defun register-char-codings (coding-system safe-chars) |
6fb7766c DL |
573 | "Add entries for CODING-SYSTEM to `char-coding-system-table'. |
574 | If SAFE-CHARS is a char-table, its non-nil entries specify characters | |
575 | which CODING-SYSTEM encodes safely. If SAFE-CHARS is t, register | |
576 | CODING-SYSTEM as a general one which can encode all characters." | |
577 | (let ((general (char-table-extra-slot char-coding-system-table 0)) | |
578 | ;; Charsets which have some members in the table, but not all | |
579 | ;; of them (i.e. not just a generic character): | |
580 | (partials (char-table-extra-slot char-coding-system-table 1))) | |
c11a8f77 KH |
581 | (if (eq safe-chars t) |
582 | (or (memq coding-system general) | |
583 | (set-char-table-extra-slot char-coding-system-table 0 | |
584 | (cons coding-system general))) | |
585 | (map-char-table | |
6fb7766c DL |
586 | (lambda (key val) |
587 | (if (and (>= key 128) val) | |
588 | (let ((codings (aref char-coding-system-table key)) | |
589 | (charset (char-charset key))) | |
590 | (unless (memq coding-system codings) | |
591 | (if (and (generic-char-p key) | |
592 | (memq charset partials)) | |
593 | ;; The generic char would clobber individual | |
594 | ;; entries already in the table. First save the | |
595 | ;; separate existing entries for all chars of the | |
596 | ;; charset (with the generic entry added, if | |
597 | ;; necessary). | |
598 | (let (entry existing) | |
599 | (map-charset-chars | |
600 | (lambda (start end) | |
601 | (while (<= start end) | |
602 | (setq entry (aref char-coding-system-table start)) | |
603 | (when entry | |
604 | (push (cons | |
605 | start | |
606 | (if (memq coding-system entry) | |
607 | entry | |
608 | (cons coding-system entry))) | |
609 | existing)) | |
610 | (setq start (1+ start)))) | |
611 | charset) | |
612 | ;; Update the generic entry. | |
613 | (aset char-coding-system-table key | |
614 | (cons coding-system codings)) | |
615 | ;; Override with the saved entries. | |
616 | (dolist (elt existing) | |
617 | (aset char-coding-system-table (car elt) (cdr elt)))) | |
618 | (aset char-coding-system-table key | |
619 | (cons coding-system codings)) | |
620 | (unless (or (memq charset partials) | |
621 | (generic-char-p key)) | |
622 | (push charset partials))))))) | |
623 | safe-chars) | |
624 | (set-char-table-extra-slot char-coding-system-table 1 partials)))) | |
c11a8f77 KH |
625 | |
626 | ||
6e9722b0 | 627 | (defun make-subsidiary-coding-system (coding-system) |
521d4010 | 628 | "Make subsidiary coding systems (eol-type variants) of CODING-SYSTEM." |
0269ddfb KH |
629 | (let ((coding-spec (coding-system-spec coding-system)) |
630 | (subsidiaries (vector (intern (format "%s-unix" coding-system)) | |
8057896b KH |
631 | (intern (format "%s-dos" coding-system)) |
632 | (intern (format "%s-mac" coding-system)))) | |
0269ddfb KH |
633 | (i 0) |
634 | temp) | |
8057896b | 635 | (while (< i 3) |
0269ddfb | 636 | (put (aref subsidiaries i) 'coding-system coding-spec) |
8057896b | 637 | (put (aref subsidiaries i) 'eol-type i) |
857ea15c | 638 | (add-to-coding-system-list (aref subsidiaries i)) |
0269ddfb KH |
639 | (setq coding-system-alist |
640 | (cons (list (symbol-name (aref subsidiaries i))) | |
641 | coding-system-alist)) | |
8057896b KH |
642 | (setq i (1+ i))) |
643 | subsidiaries)) | |
4ed46869 | 644 | |
50c29104 KH |
645 | (defun transform-make-coding-system-args (name type &optional doc-string props) |
646 | "For internal use only. | |
647 | Transform XEmacs style args for `make-coding-system' to Emacs style. | |
648 | Value is a list of transformed arguments." | |
649 | (let ((mnemonic (string-to-char (or (plist-get props 'mnemonic) "?"))) | |
650 | (eol-type (plist-get props 'eol-type)) | |
651 | properties tmp) | |
652 | (cond | |
653 | ((eq eol-type 'lf) (setq eol-type 'unix)) | |
654 | ((eq eol-type 'crlf) (setq eol-type 'dos)) | |
655 | ((eq eol-type 'cr) (setq eol-type 'mac))) | |
656 | (if (setq tmp (plist-get props 'post-read-conversion)) | |
657 | (setq properties (plist-put properties 'post-read-conversion tmp))) | |
658 | (if (setq tmp (plist-get props 'pre-write-conversion)) | |
659 | (setq properties (plist-put properties 'pre-write-conversion tmp))) | |
660 | (cond | |
f4a012a6 KH |
661 | ((eq type 'shift-jis) |
662 | `(,name 1 ,mnemonic ,doc-string () ,properties ,eol-type)) | |
663 | ((eq type 'iso2022) ; This is not perfect. | |
664 | (if (plist-get props 'escape-quoted) | |
665 | (error "escape-quoted is not supported: %S" | |
666 | `(,name ,type ,doc-string ,props))) | |
667 | (let ((g0 (plist-get props 'charset-g0)) | |
668 | (g1 (plist-get props 'charset-g1)) | |
669 | (g2 (plist-get props 'charset-g2)) | |
670 | (g3 (plist-get props 'charset-g3)) | |
671 | (use-roman | |
672 | (and | |
673 | (eq (cadr (assoc 'latin-jisx0201 | |
674 | (plist-get props 'input-charset-conversion))) | |
675 | 'ascii) | |
676 | (eq (cadr (assoc 'ascii | |
677 | (plist-get props 'output-charset-conversion))) | |
678 | 'latin-jisx0201))) | |
679 | (use-oldjis | |
680 | (and | |
681 | (eq (cadr (assoc 'japanese-jisx0208-1978 | |
682 | (plist-get props 'input-charset-conversion))) | |
683 | 'japanese-jisx0208) | |
684 | (eq (cadr (assoc 'japanese-jisx0208 | |
685 | (plist-get props 'output-charset-conversion))) | |
686 | 'japanese-jisx0208-1978)))) | |
687 | (if (charsetp g0) | |
688 | (if (plist-get props 'force-g0-on-output) | |
689 | (setq g0 `(nil ,g0)) | |
690 | (setq g0 `(,g0 t)))) | |
691 | (if (charsetp g1) | |
692 | (if (plist-get props 'force-g1-on-output) | |
693 | (setq g1 `(nil ,g1)) | |
694 | (setq g1 `(,g1 t)))) | |
695 | (if (charsetp g2) | |
696 | (if (plist-get props 'force-g2-on-output) | |
697 | (setq g2 `(nil ,g2)) | |
698 | (setq g2 `(,g2 t)))) | |
699 | (if (charsetp g3) | |
700 | (if (plist-get props 'force-g3-on-output) | |
701 | (setq g3 `(nil ,g3)) | |
702 | (setq g3 `(,g3 t)))) | |
703 | `(,name 2 ,mnemonic ,doc-string | |
704 | (,g0 ,g1 ,g2 ,g3 | |
705 | ,(plist-get props 'short) | |
706 | ,(not (plist-get props 'no-ascii-eol)) | |
707 | ,(not (plist-get props 'no-ascii-cntl)) | |
708 | ,(plist-get props 'seven) | |
709 | t | |
710 | ,(not (plist-get props 'lock-shift)) | |
711 | ,use-roman | |
712 | ,use-oldjis | |
713 | ,(plist-get props 'no-iso6429) | |
714 | nil nil nil nil) | |
715 | ,properties ,eol-type))) | |
716 | ((eq type 'big5) | |
717 | `(,name 3 ,mnemonic ,doc-string () ,properties ,eol-type)) | |
50c29104 | 718 | ((eq type 'ccl) |
f4a012a6 | 719 | `(,name 4 ,mnemonic ,doc-string |
50c29104 | 720 | (,(plist-get props 'decode) . ,(plist-get props 'encode)) |
f4a012a6 | 721 | ,properties ,eol-type)) |
50c29104 | 722 | (t |
f4a012a6 | 723 | (error "unsupported XEmacs style make-coding-style arguments: %S" |
50c29104 KH |
724 | `(,name ,type ,doc-string ,props)))))) |
725 | ||
8057896b | 726 | (defun make-coding-system (coding-system type mnemonic doc-string |
1b46a680 KH |
727 | &optional |
728 | flags | |
729 | properties | |
730 | eol-type) | |
3bb1accb | 731 | "Define a new coding system CODING-SYSTEM (symbol). |
521d4010 | 732 | Remaining arguments are TYPE, MNEMONIC, DOC-STRING, FLAGS (optional), |
d9e3229d | 733 | and PROPERTIES (optional) which construct a coding-spec of CODING-SYSTEM |
a6acd8a2 | 734 | in the following format: |
0269ddfb | 735 | [TYPE MNEMONIC DOC-STRING PLIST FLAGS] |
3bb1accb KH |
736 | |
737 | TYPE is an integer value indicating the type of the coding system as follows: | |
4ed46869 | 738 | 0: Emacs internal format, |
d9f6dfe6 | 739 | 1: Shift-JIS (or MS-Kanji) used mainly on Japanese PCs, |
4ed46869 | 740 | 2: ISO-2022 including many variants, |
d9f6dfe6 | 741 | 3: Big5 used mainly on Chinese PCs, |
cd9c3177 | 742 | 4: private, CCL programs provide encoding/decoding algorithm, |
521d4010 | 743 | 5: Raw-text, which means that text contains random 8-bit codes. |
0269ddfb | 744 | |
3bb1accb | 745 | MNEMONIC is a character to be displayed on mode line for the coding system. |
0269ddfb | 746 | |
3bb1accb | 747 | DOC-STRING is a documentation string for the coding system. |
0269ddfb | 748 | |
3bb1accb | 749 | FLAGS specifies more detailed information of the coding system as follows: |
d6d6d592 | 750 | |
3bb1accb | 751 | If TYPE is 2 (ISO-2022), FLAGS is a list of these elements: |
4ed46869 KH |
752 | CHARSET0, CHARSET1, CHARSET2, CHARSET3, SHORT-FORM, |
753 | ASCII-EOL, ASCII-CNTL, SEVEN, LOCKING-SHIFT, SINGLE-SHIFT, | |
d6d6d592 | 754 | USE-ROMAN, USE-OLDJIS, NO-ISO6429, INIT-BOL, DESIGNATION-BOL, |
850101ed | 755 | SAFE, ACCEPT-LATIN-EXTRA-CODE. |
4ed46869 KH |
756 | CHARSETn are character sets initially designated to Gn graphic registers. |
757 | If CHARSETn is nil, Gn is never used. | |
758 | If CHARSETn is t, Gn can be used but nothing designated initially. | |
759 | If CHARSETn is a list of character sets, those character sets are | |
760 | designated to Gn on output, but nothing designated to Gn initially. | |
421e3b4e | 761 | But, character set `ascii' can be designated only to G0. |
4ed46869 KH |
762 | SHORT-FORM non-nil means use short designation sequence on output. |
763 | ASCII-EOL non-nil means designate ASCII to g0 at end of line on output. | |
764 | ASCII-CNTL non-nil means designate ASCII to g0 before control codes and | |
765 | SPACE on output. | |
766 | SEVEN non-nil means use 7-bit code only on output. | |
767 | LOCKING-SHIFT non-nil means use locking-shift. | |
768 | SINGLE-SHIFT non-nil means use single-shift. | |
769 | USE-ROMAN non-nil means designate JIS0201-1976-Roman instead of ASCII. | |
770 | USE-OLDJIS non-nil means designate JIS0208-1976 instead of JIS0208-1983. | |
771 | NO-ISO6429 non-nil means not use ISO6429's direction specification. | |
69eba008 KH |
772 | INIT-BOL non-nil means any designation state is assumed to be reset |
773 | to initial at each beginning of line on output. | |
774 | DESIGNATION-BOL non-nil means designation sequences should be placed | |
775 | at beginning of line on output. | |
a6acd8a2 | 776 | SAFE non-nil means convert unsafe characters to `?' on output. |
c11a8f77 KH |
777 | Characters not specified in the property `safe-charsets' nor |
778 | `safe-chars' are unsafe. | |
850101ed RS |
779 | ACCEPT-LATIN-EXTRA-CODE non-nil means code-detection routine accepts |
780 | a code specified in `latin-extra-code-table' (which see) as a valid | |
781 | code of the coding system. | |
d6d6d592 | 782 | |
3bb1accb KH |
783 | If TYPE is 4 (private), FLAGS should be a cons of CCL programs, for |
784 | decoding and encoding. CCL programs should be specified by their | |
785 | symbols. | |
d9e3229d | 786 | |
a2852e45 KH |
787 | PROPERTIES is an alist of properties vs the corresponding values. The |
788 | following properties are recognized: | |
789 | ||
790 | o post-read-conversion | |
0bca779a | 791 | |
a2852e45 KH |
792 | The value is a function to call after some text is inserted and |
793 | decoded by the coding system itself and before any functions in | |
521d4010 DL |
794 | `after-insert-functions' are called. The argument of this |
795 | function is the same as for a function in | |
796 | `after-insert-file-functions', i.e. LENGTH of the text inserted, | |
797 | with point at the head of the text to be decoded. | |
0bca779a | 798 | |
a2852e45 | 799 | o pre-write-conversion |
0bca779a | 800 | |
a2852e45 KH |
801 | The value is a function to call after all functions in |
802 | `write-region-annotate-functions' and `buffer-file-format' are | |
803 | called, and before the text is encoded by the coding system itself. | |
a284eea3 DL |
804 | The arguments to this function are the same as those of a function |
805 | in `write-region-annotate-functions', i.e. FROM and TO, specifying | |
521d4010 | 806 | a region of text. |
0bca779a | 807 | |
a2852e45 | 808 | o translation-table-for-decode |
0bca779a | 809 | |
a2852e45 KH |
810 | The value is a translation table to be applied on decoding. See |
811 | the function `make-translation-table' for the format of translation | |
521d4010 | 812 | table. This is not applicable to type 4 (CCL-based) coding systems. |
0bca779a | 813 | |
a2852e45 | 814 | o translation-table-for-encode |
0bca779a | 815 | |
521d4010 DL |
816 | The value is a translation table to be applied on encoding. This is |
817 | not applicable to type 4 (CCL-based) coding systems. | |
0bca779a | 818 | |
a2852e45 | 819 | o safe-chars |
0bca779a | 820 | |
a2852e45 KH |
821 | The value is a char table. If a character has non-nil value in it, |
822 | the character is safely supported by the coding system. This | |
823 | overrides the specification of safe-charsets. | |
824 | ||
825 | o safe-charsets | |
521d4010 | 826 | |
a2852e45 KH |
827 | The value is a list of charsets safely supported by the coding |
828 | system. The value t means that all charsets Emacs handles are | |
829 | supported. Even if some charset is not in this list, it doesn't | |
521d4010 DL |
830 | mean that the charset can't be encoded in the coding system; |
831 | it just means that some other receiver of text encoded | |
a2852e45 | 832 | in the coding system won't be able to handle that charset. |
0bca779a | 833 | |
a2852e45 | 834 | o mime-charset |
0bca779a | 835 | |
d9f6dfe6 | 836 | The value is a symbol whose name is the `MIME-charset' parameter of |
a2852e45 | 837 | the coding system. |
0bca779a | 838 | |
a2852e45 | 839 | o valid-codes (meaningful only for a coding system based on CCL) |
0bca779a | 840 | |
a2852e45 KH |
841 | The value is a list to indicate valid byte ranges of the encoded |
842 | file. Each element of the list is an integer or a cons of integer. | |
843 | In the former case, the integer value is a valid byte code. In the | |
6fb7766c | 844 | latter case, the integers specify the range of valid byte codes. |
a2852e45 | 845 | |
d9e3229d KH |
846 | These properties are set in PLIST, a property list. This function |
847 | also sets properties `coding-category' and `alias-coding-systems' | |
848 | automatically. | |
4ed46869 | 849 | |
1b46a680 KH |
850 | EOL-TYPE specifies the EOL type of the coding-system in one of the |
851 | following formats: | |
852 | ||
853 | o symbol (unix, dos, or mac) | |
854 | ||
855 | The symbol `unix' means Unix-like EOL (LF), `dos' means | |
856 | DOS-like EOL (CRLF), and `mac' means MAC-like EOL (CR). | |
857 | ||
858 | o number (0, 1, or 2) | |
859 | ||
860 | The number 0, 1, and 2 mean UNIX, DOS, and MAC-like EOL | |
861 | respectively. | |
862 | ||
863 | o vector of coding-systems of length 3 | |
864 | ||
865 | The EOL type is detected automatically for the coding system. | |
866 | And, according to the detected EOL type, one of the coding | |
867 | systems in the vector is selected. Elements of the vector | |
4d3a9228 | 868 | corresponds to Unix-like EOL, DOS-like EOL, and Mac-like EOL |
1b46a680 KH |
869 | in this order. |
870 | ||
3bb1accb KH |
871 | Kludgy features for backward compatibility: |
872 | ||
873 | 1. If TYPE is 4 and car or cdr of FLAGS is a vector, the vector is | |
874 | treated as a compiled CCL code. | |
875 | ||
876 | 2. If PROPERTIES is just a list of character sets, the list is set as | |
877 | a value of `safe-charsets' in PLIST." | |
50c29104 KH |
878 | |
879 | ;; For compatiblity with XEmacs, we check the type of TYPE. If it | |
c3d0ee51 EZ |
880 | ;; is a symbol, perhaps, this function is called with XEmacs-style |
881 | ;; arguments. Here, try to transform that kind of arguments to | |
50c29104 KH |
882 | ;; Emacs style. |
883 | (if (symbolp type) | |
884 | (let ((args (transform-make-coding-system-args coding-system type | |
885 | mnemonic doc-string))) | |
886 | (setq coding-system (car args) | |
1053cc93 | 887 | type (nth 1 args) |
50c29104 KH |
888 | mnemonic (nth 2 args) |
889 | doc-string (nth 3 args) | |
890 | flags (nth 4 args) | |
891 | properties (nth 5 args) | |
892 | eol-type (nth 6 args)))) | |
893 | ||
0269ddfb | 894 | ;; Set a value of `coding-system' property. |
6e9722b0 | 895 | (let ((coding-spec (make-vector 5 nil)) |
d9e3229d KH |
896 | (no-initial-designation t) |
897 | (no-alternative-designation t) | |
e6cddb1a | 898 | (accept-latin-extra-code nil) |
6e9722b0 | 899 | coding-category) |
cd9c3177 | 900 | (if (or (not (integerp type)) (< type 0) (> type 5)) |
0269ddfb | 901 | (error "TYPE argument must be 0..5")) |
8057896b | 902 | (if (or (not (integerp mnemonic)) (<= mnemonic ? ) (> mnemonic 127)) |
521d4010 | 903 | (error "MNEMONIC argument must be an ASCII printable character")) |
0269ddfb KH |
904 | (aset coding-spec coding-spec-type-idx type) |
905 | (aset coding-spec coding-spec-mnemonic-idx mnemonic) | |
906 | (aset coding-spec coding-spec-doc-string-idx | |
e127a722 | 907 | (purecopy (if (stringp doc-string) doc-string ""))) |
6e9722b0 KH |
908 | (cond ((= type 0) |
909 | (setq coding-category 'coding-category-emacs-mule)) | |
910 | ((= type 1) | |
911 | (setq coding-category 'coding-category-sjis)) | |
912 | ((= type 2) ; ISO2022 | |
4ed46869 | 913 | (let ((i 0) |
6e9722b0 | 914 | (vec (make-vector 32 nil)) |
05134257 KH |
915 | (g1-designation nil) |
916 | (fl flags)) | |
4ed46869 | 917 | (while (< i 4) |
05134257 | 918 | (let ((charset (car fl))) |
6e9722b0 KH |
919 | (if (and no-initial-designation |
920 | (> i 0) | |
921 | (or (charsetp charset) | |
922 | (and (consp charset) | |
923 | (charsetp (car charset))))) | |
924 | (setq no-initial-designation nil)) | |
925 | (if (charsetp charset) | |
926 | (if (= i 1) (setq g1-designation charset)) | |
927 | (if (consp charset) | |
928 | (let ((tail charset) | |
929 | elt) | |
930 | (while tail | |
931 | (setq elt (car tail)) | |
d9e3229d KH |
932 | (if (eq elt t) |
933 | (setq no-alternative-designation nil) | |
934 | (if (and elt (not (charsetp elt))) | |
935 | (error "Invalid charset: %s" elt))) | |
6e9722b0 KH |
936 | (setq tail (cdr tail))) |
937 | (setq g1-designation (car charset))) | |
d9e3229d KH |
938 | (if charset |
939 | (if (eq charset t) | |
940 | (setq no-alternative-designation nil) | |
941 | (error "Invalid charset: %s" charset))))) | |
4ed46869 | 942 | (aset vec i charset)) |
05134257 KH |
943 | (setq fl (cdr fl) i (1+ i))) |
944 | (while (and (< i 32) fl) | |
945 | (aset vec i (car fl)) | |
e6cddb1a KH |
946 | (if (and (= i 16) ; ACCEPT-LATIN-EXTRA-CODE |
947 | (car fl)) | |
948 | (setq accept-latin-extra-code t)) | |
05134257 | 949 | (setq fl (cdr fl) i (1+ i))) |
6e9722b0 | 950 | (aset coding-spec 4 vec) |
6e9722b0 KH |
951 | (setq coding-category |
952 | (if (aref vec 8) ; Use locking-shift. | |
dc64cd19 KH |
953 | (or (and (aref vec 7) 'coding-category-iso-7-else) |
954 | 'coding-category-iso-8-else) | |
6e9722b0 KH |
955 | (if (aref vec 7) ; 7-bit only. |
956 | (if (aref vec 9) ; Use single-shift. | |
dc64cd19 | 957 | 'coding-category-iso-7-else |
d9e3229d KH |
958 | (if no-alternative-designation |
959 | 'coding-category-iso-7-tight | |
960 | 'coding-category-iso-7)) | |
3ad911d8 KH |
961 | (if (or no-initial-designation |
962 | (not no-alternative-designation)) | |
dc64cd19 | 963 | 'coding-category-iso-8-else |
6e9722b0 KH |
964 | (if (and (charsetp g1-designation) |
965 | (= (charset-dimension g1-designation) 2)) | |
966 | 'coding-category-iso-8-2 | |
967 | 'coding-category-iso-8-1))))))) | |
968 | ((= type 3) | |
969 | (setq coding-category 'coding-category-big5)) | |
970 | ((= type 4) ; private | |
c76b5c99 | 971 | (setq coding-category 'coding-category-ccl) |
3bb1accb KH |
972 | (if (not (consp flags)) |
973 | (error "Invalid FLAGS argument for TYPE 4 (CCL)") | |
974 | (let ((decoder (check-ccl-program | |
975 | (car flags) | |
976 | (intern (format "%s-decoder" coding-system)))) | |
977 | (encoder (check-ccl-program | |
978 | (cdr flags) | |
979 | (intern (format "%s-encoder" coding-system))))) | |
980 | (if (and decoder encoder) | |
981 | (aset coding-spec 4 (cons decoder encoder)) | |
982 | (error "Invalid FLAGS argument for TYPE 4 (CCL)"))))) | |
cd9c3177 KH |
983 | (t ; i.e. (= type 5) |
984 | (setq coding-category 'coding-category-raw-text))) | |
0269ddfb KH |
985 | |
986 | (let ((plist (list 'coding-category coding-category | |
d9e3229d | 987 | 'alias-coding-systems (list coding-system)))) |
0269ddfb | 988 | (if no-initial-designation |
d9e3229d KH |
989 | (plist-put plist 'no-initial-designation t)) |
990 | (if (and properties | |
991 | (or (eq properties t) | |
992 | (not (consp (car properties))))) | |
993 | ;; In the old version, the arg PROPERTIES is a list to be | |
994 | ;; set in PLIST as a value of property `safe-charsets'. | |
c11a8f77 KH |
995 | (setq properties (list (cons 'safe-charsets properties)))) |
996 | ;; In the current version PROPERTIES is a property list. | |
997 | ;; Reflect it into PLIST one by one while handling safe-chars | |
998 | ;; specially. | |
999 | (let ((safe-charsets (cdr (assq 'safe-charsets properties))) | |
1000 | (safe-chars (cdr (assq 'safe-chars properties))) | |
1001 | (l properties) | |
1002 | prop val) | |
1003 | ;; If only safe-charsets is specified, make a char-table from | |
1004 | ;; it, and store that char-table as the value of `safe-chars'. | |
1005 | (if (and (not safe-chars) safe-charsets) | |
1006 | (let (charset) | |
1007 | (if (eq safe-charsets t) | |
1008 | (setq safe-chars t) | |
1009 | (setq safe-chars (make-char-table 'safe-chars)) | |
1010 | (while safe-charsets | |
1011 | (setq charset (car safe-charsets) | |
1012 | safe-charsets (cdr safe-charsets)) | |
1013 | (cond ((eq charset 'ascii)) ; just ignore | |
1014 | ((eq charset 'eight-bit-control) | |
1015 | (let ((i 128)) | |
1016 | (while (< i 160) | |
1017 | (aset safe-chars i t) | |
1018 | (setq i (1+ i))))) | |
1019 | ((eq charset 'eight-bit-graphic) | |
1020 | (let ((i 160)) | |
1021 | (while (< i 256) | |
1022 | (aset safe-chars i t) | |
1023 | (setq i (1+ i))))) | |
1024 | (t | |
e6cddb1a KH |
1025 | (aset safe-chars (make-char charset) t)))) |
1026 | (if accept-latin-extra-code | |
1027 | (let ((i 128)) | |
1028 | (while (< i 160) | |
1029 | (if (aref latin-extra-code-table i) | |
1030 | (aset safe-chars i t)) | |
1031 | (setq i (1+ i)))))) | |
c11a8f77 KH |
1032 | (setq l (cons (cons 'safe-chars safe-chars) l)))) |
1033 | (while l | |
1034 | (setq prop (car (car l)) val (cdr (car l)) l (cdr l)) | |
1035 | (if (eq prop 'safe-chars) | |
1036 | (progn | |
cc57cc54 KH |
1037 | (if (and (symbolp val) |
1038 | (get val 'translation-table)) | |
1039 | (setq safe-chars (get val 'translation-table))) | |
1040 | (register-char-codings coding-system safe-chars) | |
1041 | (setq val safe-chars))) | |
c11a8f77 | 1042 | (plist-put plist prop val))) |
1b46a680 KH |
1043 | ;; The property `coding-category' may have been set differently |
1044 | ;; through PROPERTIES. | |
1045 | (setq coding-category (plist-get plist 'coding-category)) | |
0269ddfb | 1046 | (aset coding-spec coding-spec-plist-idx plist)) |
6e9722b0 | 1047 | (put coding-system 'coding-system coding-spec) |
6e9722b0 KH |
1048 | (put coding-category 'coding-systems |
1049 | (cons coding-system (get coding-category 'coding-systems)))) | |
4ed46869 | 1050 | |
1b46a680 | 1051 | ;; Next, set a value of `eol-type' property. |
af873cd1 | 1052 | (if (not eol-type) |
1b46a680 KH |
1053 | ;; If EOL-TYPE is nil, set a vector of subsidiary coding |
1054 | ;; systems, each corresponds to a coding system for the detected | |
1055 | ;; EOL format. | |
1056 | (setq eol-type (make-subsidiary-coding-system coding-system))) | |
1057 | (setq eol-type | |
1058 | (cond ((or (eq eol-type 'unix) (null eol-type)) | |
1059 | 0) | |
1060 | ((eq eol-type 'dos) | |
1061 | 1) | |
1062 | ((eq eol-type 'mac) | |
1063 | 2) | |
1064 | ((or (and (vectorp eol-type) | |
1065 | (= (length eol-type) 3)) | |
1066 | (and (numberp eol-type) | |
1067 | (and (>= eol-type 0) | |
1068 | (<= eol-type 2)))) | |
1069 | eol-type) | |
1070 | (t | |
1071 | (error "Invalid EOL-TYPE spec:%S" eol-type)))) | |
1072 | (put coding-system 'eol-type eol-type) | |
0269ddfb KH |
1073 | |
1074 | ;; At last, register CODING-SYSTEM in `coding-system-list' and | |
1075 | ;; `coding-system-alist'. | |
857ea15c | 1076 | (add-to-coding-system-list coding-system) |
0269ddfb | 1077 | (setq coding-system-alist (cons (list (symbol-name coding-system)) |
d9e3229d | 1078 | coding-system-alist)) |
05134257 KH |
1079 | |
1080 | ;; For a coding system of cateogory iso-8-1 and iso-8-2, create | |
1081 | ;; XXX-with-esc variants. | |
1082 | (let ((coding-category (coding-system-category coding-system))) | |
1083 | (if (or (eq coding-category 'coding-category-iso-8-1) | |
1084 | (eq coding-category 'coding-category-iso-8-2)) | |
1085 | (let ((esc (intern (concat (symbol-name coding-system) "-with-esc"))) | |
c11a8f77 KH |
1086 | (doc (format "Same as %s but can handle any charsets by ISO's escape sequences." coding-system)) |
1087 | (safe-charsets (assq 'safe-charsets properties)) | |
1088 | (mime-charset (assq 'mime-charset properties))) | |
1089 | (if safe-charsets | |
1090 | (setcdr safe-charsets t) | |
1091 | (setq properties (cons (cons 'safe-charsets t) properties))) | |
1092 | (if mime-charset | |
1093 | (setcdr mime-charset nil)) | |
05134257 KH |
1094 | (make-coding-system esc type mnemonic doc |
1095 | (if (listp (car flags)) | |
1096 | (cons (append (car flags) '(t)) (cdr flags)) | |
1097 | (cons (list (car flags) t) (cdr flags))) | |
c11a8f77 | 1098 | properties)))) |
05134257 | 1099 | |
d9e3229d | 1100 | coding-system) |
8057896b | 1101 | |
5470d5b6 KH |
1102 | (put 'safe-chars 'char-table-extra-slots 0) |
1103 | ||
a42763dc | 1104 | (defun define-coding-system-alias (alias coding-system) |
358d28fb | 1105 | "Define ALIAS as an alias for coding system CODING-SYSTEM." |
0269ddfb | 1106 | (put alias 'coding-system (coding-system-spec coding-system)) |
857ea15c | 1107 | (add-to-coding-system-list alias) |
0269ddfb KH |
1108 | (setq coding-system-alist (cons (list (symbol-name alias)) |
1109 | coding-system-alist)) | |
1110 | (let ((eol-type (coding-system-eol-type coding-system))) | |
1111 | (if (vectorp eol-type) | |
708fc465 KH |
1112 | (progn |
1113 | (nconc (coding-system-get alias 'alias-coding-systems) (list alias)) | |
1114 | (put alias 'eol-type (make-subsidiary-coding-system alias))) | |
0269ddfb | 1115 | (put alias 'eol-type eol-type)))) |
4ed46869 | 1116 | |
bbdea948 RS |
1117 | (defun merge-coding-systems (first second) |
1118 | "Fill in any unspecified aspects of coding system FIRST from SECOND. | |
1119 | Return the resulting coding system." | |
1120 | (let ((base (coding-system-base second)) | |
1121 | (eol (coding-system-eol-type second))) | |
1122 | ;; If FIRST doesn't specify text conversion, merge with that of SECOND. | |
1123 | (if (eq (coding-system-base first) 'undecided) | |
1124 | (setq first (coding-system-change-text-conversion first base))) | |
1125 | ;; If FIRST doesn't specify eol conversion, merge with that of SECOND. | |
1126 | (if (and (vectorp (coding-system-eol-type first)) | |
1127 | (numberp eol) (>= eol 0) (<= eol 2)) | |
1128 | (setq first (coding-system-change-eol-conversion | |
1129 | first eol))) | |
1130 | first)) | |
1131 | ||
4ed46869 | 1132 | (defun set-buffer-file-coding-system (coding-system &optional force) |
358d28fb RS |
1133 | "Set the file coding-system of the current buffer to CODING-SYSTEM. |
1134 | This means that when you save the buffer, it will be converted | |
1135 | according to CODING-SYSTEM. For a list of possible values of CODING-SYSTEM, | |
1136 | use \\[list-coding-systems]. | |
1137 | ||
bbdea948 RS |
1138 | If CODING-SYSTEM leaves the text conversion unspecified, or if it |
1139 | leaves the end-of-line conversion unspecified, FORCE controls what to | |
1140 | do. If FORCE is nil, get the unspecified aspect (or aspects) from the | |
1141 | buffer's previous `buffer-file-coding-system' value (if it is | |
1142 | specified there). Otherwise, levae it unspecified. | |
aeef8f07 KH |
1143 | |
1144 | This marks the buffer modified so that the succeeding \\[save-buffer] | |
1145 | surely saves the buffer with CODING-SYSTEM. From a program, if you | |
1146 | don't want to mark the buffer modified, just set the variable | |
1147 | `buffer-file-coding-system' directly." | |
bbdea948 | 1148 | (interactive "zCoding system for saving file (default, nil): \nP") |
4ed46869 | 1149 | (check-coding-system coding-system) |
36d455c4 | 1150 | (if (and coding-system buffer-file-coding-system (null force)) |
bbdea948 RS |
1151 | (setq coding-system |
1152 | (merge-coding-systems coding-system buffer-file-coding-system))) | |
4ed46869 KH |
1153 | (setq buffer-file-coding-system coding-system) |
1154 | (set-buffer-modified-p t) | |
1155 | (force-mode-line-update)) | |
1156 | ||
bbdea948 RS |
1157 | (defun revert-buffer-with-coding-system (coding-system &optional force) |
1158 | "Visit the current buffer's file again using coding system CODING-SYSTEM. | |
1159 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. | |
1160 | ||
1161 | If CODING-SYSTEM leaves the text conversion unspecified, or if it | |
1162 | leaves the end-of-line conversion unspecified, FORCE controls what to | |
1163 | do. If FORCE is nil, get the unspecified aspect (or aspects) from the | |
1164 | buffer's previous `buffer-file-coding-system' value (if it is | |
1165 | specified there). Otherwise, determine it from the file contents as | |
1166 | usual for visiting a file." | |
1167 | (interactive "zCoding system for visited file (default, nil): \nP") | |
1168 | (check-coding-system coding-system) | |
1169 | (if (and coding-system buffer-file-coding-system (null force)) | |
1170 | (setq coding-system | |
1171 | (merge-coding-systems coding-system buffer-file-coding-system))) | |
1172 | (let ((coding-system-for-read coding-system)) | |
1173 | (revert-buffer))) | |
1174 | ||
358d28fb RS |
1175 | (defvar default-terminal-coding-system nil |
1176 | "Default value for the terminal coding system. | |
1177 | This is normally set according to the selected language environment. | |
1178 | See also the command `set-terminal-coding-system'.") | |
1179 | ||
df100398 KH |
1180 | (defun set-terminal-coding-system (coding-system) |
1181 | "Set coding system of your terminal to CODING-SYSTEM. | |
358d28fb RS |
1182 | All text output to the terminal will be encoded |
1183 | with the specified coding system. | |
1184 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. | |
1185 | The default is determined by the selected language environment | |
1186 | or by the previous use of this command." | |
1187 | (interactive | |
2e02a76f RS |
1188 | (list (let ((default (if (and (not (terminal-coding-system)) |
1189 | default-terminal-coding-system) | |
1190 | default-terminal-coding-system))) | |
1191 | (read-coding-system | |
1192 | (format "Coding system for terminal display (default, %s): " | |
1193 | default) | |
1194 | default)))) | |
358d28fb RS |
1195 | (if (and (not coding-system) |
1196 | (not (terminal-coding-system))) | |
1197 | (setq coding-system default-terminal-coding-system)) | |
1198 | (if coding-system | |
521d4010 | 1199 | (setq default-terminal-coding-system coding-system)) |
df100398 KH |
1200 | (set-terminal-coding-system-internal coding-system) |
1201 | (redraw-frame (selected-frame))) | |
1202 | ||
358d28fb RS |
1203 | (defvar default-keyboard-coding-system nil |
1204 | "Default value of the keyboard coding system. | |
1205 | This is normally set according to the selected language environment. | |
1206 | See also the command `set-keyboard-coding-system'.") | |
1207 | ||
df100398 | 1208 | (defun set-keyboard-coding-system (coding-system) |
358d28fb RS |
1209 | "Set coding system for keyboard input to CODING-SYSTEM. |
1210 | In addition, this command enables Encoded-kbd minor mode. | |
6d34f495 DL |
1211 | \(If CODING-SYSTEM is nil, Encoded-kbd mode is turned off -- see |
1212 | `encoded-kbd-mode'.) | |
358d28fb RS |
1213 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. |
1214 | The default is determined by the selected language environment | |
1215 | or by the previous use of this command." | |
1216 | (interactive | |
2e02a76f RS |
1217 | (list (let ((default (if (and (not (keyboard-coding-system)) |
1218 | default-keyboard-coding-system) | |
1219 | default-keyboard-coding-system))) | |
1220 | (read-coding-system | |
1221 | (format "Coding system for keyboard input (default, %s): " | |
1222 | default) | |
1223 | default)))) | |
358d28fb RS |
1224 | (if (and (not coding-system) |
1225 | (not (keyboard-coding-system))) | |
1226 | (setq coding-system default-keyboard-coding-system)) | |
1227 | (if coding-system | |
1228 | (setq default-keyboard-coding-system coding-system)) | |
df100398 | 1229 | (set-keyboard-coding-system-internal coding-system) |
b23bad0b | 1230 | (setq keyboard-coding-system coding-system) |
df100398 KH |
1231 | (encoded-kbd-mode (if coding-system 1 0))) |
1232 | ||
6d34f495 DL |
1233 | (defcustom keyboard-coding-system nil |
1234 | "Specify coding system for keyboard input. | |
1235 | If you set this on a terminal which can't distinguish Meta keys from | |
1236 | 8-bit characters, you will have to use ESC to type Meta characters. | |
1237 | See Info node `Specify Coding' and Info node `Single-Byte Character Support'. | |
1238 | ||
1239 | Setting this variable directly does not take effect; | |
1240 | use either M-x customize or \\[set-keyboard-coding-system]." | |
1241 | :type '(coding-system :tag "Coding system") | |
1242 | :link '(info-link "(emacs)Specify Coding") | |
1243 | :link '(info-link "(emacs)Single-Byte Character Support") | |
1244 | :set (lambda (symbol value) | |
1245 | ;; Don't load encoded-kbd-mode unnecessarily. | |
1246 | (if (or value (boundp 'encoded-kbd-mode)) | |
1247 | (set-keyboard-coding-system value) | |
1248 | (set-default 'keyboard-coding-system nil))) ; must initialize | |
1249 | :version "21.1" | |
1250 | :group 'keyboard | |
1251 | :group 'mule) | |
1252 | ||
df100398 | 1253 | (defun set-buffer-process-coding-system (decoding encoding) |
358d28fb | 1254 | "Set coding systems for the process associated with the current buffer. |
df100398 | 1255 | DECODING is the coding system to be used to decode input from the process, |
358d28fb RS |
1256 | ENCODING is the coding system to be used to encode output to the process. |
1257 | ||
1258 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]." | |
4ed46869 | 1259 | (interactive |
83911021 | 1260 | "zCoding-system for output from the process: \nzCoding-system for input to the process: ") |
4ed46869 KH |
1261 | (let ((proc (get-buffer-process (current-buffer)))) |
1262 | (if (null proc) | |
521d4010 | 1263 | (error "No process") |
df100398 KH |
1264 | (check-coding-system decoding) |
1265 | (check-coding-system encoding) | |
1266 | (set-process-coding-system proc decoding encoding))) | |
4ed46869 KH |
1267 | (force-mode-line-update)) |
1268 | ||
d0b99881 RS |
1269 | (defalias 'set-clipboard-coding-system 'set-selection-coding-system) |
1270 | ||
14915c37 | 1271 | (defun set-selection-coding-system (coding-system) |
b25eef20 KH |
1272 | "Make CODING-SYSTEM used for communicating with other X clients . |
1273 | When sending or receiving text via cut_buffer, selection, and clipboard, | |
1274 | the text is encoded or decoded by CODING-SYSTEM." | |
a03b3ce1 | 1275 | (interactive "zCoding system for X selection: ") |
b25eef20 | 1276 | (check-coding-system coding-system) |
14915c37 | 1277 | (setq selection-coding-system coding-system)) |
b25eef20 | 1278 | |
e8dd0160 | 1279 | ;; Coding system lastly specified by the command |
a03b3ce1 KH |
1280 | ;; set-next-selection-coding-system. |
1281 | (defvar last-next-selection-coding-system nil) | |
1282 | ||
1283 | (defun set-next-selection-coding-system (coding-system) | |
1284 | "Make CODING-SYSTEM used for the next communication with other X clients. | |
1285 | This setting is effective for the next communication only." | |
1286 | (interactive | |
1287 | (list (read-coding-system | |
1288 | (if last-next-selection-coding-system | |
1289 | (format "Coding system for the next X selection (default, %S): " | |
1290 | last-next-selection-coding-system) | |
1291 | "Coding system for the next X selection: ") | |
1292 | last-next-selection-coding-system))) | |
1293 | (if coding-system | |
1294 | (setq last-next-selection-coding-system coding-system) | |
1295 | (setq coding-system last-next-selection-coding-system)) | |
1296 | (check-coding-system coding-system) | |
1297 | ||
1298 | (setq next-selection-coding-system coding-system)) | |
1299 | ||
4ed46869 | 1300 | (defun set-coding-priority (arg) |
521d4010 DL |
1301 | "Set priority of coding categories according to ARG. |
1302 | ARG is a list of coding categories ordered by priority." | |
d9e3229d KH |
1303 | (let ((l arg) |
1304 | (current-list (copy-sequence coding-category-list))) | |
e8dd0160 | 1305 | ;; Check the validity of ARG while deleting coding categories in |
d9e3229d KH |
1306 | ;; ARG from CURRENT-LIST. We assume that CODING-CATEGORY-LIST |
1307 | ;; contains all coding categories. | |
1308 | (while l | |
1309 | (if (or (null (get (car l) 'coding-category-index)) | |
1310 | (null (memq (car l) current-list))) | |
1311 | (error "Invalid or duplicated element in argument: %s" arg)) | |
1312 | (setq current-list (delq (car l) current-list)) | |
1313 | (setq l (cdr l))) | |
4ed46869 | 1314 | ;; Update `coding-category-list' and return it. |
2feaf204 KH |
1315 | (setq coding-category-list (append arg current-list)) |
1316 | (set-coding-priority-internal))) | |
4ed46869 | 1317 | |
835cbadb EZ |
1318 | ;;; X selections |
1319 | ||
1320 | (defvar non-standard-icccm-encodings-alist | |
1321 | '(("ISO8859-15" . latin-iso8859-15) | |
1322 | ("ISO8859-14" . latin-iso8859-14) | |
1323 | ("KOI8-R" . koi8-r) | |
1324 | ("BIG5-0" . big5)) | |
7f1cdd46 RS |
1325 | "Alist of font charset names defined by XLFD. |
1326 | The cdr of each element is the corresponding Emacs charset or coding system.") | |
835cbadb EZ |
1327 | |
1328 | ;; Functions to support "Non-Standard Character Set Encodings" defined | |
5c88a01e KH |
1329 | ;; by the COMPOUND-TEXT spec. |
1330 | ;; We support that by converting the leading sequence of the | |
1331 | ;; ``extended segment'' to the corresponding ISO-2022 sequences (if | |
1332 | ;; the leading sequence names an Emacs charset), or decode the segment | |
1333 | ;; (if it names a coding system). Encoding does the reverse. | |
1334 | ;; This function also supports "The UTF-8 encoding" described in the | |
1335 | ;; section 7 of the documentation fo COMPOUND-TEXT distributed with | |
1336 | ;; XFree86. | |
1337 | ||
835cbadb EZ |
1338 | (defun ctext-post-read-conversion (len) |
1339 | "Decode LEN characters encoded as Compound Text with Extended Segments." | |
1340 | (buffer-disable-undo) ; minimize consing due to insertions and deletions | |
1341 | (narrow-to-region (point) (+ (point) len)) | |
1342 | (save-match-data | |
1343 | (let ((pt (point-marker)) | |
1344 | (oldpt (point-marker)) | |
1345 | (newpt (make-marker)) | |
1346 | (modified-p (buffer-modified-p)) | |
1347 | (case-fold-search nil) | |
1348 | last-coding-system-used | |
1349 | encoding textlen chset) | |
1350 | (while (re-search-forward | |
5c88a01e | 1351 | "\\(\e\\)%/[0-4]\\([\200-\377][\200-\377]\\)\\([^\002]+\\)\002\\|\e%G[^\e]+\e%@" |
835cbadb EZ |
1352 | nil 'move) |
1353 | (set-marker newpt (point)) | |
1354 | (set-marker pt (match-beginning 0)) | |
5c88a01e KH |
1355 | (if (= (preceding-char) ?@) |
1356 | ;; We found embedded utf-8 sequence. | |
1357 | (progn | |
1358 | (delete-char -3) ; delete ESC % @ at the tail | |
1359 | (goto-char pt) | |
1360 | (delete-char 3) ; delete ESC % G at the head | |
1361 | (if (> pt oldpt) | |
1362 | (decode-coding-region oldpt pt 'ctext-no-compositions)) | |
1363 | (decode-coding-region pt newpt 'mule-utf-8) | |
1364 | (goto-char newpt) | |
1365 | (set-marker oldpt newpt)) | |
1366 | (setq encoding (match-string 3)) | |
1367 | (setq textlen (- (+ (* (- (aref (match-string 2) 0) 128) 128) | |
1368 | (- (aref (match-string 2) 1) 128)) | |
1369 | (1+ (length encoding)))) | |
1370 | (setq | |
1371 | chset (cdr (assoc-ignore-case encoding | |
1372 | non-standard-icccm-encodings-alist))) | |
1373 | (cond ((null chset) | |
1374 | ;; This charset is not supported--leave this extended | |
1375 | ;; segment unaltered and skip over it. | |
1376 | (goto-char (+ (point) textlen))) | |
1377 | ((charsetp chset) | |
1378 | ;; If it's a charset, replace the leading escape sequence | |
1379 | ;; with a standard ISO-2022 sequence. We will decode all | |
1380 | ;; such segments later, in one go, when we exit the loop | |
1381 | ;; or find an extended segment that names a coding | |
1382 | ;; system, not a charset. | |
1383 | (replace-match | |
1384 | (concat "\\1" | |
1385 | (if (= 0 (charset-iso-graphic-plane chset)) | |
1386 | ;; GL charsets | |
1387 | (if (= 1 (charset-dimension chset)) "(" "$(") | |
1388 | ;; GR charsets | |
1389 | (if (= 96 (charset-chars chset)) | |
1390 | "-" | |
1391 | (if (= 1 (charset-dimension chset)) ")" "$)"))) | |
1392 | (string (charset-iso-final-char chset))) | |
1393 | t) | |
1394 | (goto-char (+ (point) textlen))) | |
1395 | ((coding-system-p chset) | |
1396 | ;; If it's a coding system, we need to decode the segment | |
1397 | ;; right away. But first, decode what we've skipped | |
1398 | ;; across until now. | |
1399 | (when (> pt oldpt) | |
1400 | (decode-coding-region oldpt pt 'ctext-no-compositions)) | |
1401 | (delete-region pt newpt) | |
1402 | (set-marker newpt (+ newpt textlen)) | |
1403 | (decode-coding-region pt newpt chset) | |
1404 | (goto-char newpt) | |
1405 | (set-marker oldpt newpt))))) | |
835cbadb EZ |
1406 | ;; Decode what's left. |
1407 | (when (> (point) oldpt) | |
1408 | (decode-coding-region oldpt (point) 'ctext-no-compositions)) | |
5c88a01e | 1409 | ;; This buffer started as unibyte, because the string we get from |
835cbadb EZ |
1410 | ;; the X selection is a unibyte string. We must now make it |
1411 | ;; multibyte, so that the decoded text is inserted as multibyte | |
1412 | ;; into its buffer. | |
1413 | (set-buffer-multibyte t) | |
1414 | (set-buffer-modified-p modified-p) | |
1415 | (- (point-max) (point-min))))) | |
1416 | ||
c74fe809 EZ |
1417 | ;; If you add charsets here, be sure to modify the regexp used by |
1418 | ;; ctext-pre-write-conversion to look up non-standard charsets. | |
835cbadb EZ |
1419 | (defvar non-standard-designations-alist |
1420 | '(("$(0" . (big5 "big5-0" 2)) | |
1421 | ("$(1" . (big5 "big5-0" 2)) | |
1422 | ("-V" . (t "iso8859-10" 1)) | |
1423 | ("-Y" . (t "iso8859-13" 1)) | |
1424 | ("-_" . (t "iso8859-14" 1)) | |
1425 | ("-b" . (t "iso8859-15" 1)) | |
1426 | ("-f" . (t "iso8859-16" 1))) | |
1427 | "Alist of ctext control sequences that introduce character sets which | |
1428 | are not in the list of approved ICCCM encodings, and the corresponding | |
1429 | coding system, identifier string, and number of octets per encoded | |
1430 | character. | |
1431 | ||
1432 | Each element has the form (CTLSEQ . (ENCODING CHARSET NOCTETS)). CTLSEQ | |
1433 | is the control sequence (sans the leading ESC) that introduces the character | |
1434 | set in the text encoded by compound-text. ENCODING is a coding system | |
1435 | symbol; if it is t, it means that the ctext coding system already encodes | |
1436 | the text correctly, and only the leading control sequence needs to be altered. | |
1437 | If ENCODING is a coding system, we need to re-encode the text with that | |
1438 | coding system. CHARSET is the ICCCM name of the charset we need to put into | |
1439 | the leading control sequence. NOCTETS is the number of octets (bytes) that | |
1440 | encode each character in this charset. NOCTETS can be 0 (meaning the number | |
1441 | of octets per character is variable), 1, 2, 3, or 4.") | |
1442 | ||
1443 | (defun ctext-pre-write-conversion (from to) | |
5dde3c71 EZ |
1444 | "Encode characters between FROM and TO as Compound Text w/Extended Segments. |
1445 | ||
1446 | If FROM is a string, or if the current buffer is not the one set up for us | |
1447 | by run_pre_post_conversion_on_str, generate a new temp buffer, insert the | |
1448 | text, and convert it in the temporary buffer. Otherwise, convert in-place." | |
1449 | (cond ((and (string= (buffer-name) " *code-converting-work*") | |
1450 | (not (stringp from))) | |
1451 | ; Minimize consing due to subsequent insertions and deletions. | |
1452 | (buffer-disable-undo) | |
1453 | (narrow-to-region from to)) | |
1454 | (t | |
1455 | (let ((buf (current-buffer))) | |
1456 | (set-buffer (generate-new-buffer " *temp")) | |
1457 | (buffer-disable-undo) | |
1458 | (if (stringp from) | |
1459 | (insert from) | |
c74fe809 EZ |
1460 | (insert-buffer-substring buf from to)) |
1461 | (setq from (point-min) to (point-max))))) | |
835cbadb EZ |
1462 | (encode-coding-region from to 'ctext-no-compositions) |
1463 | ;; Replace ISO-2022 charset designations with extended segments, for | |
1464 | ;; those charsets that are not part of the official X registry. | |
1465 | (save-match-data | |
1466 | (goto-char (point-min)) | |
1467 | (let ((newpt (make-marker)) | |
1468 | (case-fold-search nil) | |
1469 | pt desig encode-info encoding chset noctets textlen) | |
1470 | (set-buffer-multibyte nil) | |
5dde3c71 EZ |
1471 | ;; The regexp below finds the leading sequences for big5 and |
1472 | ;; iso8859-1[03-6] charsets. | |
835cbadb EZ |
1473 | (while (re-search-forward "\e\\(\$([01]\\|-[VY_bf]\\)" nil 'move) |
1474 | (setq desig (match-string 1) | |
1475 | pt (point-marker) | |
1476 | encode-info (cdr (assoc desig non-standard-designations-alist)) | |
1477 | encoding (car encode-info) | |
1478 | chset (cadr encode-info) | |
1479 | noctets (car (cddr encode-info))) | |
1480 | (skip-chars-forward "^\e") | |
1481 | (set-marker newpt (point)) | |
1482 | (cond | |
1483 | ((eq encoding t) ; only the leading sequence needs to be changed | |
1484 | (setq textlen (+ (- newpt pt) (length chset) 1)) | |
5dde3c71 | 1485 | ;; Generate the ICCCM control sequence for an extended segment. |
835cbadb EZ |
1486 | (replace-match (format "\e%%/%d%c%c%s\ 2" |
1487 | noctets | |
1488 | (+ (/ textlen 128) 128) | |
1489 | (+ (% textlen 128) 128) | |
1490 | chset) | |
1491 | t t)) | |
1492 | ((coding-system-p encoding) ; need to recode the entire segment... | |
1493 | (set-marker pt (match-beginning 0)) | |
1494 | (decode-coding-region pt newpt 'ctext-no-compositions) | |
1495 | (set-buffer-multibyte t) | |
1496 | (encode-coding-region pt newpt encoding) | |
1497 | (set-buffer-multibyte nil) | |
1498 | (setq textlen (+ (- newpt pt) (length chset) 1)) | |
1499 | (goto-char pt) | |
1500 | (insert (format "\e%%/%d%c%c%s\ 2" | |
1501 | noctets | |
1502 | (+ (/ textlen 128) 128) | |
1503 | (+ (% textlen 128) 128) | |
1504 | chset)))) | |
1505 | (goto-char newpt)))) | |
1506 | (set-buffer-multibyte t) | |
5dde3c71 | 1507 | ;; Must return nil, as build_annotations_2 expects that. |
835cbadb EZ |
1508 | nil) |
1509 | ||
4ed46869 KH |
1510 | ;;; FILE I/O |
1511 | ||
e76938e7 | 1512 | (defcustom auto-coding-alist |
0bca779a | 1513 | '(("\\.\\(arc\\|zip\\|lzh\\|zoo\\|jar\\|sx[dmicw]\\|tar\\|tgz\\)\\'" . no-conversion) |
6caef2da | 1514 | ("\\.\\(gz\\|Z\\|bz\\|bz2\\|gpg\\)\\'" . no-conversion) |
45885400 | 1515 | ("/#[^/]+#\\'" . emacs-mule)) |
835f49b8 KH |
1516 | "Alist of filename patterns vs corresponding coding systems. |
1517 | Each element looks like (REGEXP . CODING-SYSTEM). | |
558b0c86 | 1518 | A file whose name matches REGEXP is decoded by CODING-SYSTEM on reading. |
835f49b8 | 1519 | |
7fed493a RS |
1520 | The settings in this alist take priority over `coding:' tags |
1521 | in the file (see the function `set-auto-coding') | |
e76938e7 DL |
1522 | and the contents of `file-coding-system-alist'." |
1523 | :group 'files | |
1524 | :group 'mule | |
1525 | :type '(repeat (cons (regexp :tag "File name regexp") | |
1526 | (symbol :tag "Coding system")))) | |
835f49b8 | 1527 | |
502522b2 GM |
1528 | (defcustom auto-coding-regexp-alist |
1529 | '(("^BABYL OPTIONS:[ \t]*-\\*-[ \t]*rmail[ \t]*-\\*-" . no-conversion)) | |
1530 | "Alist of patterns vs corresponding coding systems. | |
1531 | Each element looks like (REGEXP . CODING-SYSTEM). | |
1532 | A file whose first bytes match REGEXP is decoded by CODING-SYSTEM on reading. | |
1533 | ||
1534 | The settings in this alist take priority over `coding:' tags | |
1535 | in the file (see the function `set-auto-coding') | |
1536 | and the contents of `file-coding-system-alist'." | |
1537 | :group 'files | |
1538 | :group 'mule | |
1539 | :type '(repeat (cons (regexp :tag "Regexp") | |
1540 | (symbol :tag "Coding system")))) | |
1541 | ||
d9f6dfe6 | 1542 | ;; See the bottom of this file for built-in auto coding functions. |
447404a3 CW |
1543 | (defcustom auto-coding-functions '(sgml-xml-auto-coding-function |
1544 | sgml-html-meta-auto-coding-function) | |
d9f6dfe6 CW |
1545 | "A list of functions which attempt to determine a coding system. |
1546 | ||
1547 | Each function in this list should be written to operate on the current | |
1548 | buffer, but should not modify it in any way. It should take one | |
1549 | argument SIZE, past which it should not search. If a function | |
1550 | succeeds in determining a coding system, it should return that coding | |
1551 | system. Otherwise, it should return nil. | |
1552 | ||
447404a3 CW |
1553 | Any `coding:' tags present have a higher priority than the |
1554 | functions in this list." | |
d9f6dfe6 CW |
1555 | :group 'files |
1556 | :group 'mule | |
1557 | :type '(repeat function)) | |
1558 | ||
1c4cc63a KH |
1559 | (defvar set-auto-coding-for-load nil |
1560 | "Non-nil means look for `load-coding' property instead of `coding'. | |
1561 | This is used for loading and byte-compiling Emacs Lisp files.") | |
1562 | ||
8a592131 RS |
1563 | (defun auto-coding-alist-lookup (filename) |
1564 | "Return the coding system specified by `auto-coding-alist' for FILENAME." | |
1565 | (let ((alist auto-coding-alist) | |
ca128d75 | 1566 | (case-fold-search (memq system-type '(vax-vms windows-nt ms-dos))) |
8a592131 RS |
1567 | coding-system) |
1568 | (while (and alist (not coding-system)) | |
1569 | (if (string-match (car (car alist)) filename) | |
1570 | (setq coding-system (cdr (car alist))) | |
1571 | (setq alist (cdr alist)))) | |
1572 | coding-system)) | |
1573 | ||
835f49b8 KH |
1574 | (defun set-auto-coding (filename size) |
1575 | "Return coding system for a file FILENAME of which SIZE bytes follow point. | |
1c4cc63a KH |
1576 | These bytes should include at least the first 1k of the file |
1577 | and the last 3k of the file, but the middle may be omitted. | |
63561304 | 1578 | |
d21ba5e0 DL |
1579 | The function checks FILENAME against the variable `auto-coding-alist'. |
1580 | If FILENAME doesn't match any entries in the variable, it checks the | |
502522b2 | 1581 | contents of the current buffer following point against |
447404a3 | 1582 | `auto-coding-regexp-alist'. If no match is found, it checks for a |
502522b2 | 1583 | `coding:' tag in the first one or two lines following point. If no |
d21ba5e0 | 1584 | `coding:' tag is found, it checks any local variables list in the last |
447404a3 | 1585 | 3K bytes out of the SIZE bytes. Finally, if none of these methods |
d21ba5e0 DL |
1586 | succeed, it checks to see if any function in `auto-coding-functions' |
1587 | gives a match. | |
63561304 | 1588 | |
d21ba5e0 DL |
1589 | The return value is the specified coding system, or nil if nothing is |
1590 | specified. | |
87aba788 | 1591 | |
ba74e833 | 1592 | The variable `set-auto-coding-function' (which see) is set to this |
87aba788 | 1593 | function by default." |
502522b2 | 1594 | (or (auto-coding-alist-lookup filename) |
447404a3 CW |
1595 | ;; Try using `auto-coding-regexp-alist'. |
1596 | (save-excursion | |
1597 | (let ((alist auto-coding-regexp-alist) | |
1598 | coding-system) | |
1599 | (while (and alist (not coding-system)) | |
1600 | (let ((regexp (car (car alist)))) | |
1601 | (when (re-search-forward regexp (+ (point) size) t) | |
1602 | (setq coding-system (cdr (car alist))))) | |
0bca779a | 1603 | (setq alist (cdr alist))) |
447404a3 | 1604 | coding-system)) |
502522b2 GM |
1605 | (let* ((case-fold-search t) |
1606 | (head-start (point)) | |
1607 | (head-end (+ head-start (min size 1024))) | |
1608 | (tail-start (+ head-start (max (- size 3072) 0))) | |
1609 | (tail-end (+ head-start size)) | |
1610 | coding-system head-found tail-found pos) | |
1611 | ;; Try a short cut by searching for the string "coding:" | |
1612 | ;; and for "unibyte:" at the head and tail of SIZE bytes. | |
1613 | (setq head-found (or (search-forward "coding:" head-end t) | |
1614 | (search-forward "unibyte:" head-end t))) | |
1615 | (if (and head-found (> head-found tail-start)) | |
1616 | ;; Head and tail are overlapped. | |
1617 | (setq tail-found head-found) | |
1618 | (goto-char tail-start) | |
1619 | (setq tail-found (or (search-forward "coding:" tail-end t) | |
1620 | (search-forward "unibyte:" tail-end t)))) | |
1621 | ||
1622 | ;; At first check the head. | |
1623 | (when head-found | |
1624 | (goto-char head-start) | |
6b66d028 RS |
1625 | (setq head-end (set-auto-mode-1)) |
1626 | (setq head-start (point)) | |
1d8e9a7c | 1627 | (when (and head-end (< head-found head-end)) |
835f49b8 | 1628 | (goto-char head-start) |
502522b2 GM |
1629 | (when (and set-auto-coding-for-load |
1630 | (re-search-forward | |
6b66d028 | 1631 | "\\(.*;\\)?[ \t]*unibyte:[ \t]*\\([^ ;]+\\)" |
502522b2 GM |
1632 | head-end t)) |
1633 | (setq coding-system 'raw-text)) | |
1634 | (when (and (not coding-system) | |
1635 | (re-search-forward | |
6b66d028 | 1636 | "\\(.*;\\)?[ \t]*coding:[ \t]*\\([^ ;]+\\)" |
502522b2 GM |
1637 | head-end t)) |
1638 | (setq coding-system (intern (match-string 2))) | |
1639 | (or (coding-system-p coding-system) | |
1640 | (setq coding-system nil))))) | |
1641 | ||
1642 | ;; If no coding: tag in the head, check the tail. | |
1643 | (when (and tail-found (not coding-system)) | |
1644 | (goto-char tail-start) | |
1645 | (search-forward "\n\^L" nil t) | |
1646 | (if (re-search-forward | |
1647 | "^\\(.*\\)[ \t]*Local Variables:[ \t]*\\(.*\\)$" tail-end t) | |
1648 | ;; The prefix is what comes before "local variables:" in its | |
1649 | ;; line. The suffix is what comes after "local variables:" | |
1650 | ;; in its line. | |
1651 | (let* ((prefix (regexp-quote (match-string 1))) | |
1652 | (suffix (regexp-quote (match-string 2))) | |
1653 | (re-coding | |
1654 | (concat | |
1655 | "^" prefix | |
cfe98f50 GM |
1656 | ;; N.B. without the \n below, the regexp can |
1657 | ;; eat newlines. | |
1658 | "[ \t]*coding[ \t]*:[ \t]*\\([^ \t\n]+\\)[ \t]*" | |
502522b2 GM |
1659 | suffix "$")) |
1660 | (re-unibyte | |
1661 | (concat | |
1662 | "^" prefix | |
cfe98f50 | 1663 | "[ \t]*unibyte[ \t]*:[ \t]*\\([^ \t\n]+\\)[ \t]*" |
502522b2 GM |
1664 | suffix "$")) |
1665 | (re-end | |
cfe98f50 | 1666 | (concat "^" prefix "[ \t]*End *:[ \t]*" suffix "$")) |
502522b2 GM |
1667 | (pos (point))) |
1668 | (re-search-forward re-end tail-end 'move) | |
1669 | (setq tail-end (point)) | |
1670 | (goto-char pos) | |
1671 | (when (and set-auto-coding-for-load | |
1672 | (re-search-forward re-unibyte tail-end t)) | |
1673 | (setq coding-system 'raw-text)) | |
1674 | (when (and (not coding-system) | |
1675 | (re-search-forward re-coding tail-end t)) | |
1676 | (setq coding-system (intern (match-string 1))) | |
1677 | (or (coding-system-p coding-system) | |
1678 | (setq coding-system nil)))))) | |
447404a3 CW |
1679 | coding-system) |
1680 | ;; Finally, try all the `auto-coding-functions'. | |
1681 | (let ((funcs auto-coding-functions) | |
1682 | (coding-system nil)) | |
1683 | (while (and funcs (not coding-system)) | |
1684 | (setq coding-system (condition-case e | |
1685 | (save-excursion | |
1686 | (goto-char (point-min)) | |
1687 | (funcall (pop funcs) size)) | |
1688 | (error nil)))) | |
502522b2 | 1689 | coding-system))) |
63561304 KH |
1690 | |
1691 | (setq set-auto-coding-function 'set-auto-coding) | |
87aba788 | 1692 | |
4ed46869 | 1693 | (defun after-insert-file-set-buffer-file-coding-system (inserted) |
521d4010 | 1694 | "Set `buffer-file-coding-system' of current buffer after text is inserted." |
4ed46869 KH |
1695 | (if last-coding-system-used |
1696 | (let ((coding-system | |
1697 | (find-new-buffer-file-coding-system last-coding-system-used)) | |
1698 | (modified-p (buffer-modified-p))) | |
0269ddfb | 1699 | (when coding-system |
71983219 | 1700 | (set-buffer-file-coding-system coding-system t) |
d0c26c63 KH |
1701 | (if (and enable-multibyte-characters |
1702 | (or (eq coding-system 'no-conversion) | |
ccb77b4e | 1703 | (eq (coding-system-type coding-system) 5)) |
136e48e4 KH |
1704 | ;; If buffer was unmodified and the size is the |
1705 | ;; same as INSERTED, we must be visiting it. | |
1706 | (not modified-p) | |
1707 | (= (buffer-size) inserted)) | |
ccb77b4e RS |
1708 | ;; For coding systems no-conversion and raw-text..., |
1709 | ;; edit the buffer as unibyte. | |
d0c26c63 KH |
1710 | (let ((pos-byte (position-bytes (+ (point) inserted)))) |
1711 | (set-buffer-multibyte nil) | |
1712 | (setq inserted (- pos-byte (position-bytes (point)))))) | |
0269ddfb | 1713 | (set-buffer-modified-p modified-p)))) |
d0c26c63 | 1714 | inserted) |
4ed46869 | 1715 | |
84c9d215 KH |
1716 | (add-hook 'after-insert-file-functions |
1717 | 'after-insert-file-set-buffer-file-coding-system) | |
4ed46869 | 1718 | |
8057896b | 1719 | ;; The coding-spec and eol-type of coding-system returned is decided |
4ed46869 KH |
1720 | ;; independently in the following order. |
1721 | ;; 1. That of buffer-file-coding-system locally bound. | |
1722 | ;; 2. That of CODING. | |
1723 | ||
1724 | (defun find-new-buffer-file-coding-system (coding) | |
1725 | "Return a coding system for a buffer when a file of CODING is inserted. | |
a73a8c89 KH |
1726 | The local variable `buffer-file-coding-system' of the current buffer |
1727 | is set to the returned value. | |
509064c5 | 1728 | Return nil if there's no need to set `buffer-file-coding-system'." |
4ed46869 | 1729 | (let (local-coding local-eol |
b685f8d6 | 1730 | found-coding found-eol |
4ed46869 KH |
1731 | new-coding new-eol) |
1732 | (if (null coding) | |
1733 | ;; Nothing found about coding. | |
1734 | nil | |
1735 | ||
b685f8d6 RS |
1736 | ;; Get information of `buffer-file-coding-system' in LOCAL-EOL |
1737 | ;; and LOCAL-CODING. | |
1738 | (setq local-eol (coding-system-eol-type buffer-file-coding-system)) | |
1739 | (if (null (numberp local-eol)) | |
1740 | ;; But eol-type is not yet set. | |
1741 | (setq local-eol nil)) | |
0269ddfb KH |
1742 | (if (and buffer-file-coding-system |
1743 | (not (eq (coding-system-type buffer-file-coding-system) t))) | |
1744 | ;; This is not `undecided'. | |
1745 | (setq local-coding (coding-system-base buffer-file-coding-system))) | |
b685f8d6 RS |
1746 | |
1747 | (if (and (local-variable-p 'buffer-file-coding-system) | |
1748 | local-eol local-coding) | |
4ed46869 KH |
1749 | ;; The current buffer has already set full coding-system, we |
1750 | ;; had better not change it. | |
1751 | nil | |
1752 | ||
8057896b | 1753 | (setq found-eol (coding-system-eol-type coding)) |
4ed46869 | 1754 | (if (null (numberp found-eol)) |
be02cd54 EZ |
1755 | ;; But eol-type is not found. |
1756 | ;; If EOL conversions are inhibited, force unix eol-type. | |
1757 | (setq found-eol (if inhibit-eol-conversion 0))) | |
c76b5c99 KH |
1758 | (if (eq (coding-system-type coding) t) |
1759 | (setq found-coding 'undecided) | |
1760 | (setq found-coding (coding-system-base coding))) | |
1761 | ||
1762 | (if (and (not found-eol) (eq found-coding 'undecided)) | |
1763 | ;; No valid coding information found. | |
1764 | nil | |
1765 | ||
1766 | ;; Some coding information (eol or text) found. | |
1767 | ||
1768 | ;; The local setting takes precedence over the found one. | |
1769 | (setq new-coding (if (local-variable-p 'buffer-file-coding-system) | |
1770 | (or local-coding found-coding) | |
1771 | (or found-coding local-coding))) | |
1772 | (setq new-eol (if (local-variable-p 'buffer-file-coding-system) | |
1773 | (or local-eol found-eol) | |
1774 | (or found-eol local-eol))) | |
1775 | ||
1776 | (let ((eol-type (coding-system-eol-type new-coding))) | |
1777 | (if (and (numberp new-eol) (vectorp eol-type)) | |
1778 | (aref eol-type new-eol) | |
1779 | new-coding))))))) | |
4ed46869 | 1780 | |
fe831d33 GV |
1781 | (defun modify-coding-system-alist (target-type regexp coding-system) |
1782 | "Modify one of look up tables for finding a coding system on I/O operation. | |
8c453b46 RS |
1783 | There are three of such tables, `file-coding-system-alist', |
1784 | `process-coding-system-alist', and `network-coding-system-alist'. | |
fe831d33 GV |
1785 | |
1786 | TARGET-TYPE specifies which of them to modify. | |
8c453b46 RS |
1787 | If it is `file', it affects `file-coding-system-alist' (which see). |
1788 | If it is `process', it affects `process-coding-system-alist' (which see). | |
e8dd0160 | 1789 | If it is `network', it affects `network-coding-system-alist' (which see). |
fe831d33 GV |
1790 | |
1791 | REGEXP is a regular expression matching a target of I/O operation. | |
1792 | The target is a file name if TARGET-TYPE is `file', a program name if | |
1793 | TARGET-TYPE is `process', or a network service name or a port number | |
1794 | to connect to if TARGET-TYPE is `network'. | |
1795 | ||
1796 | CODING-SYSTEM is a coding system to perform code conversion on the I/O | |
8c453b46 RS |
1797 | operation, or a cons cell (DECODING . ENCODING) specifying the coding systems |
1798 | for decoding and encoding respectively, | |
1799 | or a function symbol which, when called, returns such a cons cell." | |
fe831d33 GV |
1800 | (or (memq target-type '(file process network)) |
1801 | (error "Invalid target type: %s" target-type)) | |
1802 | (or (stringp regexp) | |
1803 | (and (eq target-type 'network) (integerp regexp)) | |
1804 | (error "Invalid regular expression: %s" regexp)) | |
1805 | (if (symbolp coding-system) | |
1806 | (if (not (fboundp coding-system)) | |
1807 | (progn | |
1808 | (check-coding-system coding-system) | |
1809 | (setq coding-system (cons coding-system coding-system)))) | |
1810 | (check-coding-system (car coding-system)) | |
1811 | (check-coding-system (cdr coding-system))) | |
1812 | (cond ((eq target-type 'file) | |
1813 | (let ((slot (assoc regexp file-coding-system-alist))) | |
1814 | (if slot | |
1815 | (setcdr slot coding-system) | |
1816 | (setq file-coding-system-alist | |
1817 | (cons (cons regexp coding-system) | |
1818 | file-coding-system-alist))))) | |
1819 | ((eq target-type 'process) | |
1820 | (let ((slot (assoc regexp process-coding-system-alist))) | |
1821 | (if slot | |
1822 | (setcdr slot coding-system) | |
1823 | (setq process-coding-system-alist | |
1824 | (cons (cons regexp coding-system) | |
1825 | process-coding-system-alist))))) | |
1826 | (t | |
1827 | (let ((slot (assoc regexp network-coding-system-alist))) | |
1828 | (if slot | |
1829 | (setcdr slot coding-system) | |
1830 | (setq network-coding-system-alist | |
1831 | (cons (cons regexp coding-system) | |
1832 | network-coding-system-alist))))))) | |
1833 | ||
b25eef20 | 1834 | (defun make-translation-table (&rest args) |
a284eea3 | 1835 | "Make a translation table from arguments. |
d38b07f9 | 1836 | A translation table is a char table intended for character |
a284eea3 DL |
1837 | translation in CCL programs. |
1838 | ||
d38b07f9 | 1839 | Each argument is a list of elements of the form (FROM . TO), where FROM |
a284eea3 | 1840 | is a character to be translated to TO. |
13d5617d | 1841 | |
452fdb31 | 1842 | FROM can be a generic character (see `make-char'). In this case, TO is |
d38b07f9 | 1843 | a generic character containing the same number of characters, or an |
452fdb31 | 1844 | ordinary character. If FROM and TO are both generic characters, all |
b25eef20 | 1845 | characters belonging to FROM are translated to characters belonging to TO |
4e003d37 KH |
1846 | without changing their position code(s). |
1847 | ||
1848 | The arguments and forms in each argument are processed in the given | |
1849 | order, and if a previous form already translates TO to some other | |
1850 | character, say TO-ALT, FROM is also translated to TO-ALT." | |
f967223b | 1851 | (let ((table (make-char-table 'translation-table)) |
a73a8c89 KH |
1852 | revlist) |
1853 | (while args | |
1854 | (let ((elts (car args))) | |
1855 | (while elts | |
13d5617d KH |
1856 | (let* ((from (car (car elts))) |
1857 | (from-i 0) ; degree of freedom of FROM | |
1858 | (from-rev (nreverse (split-char from))) | |
1859 | (to (cdr (car elts))) | |
1860 | (to-i 0) ; degree of freedom of TO | |
1861 | (to-rev (nreverse (split-char to)))) | |
1862 | ;; Check numbers of heading 0s in FROM-REV and TO-REV. | |
1863 | (while (eq (car from-rev) 0) | |
1864 | (setq from-i (1+ from-i) from-rev (cdr from-rev))) | |
1865 | (while (eq (car to-rev) 0) | |
1866 | (setq to-i (1+ to-i) to-rev (cdr to-rev))) | |
1867 | (if (and (/= from-i to-i) (/= to-i 0)) | |
1868 | (error "Invalid character pair (%d . %d)" from to)) | |
b25eef20 KH |
1869 | ;; If we have already translated TO to TO-ALT, FROM should |
1870 | ;; also be translated to TO-ALT. But, this is only if TO | |
1871 | ;; is a generic character or TO-ALT is not a generic | |
13d5617d KH |
1872 | ;; character. |
1873 | (let ((to-alt (aref table to))) | |
1874 | (if (and to-alt | |
1875 | (or (> to-i 0) (not (generic-char-p to-alt)))) | |
1876 | (setq to to-alt))) | |
1877 | (if (> from-i 0) | |
1878 | (set-char-table-default table from to) | |
1879 | (aset table from to)) | |
b25eef20 KH |
1880 | ;; If we have already translated some chars to FROM, they |
1881 | ;; should also be translated to TO. | |
a73a8c89 KH |
1882 | (let ((l (assq from revlist))) |
1883 | (if l | |
1884 | (let ((ch (car l))) | |
1885 | (setcar l to) | |
1886 | (setq l (cdr l)) | |
1887 | (while l | |
1888 | (aset table ch to) | |
1889 | (setq l (cdr l)) )))) | |
1890 | ;; Now update REVLIST. | |
1891 | (let ((l (assq to revlist))) | |
1892 | (if l | |
1893 | (setcdr l (cons from (cdr l))) | |
1894 | (setq revlist (cons (list to from) revlist))))) | |
1895 | (setq elts (cdr elts)))) | |
1896 | (setq args (cdr args))) | |
1897 | ;; Return TABLE just created. | |
1898 | table)) | |
1899 | ||
c76b5c99 KH |
1900 | (defun make-translation-table-from-vector (vec) |
1901 | "Make translation table from decoding vector VEC. | |
9e3b6057 DL |
1902 | VEC is an array of 256 elements to map unibyte codes to multibyte |
1903 | characters. Elements may be nil for undefined code points. | |
c76b5c99 KH |
1904 | See also the variable `nonascii-translation-table'." |
1905 | (let ((table (make-char-table 'translation-table)) | |
1906 | (rev-table (make-char-table 'translation-table)) | |
c76b5c99 | 1907 | ch) |
9e3b6057 | 1908 | (dotimes (i 256) |
c76b5c99 | 1909 | (setq ch (aref vec i)) |
9e3b6057 DL |
1910 | (when ch |
1911 | (aset table i ch) | |
1912 | (if (>= ch 256) | |
1913 | (aset rev-table ch i)))) | |
c76b5c99 KH |
1914 | (set-char-table-extra-slot table 0 rev-table) |
1915 | table)) | |
1916 | ||
f967223b | 1917 | (defun define-translation-table (symbol &rest args) |
a284eea3 DL |
1918 | "Define SYMBOL as the name of translation table made by ARGS. |
1919 | This sets up information so that the table can be used for | |
1920 | translations in a CCL program. | |
b25eef20 | 1921 | |
a284eea3 DL |
1922 | If the first element of ARGS is a char-table whose purpose is |
1923 | `translation-table', just define SYMBOL to name it. (Note that this | |
1924 | function does not bind SYMBOL.) | |
007c79c8 | 1925 | |
a284eea3 | 1926 | Any other ARGS should be suitable as arguments of the function |
007c79c8 | 1927 | `make-translation-table' (which see). |
b25eef20 | 1928 | |
452fdb31 | 1929 | This function sets properties `translation-table' and |
521d4010 DL |
1930 | `translation-table-id' of SYMBOL to the created table itself and the |
1931 | identification number of the table respectively. It also registers | |
1932 | the table in `translation-table-vector'." | |
007c79c8 KH |
1933 | (let ((table (if (and (char-table-p (car args)) |
1934 | (eq (char-table-subtype (car args)) | |
1935 | 'translation-table)) | |
1936 | (car args) | |
1937 | (apply 'make-translation-table args))) | |
f967223b | 1938 | (len (length translation-table-vector)) |
d9e3229d | 1939 | (id 0) |
b25eef20 | 1940 | (done nil)) |
f967223b | 1941 | (put symbol 'translation-table table) |
b25eef20 KH |
1942 | (while (not done) |
1943 | (if (>= id len) | |
f967223b KH |
1944 | (setq translation-table-vector |
1945 | (vconcat translation-table-vector (make-vector len nil)))) | |
1946 | (let ((slot (aref translation-table-vector id))) | |
b25eef20 KH |
1947 | (if (or (not slot) |
1948 | (eq (car slot) symbol)) | |
1949 | (progn | |
f967223b | 1950 | (aset translation-table-vector id (cons symbol table)) |
007c79c8 KH |
1951 | (setq done t)) |
1952 | (setq id (1+ id))))) | |
f967223b | 1953 | (put symbol 'translation-table-id id) |
d9e3229d KH |
1954 | id)) |
1955 | ||
35554641 KH |
1956 | (put 'with-category-table 'lisp-indent-function 1) |
1957 | ||
1958 | (defmacro with-category-table (category-table &rest body) | |
1959 | `(let ((current-category-table (category-table))) | |
1960 | (set-category-table ,category-table) | |
1961 | (unwind-protect | |
1962 | (progn ,@body) | |
1963 | (set-category-table current-category-table)))) | |
1964 | ||
394e4eb0 DL |
1965 | (defun define-translation-hash-table (symbol table) |
1966 | "Define SYMBOL as the name of the hash translation TABLE for use in CCL. | |
1967 | ||
1968 | Analogous to `define-translation-table', but updates | |
1969 | `translation-hash-table-vector' and the table is for use in the CCL | |
1970 | `lookup-integer' and `lookup-character' functions." | |
1971 | (unless (and (symbolp symbol) | |
1972 | (hash-table-p table)) | |
1973 | (error "Bad args to define-translation-hash-table")) | |
1974 | (let ((len (length translation-hash-table-vector)) | |
1975 | (id 0) | |
1976 | done) | |
1977 | (put symbol 'translation-hash-table table) | |
1978 | (while (not done) | |
1979 | (if (>= id len) | |
1980 | (setq translation-hash-table-vector | |
1981 | (vconcat translation-hash-table-vector [nil]))) | |
1982 | (let ((slot (aref translation-hash-table-vector id))) | |
1983 | (if (or (not slot) | |
1984 | (eq (car slot) symbol)) | |
1985 | (progn | |
1986 | (aset translation-hash-table-vector id (cons symbol table)) | |
1987 | (setq done t)) | |
1988 | (setq id (1+ id))))) | |
1989 | (put symbol 'translation-hash-table-id id) | |
1990 | id)) | |
1991 | ||
69eba008 KH |
1992 | ;;; Initialize some variables. |
1993 | ||
1994 | (put 'use-default-ascent 'char-table-extra-slots 0) | |
1995 | (setq use-default-ascent (make-char-table 'use-default-ascent)) | |
d6d6d592 KH |
1996 | (put 'ignore-relative-composition 'char-table-extra-slots 0) |
1997 | (setq ignore-relative-composition | |
1998 | (make-char-table 'ignore-relative-composition)) | |
69eba008 | 1999 | |
d9f6dfe6 CW |
2000 | |
2001 | ;;; Built-in auto-coding-functions: | |
2002 | ||
2003 | (defun sgml-xml-auto-coding-function (size) | |
2004 | "Determine whether the buffer is XML, and if so, its encoding. | |
2005 | This function is intended to be added to `auto-coding-functions'." | |
623b1d5e | 2006 | (when (re-search-forward "\\`[[:space:]\n]*<\\?xml" nil t) |
d9f6dfe6 CW |
2007 | (let ((end (save-excursion |
2008 | ;; This is a hack. | |
2f4e9c47 | 2009 | (re-search-forward "\"\\s-*\\?>" size t)))) |
d9f6dfe6 CW |
2010 | (when end |
2011 | (if (re-search-forward "encoding=\"\\(.+?\\)\"" end t) | |
447404a3 CW |
2012 | (let* ((match (match-string 1)) |
2013 | (sym (intern (downcase match)))) | |
2014 | (if (coding-system-p sym) | |
2015 | sym | |
2016 | (message "Warning: unknown coding system \"%s\"" match) | |
2017 | nil)) | |
d9f6dfe6 CW |
2018 | 'utf-8))))) |
2019 | ||
447404a3 CW |
2020 | (defun sgml-html-meta-auto-coding-function (size) |
2021 | "If the buffer has an HTML meta tag, use it to determine encoding. | |
2022 | This function is intended to be added to `auto-coding-functions'." | |
2023 | (setq size (min size | |
2024 | ;; Only search forward 10 lines | |
2025 | (save-excursion | |
2026 | (forward-line 10) | |
2027 | (point)))) | |
2028 | (when (and (search-forward "<html>" size t) | |
2029 | (re-search-forward "<meta\\s-+http-equiv=\"content-type\"\\s-+content=\"text/\\sw+;\\s-*charset=\\(.+?\\)\"" size t)) | |
2030 | (let* ((match (match-string 1)) | |
2031 | (sym (intern (downcase match)))) | |
2032 | (if (coding-system-p sym) | |
2033 | sym | |
2034 | (message "Warning: unknown coding system \"%s\"" match) | |
2035 | nil)))) | |
0bca779a | 2036 | |
69eba008 | 2037 | ;;; |
4ed46869 KH |
2038 | (provide 'mule) |
2039 | ||
2040 | ;;; mule.el ends here |