Commit | Line | Data |
---|---|---|
4ed46869 KH |
1 | ;;; mule.el --- basic commands for mulitilingual environment |
2 | ||
4ed46869 | 3 | ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN. |
58cfed09 | 4 | ;; Licensed to the Free Software Foundation. |
4ed46869 KH |
5 | |
6 | ;; Keywords: mule, multilingual, character set, coding system | |
7 | ||
8 | ;; This file is part of GNU Emacs. | |
9 | ||
10 | ;; GNU Emacs is free software; you can redistribute it and/or modify | |
11 | ;; it under the terms of the GNU General Public License as published by | |
12 | ;; the Free Software Foundation; either version 2, or (at your option) | |
13 | ;; any later version. | |
14 | ||
15 | ;; GNU Emacs is distributed in the hope that it will be useful, | |
16 | ;; but WITHOUT ANY WARRANTY; without even the implied warranty of | |
17 | ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
18 | ;; GNU General Public License for more details. | |
19 | ||
20 | ;; You should have received a copy of the GNU General Public License | |
369314dc KH |
21 | ;; along with GNU Emacs; see the file COPYING. If not, write to the |
22 | ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
23 | ;; Boston, MA 02111-1307, USA. | |
4ed46869 KH |
24 | |
25 | ;;; Code: | |
26 | ||
d2542b42 | 27 | (defconst mule-version "5.0 (SAKAKI)" "\ |
4ed46869 KH |
28 | Version number and name of this version of MULE (multilingual environment).") |
29 | ||
35554641 | 30 | (defconst mule-version-date "1999.12.7" "\ |
4ed46869 KH |
31 | Distribution date of this version of MULE (multilingual environment).") |
32 | ||
33 | (defun load-with-code-conversion (fullname file &optional noerror nomessage) | |
0f69cb38 KH |
34 | "Execute a file of Lisp code named FILE whose absolute name is FULLNAME. |
35 | The file contents are decoded before evaluation if necessary. | |
4ed46869 KH |
36 | If optional second arg NOERROR is non-nil, |
37 | report no error if FILE doesn't exist. | |
38 | Print messages at start and end of loading unless | |
39 | optional third arg NOMESSAGE is non-nil. | |
40 | Return t if file exists." | |
41 | (if (null (file-readable-p fullname)) | |
42 | (and (null noerror) | |
43 | (signal 'file-error (list "Cannot open load file" file))) | |
44 | ;; Read file with code conversion, and then eval. | |
45 | (let* ((buffer | |
46 | ;; To avoid any autoloading, set default-major-mode to | |
47 | ;; fundamental-mode. | |
88162676 RS |
48 | ;; So that we don't get completely screwed if the |
49 | ;; file is encoded in some complicated character set, | |
50 | ;; read it with real decoding, as a multibyte buffer, | |
51 | ;; even if this is a --unibyte Emacs session. | |
52 | (let ((default-major-mode 'fundamental-mode) | |
53 | (default-enable-multibyte-characters t)) | |
4ed46869 KH |
54 | ;; We can't use `generate-new-buffer' because files.el |
55 | ;; is not yet loaded. | |
56 | (get-buffer-create (generate-new-buffer-name " *load*")))) | |
db5cae4b SM |
57 | (load-in-progress t) |
58 | (source (save-match-data (string-match "\\.el\\'" fullname)))) | |
59 | (unless nomessage | |
60 | (if source | |
61 | (message "Loading %s (source)..." file) | |
62 | (message "Loading %s..." file))) | |
63 | (when purify-flag | |
64 | (setq preloaded-file-list (cons file preloaded-file-list))) | |
4ed46869 | 65 | (unwind-protect |
a6acd8a2 | 66 | (let ((load-file-name fullname) |
1c4cc63a | 67 | (set-auto-coding-for-load t) |
a6acd8a2 | 68 | (inhibit-file-name-operation nil)) |
4ed46869 KH |
69 | (save-excursion |
70 | (set-buffer buffer) | |
71 | (insert-file-contents fullname) | |
7d276780 EZ |
72 | ;; If the loaded file was inserted with no-conversion or |
73 | ;; raw-text coding system, make the buffer unibyte. | |
74 | ;; Otherwise, eval-buffer might try to interpret random | |
75 | ;; binary junk as multibyte characters. | |
76 | (if (and enable-multibyte-characters | |
77 | (or (eq (coding-system-type last-coding-system-used) 5) | |
78 | (eq last-coding-system-used 'no-conversion))) | |
79 | (set-buffer-multibyte nil)) | |
4ed46869 KH |
80 | ;; Make `kill-buffer' quiet. |
81 | (set-buffer-modified-p nil)) | |
0f69cb38 | 82 | ;; Have the original buffer current while we eval. |
88162676 RS |
83 | (eval-buffer buffer nil file |
84 | ;; If this Emacs is running with --unibyte, | |
85 | ;; convert multibyte strings to unibyte | |
86 | ;; after reading them. | |
ba74e833 | 87 | ;; (not default-enable-multibyte-characters) |
8dd08b5b | 88 | nil t |
ba74e833 | 89 | )) |
cfc70cdf RS |
90 | (let (kill-buffer-hook kill-buffer-query-functions) |
91 | (kill-buffer buffer))) | |
4ed46869 | 92 | (let ((hook (assoc file after-load-alist))) |
db5cae4b SM |
93 | (when hook |
94 | (mapcar (function eval) (cdr hook)))) | |
95 | (unless (or nomessage noninteractive) | |
96 | (if source | |
97 | (message "Loading %s (source)...done" file) | |
98 | (message "Loading %s...done" file))) | |
4ed46869 KH |
99 | t))) |
100 | ||
101 | ;; API (Application Program Interface) for charsets. | |
102 | ||
2e02a76f RS |
103 | ;; Return t if OBJ is a quoted symbol |
104 | ;; and the symbol is the name of a standard charset. | |
105 | (defsubst charset-quoted-standard-p (obj) | |
106 | (and (listp obj) (eq (car obj) 'quote) | |
107 | (symbolp (car-safe (cdr obj))) | |
108 | (let ((vector (get (car-safe (cdr obj)) 'charset))) | |
109 | (and (vectorp vector) | |
110 | (< (aref vector 0) 160))))) | |
4ed46869 KH |
111 | |
112 | (defsubst charsetp (object) | |
e4a7fdfc | 113 | "T if OBJECT is a charset." |
4ed46869 KH |
114 | (and (symbolp object) (vectorp (get object 'charset)))) |
115 | ||
116 | (defsubst charset-info (charset) | |
117 | "Return a vector of information of CHARSET. | |
118 | The elements of the vector are: | |
119 | CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION, | |
120 | LEADING-CODE-BASE, LEADING-CODE-EXT, | |
121 | ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE, | |
122 | REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION, | |
123 | PLIST, | |
124 | where | |
125 | CHARSET-ID (integer) is the identification number of the charset. | |
aa8bb645 KH |
126 | BYTES (integer) is the length of multi-byte form of a character in |
127 | the charset: one of 1, 2, 3, and 4. | |
4ed46869 KH |
128 | DIMENSION (integer) is the number of bytes to represent a character of |
129 | the charset: 1 or 2. | |
130 | CHARS (integer) is the number of characters in a dimension: 94 or 96. | |
4ed46869 KH |
131 | WIDTH (integer) is the number of columns a character in the charset |
132 | occupies on the screen: one of 0, 1, and 2. | |
133 | DIRECTION (integer) is the rendering direction of characters in the | |
08b4ace5 KH |
134 | charset when rendering. If 0, render from left to right, else |
135 | render from right to left. | |
4ed46869 KH |
136 | LEADING-CODE-BASE (integer) is the base leading-code for the |
137 | charset. | |
138 | LEADING-CODE-EXT (integer) is the extended leading-code for the | |
139 | charset. All charsets of less than 0xA0 has the value 0. | |
140 | ISO-FINAL-CHAR (character) is the final character of the | |
141 | corresponding ISO 2022 charset. | |
142 | ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked | |
143 | while encoding to variants of ISO 2022 coding system, one of the | |
144 | following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR). | |
145 | REVERSE-CHARSET (integer) is the charset which differs only in | |
146 | LEFT-TO-RIGHT value from the charset. If there's no such a | |
147 | charset, the value is -1. | |
148 | SHORT-NAME (string) is the short name to refer to the charset. | |
149 | LONG-NAME (string) is the long name to refer to the charset | |
150 | DESCRIPTION (string) is the description string of the charset. | |
151 | PLIST (property list) may contain any type of information a user | |
152 | want to put and get by functions `put-charset-property' and | |
153 | `get-charset-property' respectively." | |
154 | (get charset 'charset)) | |
155 | ||
40c81f74 PE |
156 | ;; It is better not to use backquote in this file, |
157 | ;; because that makes a bootstrapping problem | |
158 | ;; if you need to recompile all the Lisp files using interpreted code. | |
159 | ||
4ed46869 KH |
160 | (defmacro charset-id (charset) |
161 | "Return charset identification number of CHARSET." | |
2e02a76f | 162 | (if (charset-quoted-standard-p charset) |
4ed46869 | 163 | (aref (charset-info (nth 1 charset)) 0) |
40c81f74 | 164 | (list 'aref (list 'charset-info charset) 0))) |
4ed46869 KH |
165 | |
166 | (defmacro charset-bytes (charset) | |
900dc6e3 KH |
167 | "Return bytes of CHARSET. |
168 | See the function `charset-info' for more detail." | |
2e02a76f | 169 | (if (charset-quoted-standard-p charset) |
4ed46869 | 170 | (aref (charset-info (nth 1 charset)) 1) |
40c81f74 | 171 | (list 'aref (list 'charset-info charset) 1))) |
4ed46869 KH |
172 | |
173 | (defmacro charset-dimension (charset) | |
900dc6e3 KH |
174 | "Return dimension of CHARSET. |
175 | See the function `charset-info' for more detail." | |
2e02a76f | 176 | (if (charset-quoted-standard-p charset) |
4ed46869 | 177 | (aref (charset-info (nth 1 charset)) 2) |
40c81f74 | 178 | (list 'aref (list 'charset-info charset) 2))) |
4ed46869 KH |
179 | |
180 | (defmacro charset-chars (charset) | |
900dc6e3 KH |
181 | "Return character numbers contained in a dimension of CHARSET. |
182 | See the function `charset-info' for more detail." | |
2e02a76f | 183 | (if (charset-quoted-standard-p charset) |
4ed46869 | 184 | (aref (charset-info (nth 1 charset)) 3) |
40c81f74 | 185 | (list 'aref (list 'charset-info charset) 3))) |
4ed46869 KH |
186 | |
187 | (defmacro charset-width (charset) | |
900dc6e3 KH |
188 | "Return width (how many column occupied on a screen) of CHARSET. |
189 | See the function `charset-info' for more detail." | |
2e02a76f | 190 | (if (charset-quoted-standard-p charset) |
4ed46869 | 191 | (aref (charset-info (nth 1 charset)) 4) |
40c81f74 | 192 | (list 'aref (list 'charset-info charset) 4))) |
4ed46869 KH |
193 | |
194 | (defmacro charset-direction (charset) | |
900dc6e3 KH |
195 | "Return direction of CHARSET. |
196 | See the function `charset-info' for more detail." | |
2e02a76f | 197 | (if (charset-quoted-standard-p charset) |
4ed46869 | 198 | (aref (charset-info (nth 1 charset)) 5) |
40c81f74 | 199 | (list 'aref (list 'charset-info charset) 5))) |
4ed46869 KH |
200 | |
201 | (defmacro charset-iso-final-char (charset) | |
900dc6e3 KH |
202 | "Return final char of CHARSET. |
203 | See the function `charset-info' for more detail." | |
2e02a76f | 204 | (if (charset-quoted-standard-p charset) |
4ed46869 | 205 | (aref (charset-info (nth 1 charset)) 8) |
40c81f74 | 206 | (list 'aref (list 'charset-info charset) 8))) |
4ed46869 KH |
207 | |
208 | (defmacro charset-iso-graphic-plane (charset) | |
900dc6e3 KH |
209 | "Return graphic plane of CHARSET. |
210 | See the function `charset-info' for more detail." | |
2e02a76f | 211 | (if (charset-quoted-standard-p charset) |
4ed46869 | 212 | (aref (charset-info (nth 1 charset)) 9) |
40c81f74 | 213 | (list 'aref (list 'charset-info charset) 9))) |
4ed46869 KH |
214 | |
215 | (defmacro charset-reverse-charset (charset) | |
900dc6e3 KH |
216 | "Return reverse charset of CHARSET. |
217 | See the function `charset-info' for more detail." | |
2e02a76f | 218 | (if (charset-quoted-standard-p charset) |
4ed46869 | 219 | (aref (charset-info (nth 1 charset)) 10) |
40c81f74 | 220 | (list 'aref (list 'charset-info charset) 10))) |
4ed46869 KH |
221 | |
222 | (defmacro charset-short-name (charset) | |
900dc6e3 KH |
223 | "Return short name of CHARSET. |
224 | See the function `charset-info' for more detail." | |
2e02a76f | 225 | (if (charset-quoted-standard-p charset) |
4ed46869 | 226 | (aref (charset-info (nth 1 charset)) 11) |
40c81f74 | 227 | (list 'aref (list 'charset-info charset) 11))) |
4ed46869 KH |
228 | |
229 | (defmacro charset-long-name (charset) | |
900dc6e3 KH |
230 | "Return long name of CHARSET. |
231 | See the function `charset-info' for more detail." | |
2e02a76f | 232 | (if (charset-quoted-standard-p charset) |
4ed46869 | 233 | (aref (charset-info (nth 1 charset)) 12) |
40c81f74 | 234 | (list 'aref (list 'charset-info charset) 12))) |
4ed46869 KH |
235 | |
236 | (defmacro charset-description (charset) | |
0611934b | 237 | "Return description of CHARSET. |
900dc6e3 | 238 | See the function `charset-info' for more detail." |
2e02a76f | 239 | (if (charset-quoted-standard-p charset) |
4ed46869 | 240 | (aref (charset-info (nth 1 charset)) 13) |
40c81f74 | 241 | (list 'aref (list 'charset-info charset) 13))) |
4ed46869 KH |
242 | |
243 | (defmacro charset-plist (charset) | |
900dc6e3 KH |
244 | "Return list charset property of CHARSET. |
245 | See the function `charset-info' for more detail." | |
40c81f74 PE |
246 | (list 'aref |
247 | (if (charset-quoted-standard-p charset) | |
248 | (charset-info (nth 1 charset)) | |
249 | (list 'charset-info charset)) | |
250 | 14)) | |
4ed46869 KH |
251 | |
252 | (defun set-charset-plist (charset plist) | |
e8dd0160 | 253 | "Set CHARSET's property list to PLIST, and return PLIST." |
4ed46869 KH |
254 | (aset (charset-info charset) 14 plist)) |
255 | ||
d3e7e7cf EZ |
256 | (defun make-char (charset &optional code1 code2) |
257 | "Return a character of CHARSET whose position codes are CODE1 and CODE2. | |
f98e2797 | 258 | CODE1 and CODE2 are optional, but if you don't supply |
a9fb0b58 KH |
259 | sufficient position codes, return a generic character which stands for |
260 | all characters or group of characters in the character set. | |
d3e7e7cf | 261 | A generic character can be used to index a char table (e.g. syntax-table). |
9ff05eae KH |
262 | |
263 | Such character sets as ascii, eight-bit-control, and eight-bit-graphic | |
264 | don't have corresponding generic characters. If CHARSET is one of | |
265 | them and you don't supply CODE1, return the character of the smallest | |
266 | code in CHARSET. | |
267 | ||
d3e7e7cf | 268 | If CODE1 or CODE2 are invalid (out of range), this function signals an error." |
636799f2 | 269 | (make-char-internal (charset-id charset) code1 code2)) |
a6acd8a2 KH |
270 | |
271 | (put 'make-char 'byte-compile | |
272 | (function | |
273 | (lambda (form) | |
274 | (let ((charset (nth 1 form))) | |
275 | (if (charset-quoted-standard-p charset) | |
276 | (byte-compile-normal-call | |
277 | (cons 'make-char-internal | |
278 | (cons (charset-id (nth 1 charset)) (nthcdr 2 form)))) | |
279 | (byte-compile-normal-call | |
280 | (cons 'make-char-internal | |
281 | (cons (list 'charset-id charset) (nthcdr 2 form))))))))) | |
4ed46869 | 282 | |
d3675a42 | 283 | (defun charset-list () |
900dc6e3 KH |
284 | "Return list of charsets ever defined. |
285 | ||
d3675a42 | 286 | This function is provided for backward compatibility. |
900dc6e3 | 287 | Now we have the variable `charset-list'." |
d3675a42 KH |
288 | charset-list) |
289 | ||
13d5617d KH |
290 | (defsubst generic-char-p (char) |
291 | "Return t if and only if CHAR is a generic character. | |
292 | See also the documentation of make-char." | |
1d935669 KH |
293 | (and (>= char 0400) |
294 | (let ((l (split-char char))) | |
295 | (and (or (= (nth 1 l) 0) (eq (nth 2 l) 0)) | |
296 | (not (eq (car l) 'composition)))))) | |
69eba008 | 297 | |
cc57cc54 | 298 | (defun decode-char (ccs code-point &optional restriction) |
7f341749 | 299 | "Return character specified by coded character set CCS and CODE-POINT in it. |
cc57cc54 | 300 | Return nil if such a character is not supported. |
e76938e7 DL |
301 | Currently the only supported coded character set is `ucs' (ISO/IEC |
302 | 10646: Universal Multi-Octet Coded Character Set). | |
cc57cc54 KH |
303 | |
304 | Optional argument RESTRICTION specifies a way to map the pair of CCS | |
305 | and CODE-POINT to a chracter. Currently not supported and just ignored." | |
306 | (cond ((eq ccs 'ucs) | |
307 | (cond ((< code-point 128) | |
308 | code-point) | |
309 | ((< code-point 256) | |
310 | (make-char 'latin-iso8859-1 code-point)) | |
311 | ((< code-point #x2500) | |
312 | (setq code-point (- code-point #x0100)) | |
313 | (make-char 'mule-unicode-0100-24ff | |
314 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
315 | ((< code-point #x33ff) | |
316 | (setq code-point (- code-point #x2500)) | |
317 | (make-char 'mule-unicode-2500-33ff | |
318 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
319 | ((and (>= code-point #xe000) (< code-point #x10000)) | |
320 | (setq code-point (- code-point #xe000)) | |
321 | (make-char 'mule-unicode-e000-ffff | |
322 | (+ (/ code-point 96) 32) (+ (% code-point 96) 32))) | |
323 | )))) | |
324 | ||
325 | (defun encode-char (char ccs &optional restriction) | |
7f341749 | 326 | "Return code-point in coded character set CCS that corresponds to CHAR. |
cc57cc54 | 327 | Return nil if CHAR is not included in CCS. |
e76938e7 DL |
328 | Currently the only supported coded character set is `ucs' (ISO/IEC |
329 | 10646: Universal Multi-Octet Coded Character Set). | |
7f341749 KH |
330 | |
331 | CHAR should be in one of these charsets: | |
cc57cc54 KH |
332 | ascii, latin-iso8859-1, mule-unicode-0100-24ff, mule-unicode-2500-33ff, |
333 | mule-unicode-e000-ffff | |
334 | Otherwise, return nil. | |
335 | ||
336 | Optional argument RESTRICTION specifies a way to map CHAR to a | |
337 | code-point in CCS. Currently not supported and just ignored." | |
338 | (let* ((split (split-char char)) | |
339 | (charset (car split))) | |
340 | (cond ((eq ccs 'ucs) | |
341 | (cond ((eq charset 'ascii) | |
342 | char) | |
343 | ((eq charset 'latin-iso8859-1) | |
344 | (+ (nth 1 split) 128)) | |
345 | ((eq charset 'mule-unicode-0100-24ff) | |
346 | (+ #x0100 (+ (* (- (nth 1 split) 32) 96) | |
347 | (- (nth 2 split) 32)))) | |
348 | ((eq charset 'mule-unicode-2500-33ff) | |
349 | (+ #x2500 (+ (* (- (nth 1 split) 32) 96) | |
350 | (- (nth 2 split) 32)))) | |
351 | ((eq charset 'mule-unicode-e000-ffff) | |
352 | (+ #xe000 (+ (* (- (nth 1 split) 32) 96) | |
353 | (- (nth 2 split) 32))))))))) | |
354 | ||
0269ddfb | 355 | \f |
e76938e7 | 356 | ;; Coding system stuff |
4ed46869 | 357 | |
8057896b | 358 | ;; Coding system is a symbol that has the property `coding-system'. |
4ed46869 | 359 | ;; |
8057896b KH |
360 | ;; The value of the property `coding-system' is a vector of the |
361 | ;; following format: | |
d3675a42 | 362 | ;; [TYPE MNEMONIC DOC-STRING PLIST FLAGS] |
8057896b | 363 | ;; We call this vector as coding-spec. See comments in src/coding.c |
d3675a42 | 364 | ;; for more detail. |
8057896b KH |
365 | |
366 | (defconst coding-spec-type-idx 0) | |
367 | (defconst coding-spec-mnemonic-idx 1) | |
368 | (defconst coding-spec-doc-string-idx 2) | |
0269ddfb | 369 | (defconst coding-spec-plist-idx 3) |
8057896b KH |
370 | (defconst coding-spec-flags-idx 4) |
371 | ||
a6acd8a2 KH |
372 | ;; PLIST is a property list of a coding system. To share PLIST among |
373 | ;; alias coding systems, a coding system has PLIST in coding-spec | |
374 | ;; instead of having it in normal property list of Lisp symbol. | |
375 | ;; Here's a list of coding system properties currently being used. | |
0269ddfb KH |
376 | ;; |
377 | ;; o coding-category | |
378 | ;; | |
379 | ;; The value is a coding category the coding system belongs to. The | |
1b46a680 KH |
380 | ;; function `make-coding-system' sets this value automatically |
381 | ;; unless its argument PROPERTIES specifies this property. | |
4ed46869 | 382 | ;; |
0269ddfb | 383 | ;; o alias-coding-systems |
4ed46869 | 384 | ;; |
0269ddfb KH |
385 | ;; The value is a list of coding systems of the same alias group. The |
386 | ;; first element is the coding system made at first, which we call as | |
1b46a680 KH |
387 | ;; `base coding system'. The function `make-coding-system' sets this |
388 | ;; value automatically and `define-coding-system-alias' updates it. | |
0269ddfb KH |
389 | ;; |
390 | ;; o post-read-conversion | |
391 | ;; | |
392 | ;; The value is a function to call after some text is inserted and | |
393 | ;; decoded by the coding system itself and before any functions in | |
394 | ;; `after-insert-functions' are called. The arguments to this | |
395 | ;; function is the same as those of a function in | |
4ed46869 KH |
396 | ;; `after-insert-functions', i.e. LENGTH of a text while putting point |
397 | ;; at the head of the text to be decoded | |
398 | ;; | |
0269ddfb KH |
399 | ;; o pre-write-conversion |
400 | ;; | |
401 | ;; The value is a function to call after all functions in | |
402 | ;; `write-region-annotate-functions' and `buffer-file-format' are | |
403 | ;; called, and before the text is encoded by the coding system itself. | |
404 | ;; The arguments to this function is the same as those of a function | |
405 | ;; in `write-region-annotate-functions', i.e. FROM and TO specifying | |
406 | ;; region of a text. | |
407 | ;; | |
f967223b | 408 | ;; o translation-table-for-decode |
0269ddfb | 409 | ;; |
f967223b KH |
410 | ;; The value is a translation table to be applied on decoding. See |
411 | ;; the function `make-translation-table' for the format of translation | |
412 | ;; table. | |
0269ddfb | 413 | ;; |
f967223b | 414 | ;; o translation-table-for-encode |
0269ddfb | 415 | ;; |
f967223b | 416 | ;; The value is a translation table to be applied on encoding. |
a6acd8a2 | 417 | ;; |
c11a8f77 KH |
418 | ;; o safe-chars |
419 | ;; | |
420 | ;; The value is a char table. If a character has non-nil value in it, | |
421 | ;; the character is safely supported by the coding system. This | |
422 | ;; overrides the specification of safe-charsets. | |
423 | ||
a6acd8a2 KH |
424 | ;; o safe-charsets |
425 | ;; | |
426 | ;; The value is a list of charsets safely supported by the coding | |
427 | ;; system. The value t means that all charsets Emacs handles are | |
428 | ;; supported. Even if some charset is not in this list, it doesn't | |
429 | ;; mean that the charset can't be encoded in the coding system, | |
430 | ;; instead, it just means that some other receiver of a text encoded | |
431 | ;; in the coding system won't be able to handle that charset. | |
b25eef20 KH |
432 | ;; |
433 | ;; o mime-charset | |
434 | ;; | |
435 | ;; The value is a symbol of which name is `MIME-charset' parameter of | |
436 | ;; the coding system. | |
c76b5c99 | 437 | ;; |
1322c8ca KH |
438 | ;; o charset-origin-alist |
439 | ;; | |
440 | ;; The value is a list of this form: | |
441 | ;; (CHARSET EXTERNAL-CHARSET-NAME ENCODING-FUNCTION). | |
442 | ;; ENCODING-FUNCTION is a function to encode a character in CHARSET | |
443 | ;; to the code in EXTERNAL-CHARSET-NAME. The command what-cursor-position | |
444 | ;; uses this information of the buffer-file-coding-system. | |
caa85ad9 KH |
445 | ;; ENCODING-FUNCTION may be a translation table or a symbol whose |
446 | ;; property `translation-table' is a translation table. In these case, | |
447 | ;; the translation table is used to encode the character. | |
1322c8ca | 448 | ;; |
c76b5c99 KH |
449 | ;; o valid-codes (meaningful only for a coding system based on CCL) |
450 | ;; | |
451 | ;; The value is a list to indicate valid byte ranges of the encoded | |
452 | ;; file. Each element of the list is an integer or a cons of integer. | |
453 | ;; In the former case, the integer value is a valid byte code. In the | |
e8dd0160 | 454 | ;; latter case, the integers specifies the range of valid byte codes. |
c76b5c99 | 455 | |
0269ddfb KH |
456 | |
457 | ;; Return coding-spec of CODING-SYSTEM | |
458 | (defsubst coding-system-spec (coding-system) | |
459 | (get (check-coding-system coding-system) 'coding-system)) | |
4ed46869 | 460 | |
4ed46869 | 461 | (defun coding-system-type (coding-system) |
0269ddfb KH |
462 | "Return the coding type of CODING-SYSTEM. |
463 | A coding type is an integer value indicating the encoding method | |
464 | of CODING-SYSTEM. See the function `make-coding-system' for more detail." | |
465 | (aref (coding-system-spec coding-system) coding-spec-type-idx)) | |
4ed46869 | 466 | |
4ed46869 | 467 | (defun coding-system-mnemonic (coding-system) |
0269ddfb | 468 | "Return the mnemonic character of CODING-SYSTEM. |
6e2c8840 KH |
469 | The mnemonic character of a coding system is used in mode line |
470 | to indicate the coding system. If the arg is nil, return ?-." | |
471 | (let ((spec (coding-system-spec coding-system))) | |
472 | (if spec (aref spec coding-spec-mnemonic-idx) ?-))) | |
4ed46869 | 473 | |
8057896b | 474 | (defun coding-system-doc-string (coding-system) |
0269ddfb KH |
475 | "Return the documentation string for CODING-SYSTEM." |
476 | (aref (coding-system-spec coding-system) coding-spec-doc-string-idx)) | |
4ed46869 | 477 | |
d3675a42 | 478 | (defun coding-system-plist (coding-system) |
0269ddfb KH |
479 | "Return the property list of CODING-SYSTEM." |
480 | (aref (coding-system-spec coding-system) coding-spec-plist-idx)) | |
d3675a42 | 481 | |
4ed46869 | 482 | (defun coding-system-flags (coding-system) |
0269ddfb KH |
483 | "Return `flags' of CODING-SYSTEM. |
484 | A `flags' of a coding system is a vector of length 32 indicating detailed | |
485 | information of a coding system. See the function `make-coding-system' | |
486 | for more detail." | |
487 | (aref (coding-system-spec coding-system) coding-spec-flags-idx)) | |
488 | ||
489 | (defun coding-system-get (coding-system prop) | |
490 | "Extract a value from CODING-SYSTEM's property list for property PROP." | |
491 | (plist-get (coding-system-plist coding-system) prop)) | |
492 | ||
493 | (defun coding-system-put (coding-system prop val) | |
494 | "Change value in CODING-SYSTEM's property list PROP to VAL." | |
495 | (let ((plist (coding-system-plist coding-system))) | |
496 | (if plist | |
497 | (plist-put plist prop val) | |
498 | (aset (coding-system-spec coding-system) coding-spec-plist-idx | |
499 | (list prop val))))) | |
500 | ||
501 | (defun coding-system-category (coding-system) | |
502 | "Return the coding category of CODING-SYSTEM." | |
503 | (coding-system-get coding-system 'coding-category)) | |
504 | ||
505 | (defun coding-system-base (coding-system) | |
506 | "Return the base coding system of CODING-SYSTEM. | |
d9e3229d KH |
507 | A base coding system is what made by `make-coding-system'. |
508 | Any alias nor subsidiary coding systems are not base coding system." | |
0269ddfb KH |
509 | (car (coding-system-get coding-system 'alias-coding-systems))) |
510 | ||
511 | (defalias 'coding-system-parent 'coding-system-base) | |
2598a293 | 512 | (make-obsolete 'coding-system-parent 'coding-system-base "20.3") |
0269ddfb KH |
513 | |
514 | ;; Coding system also has a property `eol-type'. | |
515 | ;; | |
516 | ;; This property indicates how the coding system handles end-of-line | |
517 | ;; format. The value is integer 0, 1, 2, or a vector of three coding | |
518 | ;; systems. Each integer value 0, 1, and 2 indicates the format of | |
519 | ;; end-of-line LF, CRLF, and CR respectively. A vector value | |
520 | ;; indicates that the format of end-of-line should be detected | |
521 | ;; automatically. Nth element of the vector is the subsidiary coding | |
522 | ;; system whose `eol-type' property is N. | |
4ed46869 | 523 | |
8057896b | 524 | (defun coding-system-eol-type (coding-system) |
0269ddfb KH |
525 | "Return eol-type of CODING-SYSTEM. |
526 | An eol-type is integer 0, 1, 2, or a vector of coding systems. | |
4ed46869 | 527 | |
0269ddfb KH |
528 | Integer values 0, 1, and 2 indicate a format of end-of-line; LF, |
529 | CRLF, and CR respectively. | |
530 | ||
531 | A vector value indicates that a format of end-of-line should be | |
532 | detected automatically. Nth element of the vector is the subsidiary | |
533 | coding system whose eol-type is N." | |
534 | (get coding-system 'eol-type)) | |
bd882697 | 535 | |
857ea15c AS |
536 | (defun coding-system-lessp (x y) |
537 | (cond ((eq x 'no-conversion) t) | |
538 | ((eq y 'no-conversion) nil) | |
539 | ((eq x 'emacs-mule) t) | |
540 | ((eq y 'emacs-mule) nil) | |
541 | ((eq x 'undecided) t) | |
542 | ((eq y 'undecided) nil) | |
543 | (t (let ((c1 (coding-system-mnemonic x)) | |
544 | (c2 (coding-system-mnemonic y))) | |
545 | (or (< (downcase c1) (downcase c2)) | |
546 | (and (not (> (downcase c1) (downcase c2))) | |
547 | (< c1 c2))))))) | |
548 | ||
549 | ;; Add CODING-SYSTEM to coding-system-list while keeping it sorted. | |
550 | (defun add-to-coding-system-list (coding-system) | |
551 | (if (or (null coding-system-list) | |
552 | (coding-system-lessp coding-system (car coding-system-list))) | |
553 | (setq coding-system-list (cons coding-system coding-system-list)) | |
554 | (let ((len (length coding-system-list)) | |
555 | mid (tem coding-system-list)) | |
556 | (while (> len 1) | |
557 | (setq mid (nthcdr (/ len 2) tem)) | |
558 | (if (coding-system-lessp (car mid) coding-system) | |
559 | (setq tem mid | |
560 | len (- len (/ len 2))) | |
561 | (setq len (/ len 2)))) | |
562 | (setcdr tem (cons coding-system (cdr tem)))))) | |
563 | ||
80a7463d | 564 | (defun coding-system-list (&optional base-only) |
c11a8f77 KH |
565 | "Return a list of all existing non-subsidiary coding systems. |
566 | If optional arg BASE-ONLY is non-nil, only base coding systems are listed. | |
567 | The value doesn't include subsidiary coding systems which are what | |
568 | made from bases and aliases automatically for various end-of-line | |
569 | formats (e.g. iso-latin-1-unix, koi8-r-dos)." | |
80a7463d KH |
570 | (let* ((codings (copy-sequence coding-system-list)) |
571 | (tail (cons nil codings))) | |
572 | ;; Remove subsidiary coding systems (eol variants) and alias | |
573 | ;; coding systems (if necessary). | |
574 | (while (cdr tail) | |
575 | (let* ((coding (car (cdr tail))) | |
576 | (aliases (coding-system-get coding 'alias-coding-systems))) | |
577 | (if (or | |
578 | ;; CODING is an eol variant if not in ALIASES. | |
579 | (not (memq coding aliases)) | |
580 | ;; CODING is an alias if it is not car of ALIASES. | |
581 | (and base-only (not (eq coding (car aliases))))) | |
582 | (setcdr tail (cdr (cdr tail))) | |
583 | (setq tail (cdr tail))))) | |
584 | codings)) | |
585 | ||
c11a8f77 KH |
586 | (defun register-char-codings (coding-system safe-chars) |
587 | (let ((general (char-table-extra-slot char-coding-system-table 0))) | |
588 | (if (eq safe-chars t) | |
589 | (or (memq coding-system general) | |
590 | (set-char-table-extra-slot char-coding-system-table 0 | |
591 | (cons coding-system general))) | |
592 | (map-char-table | |
593 | (function | |
594 | (lambda (key val) | |
595 | (if (and (>= key 128) val) | |
596 | (let ((codings (aref char-coding-system-table key))) | |
597 | (or (memq coding-system codings) | |
598 | (aset char-coding-system-table key | |
599 | (cons coding-system codings))))))) | |
600 | safe-chars)))) | |
601 | ||
602 | ||
6e9722b0 KH |
603 | ;; Make subsidiary coding systems (eol-type variants) of CODING-SYSTEM. |
604 | (defun make-subsidiary-coding-system (coding-system) | |
0269ddfb KH |
605 | (let ((coding-spec (coding-system-spec coding-system)) |
606 | (subsidiaries (vector (intern (format "%s-unix" coding-system)) | |
8057896b KH |
607 | (intern (format "%s-dos" coding-system)) |
608 | (intern (format "%s-mac" coding-system)))) | |
0269ddfb KH |
609 | (i 0) |
610 | temp) | |
8057896b | 611 | (while (< i 3) |
0269ddfb | 612 | (put (aref subsidiaries i) 'coding-system coding-spec) |
8057896b | 613 | (put (aref subsidiaries i) 'eol-type i) |
857ea15c | 614 | (add-to-coding-system-list (aref subsidiaries i)) |
0269ddfb KH |
615 | (setq coding-system-alist |
616 | (cons (list (symbol-name (aref subsidiaries i))) | |
617 | coding-system-alist)) | |
8057896b KH |
618 | (setq i (1+ i))) |
619 | subsidiaries)) | |
4ed46869 | 620 | |
8057896b | 621 | (defun make-coding-system (coding-system type mnemonic doc-string |
1b46a680 KH |
622 | &optional |
623 | flags | |
624 | properties | |
625 | eol-type) | |
3bb1accb | 626 | "Define a new coding system CODING-SYSTEM (symbol). |
a6acd8a2 | 627 | Remaining arguments are TYPE, MNEMONIC, DOC-STRING, FLAGS (optional), |
d9e3229d | 628 | and PROPERTIES (optional) which construct a coding-spec of CODING-SYSTEM |
a6acd8a2 | 629 | in the following format: |
0269ddfb | 630 | [TYPE MNEMONIC DOC-STRING PLIST FLAGS] |
3bb1accb KH |
631 | |
632 | TYPE is an integer value indicating the type of the coding system as follows: | |
4ed46869 KH |
633 | 0: Emacs internal format, |
634 | 1: Shift-JIS (or MS-Kanji) used mainly on Japanese PC, | |
635 | 2: ISO-2022 including many variants, | |
636 | 3: Big5 used mainly on Chinese PC, | |
cd9c3177 KH |
637 | 4: private, CCL programs provide encoding/decoding algorithm, |
638 | 5: Raw-text, which means that text contains random 8-bit codes. | |
0269ddfb | 639 | |
3bb1accb | 640 | MNEMONIC is a character to be displayed on mode line for the coding system. |
0269ddfb | 641 | |
3bb1accb | 642 | DOC-STRING is a documentation string for the coding system. |
0269ddfb | 643 | |
3bb1accb | 644 | FLAGS specifies more detailed information of the coding system as follows: |
d6d6d592 | 645 | |
3bb1accb | 646 | If TYPE is 2 (ISO-2022), FLAGS is a list of these elements: |
4ed46869 KH |
647 | CHARSET0, CHARSET1, CHARSET2, CHARSET3, SHORT-FORM, |
648 | ASCII-EOL, ASCII-CNTL, SEVEN, LOCKING-SHIFT, SINGLE-SHIFT, | |
d6d6d592 | 649 | USE-ROMAN, USE-OLDJIS, NO-ISO6429, INIT-BOL, DESIGNATION-BOL, |
850101ed | 650 | SAFE, ACCEPT-LATIN-EXTRA-CODE. |
4ed46869 KH |
651 | CHARSETn are character sets initially designated to Gn graphic registers. |
652 | If CHARSETn is nil, Gn is never used. | |
653 | If CHARSETn is t, Gn can be used but nothing designated initially. | |
654 | If CHARSETn is a list of character sets, those character sets are | |
655 | designated to Gn on output, but nothing designated to Gn initially. | |
421e3b4e | 656 | But, character set `ascii' can be designated only to G0. |
4ed46869 KH |
657 | SHORT-FORM non-nil means use short designation sequence on output. |
658 | ASCII-EOL non-nil means designate ASCII to g0 at end of line on output. | |
659 | ASCII-CNTL non-nil means designate ASCII to g0 before control codes and | |
660 | SPACE on output. | |
661 | SEVEN non-nil means use 7-bit code only on output. | |
662 | LOCKING-SHIFT non-nil means use locking-shift. | |
663 | SINGLE-SHIFT non-nil means use single-shift. | |
664 | USE-ROMAN non-nil means designate JIS0201-1976-Roman instead of ASCII. | |
665 | USE-OLDJIS non-nil means designate JIS0208-1976 instead of JIS0208-1983. | |
666 | NO-ISO6429 non-nil means not use ISO6429's direction specification. | |
69eba008 KH |
667 | INIT-BOL non-nil means any designation state is assumed to be reset |
668 | to initial at each beginning of line on output. | |
669 | DESIGNATION-BOL non-nil means designation sequences should be placed | |
670 | at beginning of line on output. | |
a6acd8a2 | 671 | SAFE non-nil means convert unsafe characters to `?' on output. |
c11a8f77 KH |
672 | Characters not specified in the property `safe-charsets' nor |
673 | `safe-chars' are unsafe. | |
850101ed RS |
674 | ACCEPT-LATIN-EXTRA-CODE non-nil means code-detection routine accepts |
675 | a code specified in `latin-extra-code-table' (which see) as a valid | |
676 | code of the coding system. | |
d6d6d592 | 677 | |
3bb1accb KH |
678 | If TYPE is 4 (private), FLAGS should be a cons of CCL programs, for |
679 | decoding and encoding. CCL programs should be specified by their | |
680 | symbols. | |
d9e3229d KH |
681 | |
682 | PROPERTIES is an alist of properties vs the corresponding values. | |
683 | These properties are set in PLIST, a property list. This function | |
684 | also sets properties `coding-category' and `alias-coding-systems' | |
685 | automatically. | |
4ed46869 | 686 | |
1b46a680 KH |
687 | EOL-TYPE specifies the EOL type of the coding-system in one of the |
688 | following formats: | |
689 | ||
690 | o symbol (unix, dos, or mac) | |
691 | ||
692 | The symbol `unix' means Unix-like EOL (LF), `dos' means | |
693 | DOS-like EOL (CRLF), and `mac' means MAC-like EOL (CR). | |
694 | ||
695 | o number (0, 1, or 2) | |
696 | ||
697 | The number 0, 1, and 2 mean UNIX, DOS, and MAC-like EOL | |
698 | respectively. | |
699 | ||
700 | o vector of coding-systems of length 3 | |
701 | ||
702 | The EOL type is detected automatically for the coding system. | |
703 | And, according to the detected EOL type, one of the coding | |
704 | systems in the vector is selected. Elements of the vector | |
4d3a9228 | 705 | corresponds to Unix-like EOL, DOS-like EOL, and Mac-like EOL |
1b46a680 KH |
706 | in this order. |
707 | ||
3bb1accb KH |
708 | Kludgy features for backward compatibility: |
709 | ||
710 | 1. If TYPE is 4 and car or cdr of FLAGS is a vector, the vector is | |
711 | treated as a compiled CCL code. | |
712 | ||
713 | 2. If PROPERTIES is just a list of character sets, the list is set as | |
714 | a value of `safe-charsets' in PLIST." | |
0269ddfb | 715 | ;; Set a value of `coding-system' property. |
6e9722b0 | 716 | (let ((coding-spec (make-vector 5 nil)) |
d9e3229d KH |
717 | (no-initial-designation t) |
718 | (no-alternative-designation t) | |
6e9722b0 | 719 | coding-category) |
cd9c3177 | 720 | (if (or (not (integerp type)) (< type 0) (> type 5)) |
0269ddfb | 721 | (error "TYPE argument must be 0..5")) |
8057896b | 722 | (if (or (not (integerp mnemonic)) (<= mnemonic ? ) (> mnemonic 127)) |
e8dd0160 | 723 | (error "MNEMONIC argument must be an ASCII printable character.")) |
0269ddfb KH |
724 | (aset coding-spec coding-spec-type-idx type) |
725 | (aset coding-spec coding-spec-mnemonic-idx mnemonic) | |
726 | (aset coding-spec coding-spec-doc-string-idx | |
e127a722 | 727 | (purecopy (if (stringp doc-string) doc-string ""))) |
6e9722b0 KH |
728 | (cond ((= type 0) |
729 | (setq coding-category 'coding-category-emacs-mule)) | |
730 | ((= type 1) | |
731 | (setq coding-category 'coding-category-sjis)) | |
732 | ((= type 2) ; ISO2022 | |
4ed46869 | 733 | (let ((i 0) |
6e9722b0 | 734 | (vec (make-vector 32 nil)) |
05134257 KH |
735 | (g1-designation nil) |
736 | (fl flags)) | |
4ed46869 | 737 | (while (< i 4) |
05134257 | 738 | (let ((charset (car fl))) |
6e9722b0 KH |
739 | (if (and no-initial-designation |
740 | (> i 0) | |
741 | (or (charsetp charset) | |
742 | (and (consp charset) | |
743 | (charsetp (car charset))))) | |
744 | (setq no-initial-designation nil)) | |
745 | (if (charsetp charset) | |
746 | (if (= i 1) (setq g1-designation charset)) | |
747 | (if (consp charset) | |
748 | (let ((tail charset) | |
749 | elt) | |
750 | (while tail | |
751 | (setq elt (car tail)) | |
d9e3229d KH |
752 | (if (eq elt t) |
753 | (setq no-alternative-designation nil) | |
754 | (if (and elt (not (charsetp elt))) | |
755 | (error "Invalid charset: %s" elt))) | |
6e9722b0 KH |
756 | (setq tail (cdr tail))) |
757 | (setq g1-designation (car charset))) | |
d9e3229d KH |
758 | (if charset |
759 | (if (eq charset t) | |
760 | (setq no-alternative-designation nil) | |
761 | (error "Invalid charset: %s" charset))))) | |
4ed46869 | 762 | (aset vec i charset)) |
05134257 KH |
763 | (setq fl (cdr fl) i (1+ i))) |
764 | (while (and (< i 32) fl) | |
765 | (aset vec i (car fl)) | |
766 | (setq fl (cdr fl) i (1+ i))) | |
6e9722b0 | 767 | (aset coding-spec 4 vec) |
6e9722b0 KH |
768 | (setq coding-category |
769 | (if (aref vec 8) ; Use locking-shift. | |
dc64cd19 KH |
770 | (or (and (aref vec 7) 'coding-category-iso-7-else) |
771 | 'coding-category-iso-8-else) | |
6e9722b0 KH |
772 | (if (aref vec 7) ; 7-bit only. |
773 | (if (aref vec 9) ; Use single-shift. | |
dc64cd19 | 774 | 'coding-category-iso-7-else |
d9e3229d KH |
775 | (if no-alternative-designation |
776 | 'coding-category-iso-7-tight | |
777 | 'coding-category-iso-7)) | |
3ad911d8 KH |
778 | (if (or no-initial-designation |
779 | (not no-alternative-designation)) | |
dc64cd19 | 780 | 'coding-category-iso-8-else |
6e9722b0 KH |
781 | (if (and (charsetp g1-designation) |
782 | (= (charset-dimension g1-designation) 2)) | |
783 | 'coding-category-iso-8-2 | |
784 | 'coding-category-iso-8-1))))))) | |
785 | ((= type 3) | |
786 | (setq coding-category 'coding-category-big5)) | |
787 | ((= type 4) ; private | |
c76b5c99 | 788 | (setq coding-category 'coding-category-ccl) |
3bb1accb KH |
789 | (if (not (consp flags)) |
790 | (error "Invalid FLAGS argument for TYPE 4 (CCL)") | |
791 | (let ((decoder (check-ccl-program | |
792 | (car flags) | |
793 | (intern (format "%s-decoder" coding-system)))) | |
794 | (encoder (check-ccl-program | |
795 | (cdr flags) | |
796 | (intern (format "%s-encoder" coding-system))))) | |
797 | (if (and decoder encoder) | |
798 | (aset coding-spec 4 (cons decoder encoder)) | |
799 | (error "Invalid FLAGS argument for TYPE 4 (CCL)"))))) | |
cd9c3177 KH |
800 | (t ; i.e. (= type 5) |
801 | (setq coding-category 'coding-category-raw-text))) | |
0269ddfb KH |
802 | |
803 | (let ((plist (list 'coding-category coding-category | |
d9e3229d | 804 | 'alias-coding-systems (list coding-system)))) |
0269ddfb | 805 | (if no-initial-designation |
d9e3229d KH |
806 | (plist-put plist 'no-initial-designation t)) |
807 | (if (and properties | |
808 | (or (eq properties t) | |
809 | (not (consp (car properties))))) | |
810 | ;; In the old version, the arg PROPERTIES is a list to be | |
811 | ;; set in PLIST as a value of property `safe-charsets'. | |
c11a8f77 KH |
812 | (setq properties (list (cons 'safe-charsets properties)))) |
813 | ;; In the current version PROPERTIES is a property list. | |
814 | ;; Reflect it into PLIST one by one while handling safe-chars | |
815 | ;; specially. | |
816 | (let ((safe-charsets (cdr (assq 'safe-charsets properties))) | |
817 | (safe-chars (cdr (assq 'safe-chars properties))) | |
818 | (l properties) | |
819 | prop val) | |
820 | ;; If only safe-charsets is specified, make a char-table from | |
821 | ;; it, and store that char-table as the value of `safe-chars'. | |
822 | (if (and (not safe-chars) safe-charsets) | |
823 | (let (charset) | |
824 | (if (eq safe-charsets t) | |
825 | (setq safe-chars t) | |
826 | (setq safe-chars (make-char-table 'safe-chars)) | |
827 | (while safe-charsets | |
828 | (setq charset (car safe-charsets) | |
829 | safe-charsets (cdr safe-charsets)) | |
830 | (cond ((eq charset 'ascii)) ; just ignore | |
831 | ((eq charset 'eight-bit-control) | |
832 | (let ((i 128)) | |
833 | (while (< i 160) | |
834 | (aset safe-chars i t) | |
835 | (setq i (1+ i))))) | |
836 | ((eq charset 'eight-bit-graphic) | |
837 | (let ((i 160)) | |
838 | (while (< i 256) | |
839 | (aset safe-chars i t) | |
840 | (setq i (1+ i))))) | |
841 | (t | |
842 | (aset safe-chars (make-char charset) t))))) | |
843 | (setq l (cons (cons 'safe-chars safe-chars) l)))) | |
844 | (while l | |
845 | (setq prop (car (car l)) val (cdr (car l)) l (cdr l)) | |
846 | (if (eq prop 'safe-chars) | |
847 | (progn | |
cc57cc54 KH |
848 | (if (and (symbolp val) |
849 | (get val 'translation-table)) | |
850 | (setq safe-chars (get val 'translation-table))) | |
851 | (register-char-codings coding-system safe-chars) | |
852 | (setq val safe-chars))) | |
c11a8f77 | 853 | (plist-put plist prop val))) |
1b46a680 KH |
854 | ;; The property `coding-category' may have been set differently |
855 | ;; through PROPERTIES. | |
856 | (setq coding-category (plist-get plist 'coding-category)) | |
0269ddfb | 857 | (aset coding-spec coding-spec-plist-idx plist)) |
6e9722b0 | 858 | (put coding-system 'coding-system coding-spec) |
6e9722b0 KH |
859 | (put coding-category 'coding-systems |
860 | (cons coding-system (get coding-category 'coding-systems)))) | |
4ed46869 | 861 | |
1b46a680 | 862 | ;; Next, set a value of `eol-type' property. |
af873cd1 | 863 | (if (not eol-type) |
1b46a680 KH |
864 | ;; If EOL-TYPE is nil, set a vector of subsidiary coding |
865 | ;; systems, each corresponds to a coding system for the detected | |
866 | ;; EOL format. | |
867 | (setq eol-type (make-subsidiary-coding-system coding-system))) | |
868 | (setq eol-type | |
869 | (cond ((or (eq eol-type 'unix) (null eol-type)) | |
870 | 0) | |
871 | ((eq eol-type 'dos) | |
872 | 1) | |
873 | ((eq eol-type 'mac) | |
874 | 2) | |
875 | ((or (and (vectorp eol-type) | |
876 | (= (length eol-type) 3)) | |
877 | (and (numberp eol-type) | |
878 | (and (>= eol-type 0) | |
879 | (<= eol-type 2)))) | |
880 | eol-type) | |
881 | (t | |
882 | (error "Invalid EOL-TYPE spec:%S" eol-type)))) | |
883 | (put coding-system 'eol-type eol-type) | |
0269ddfb KH |
884 | |
885 | ;; At last, register CODING-SYSTEM in `coding-system-list' and | |
886 | ;; `coding-system-alist'. | |
857ea15c | 887 | (add-to-coding-system-list coding-system) |
0269ddfb | 888 | (setq coding-system-alist (cons (list (symbol-name coding-system)) |
d9e3229d | 889 | coding-system-alist)) |
05134257 KH |
890 | |
891 | ;; For a coding system of cateogory iso-8-1 and iso-8-2, create | |
892 | ;; XXX-with-esc variants. | |
893 | (let ((coding-category (coding-system-category coding-system))) | |
894 | (if (or (eq coding-category 'coding-category-iso-8-1) | |
895 | (eq coding-category 'coding-category-iso-8-2)) | |
896 | (let ((esc (intern (concat (symbol-name coding-system) "-with-esc"))) | |
c11a8f77 KH |
897 | (doc (format "Same as %s but can handle any charsets by ISO's escape sequences." coding-system)) |
898 | (safe-charsets (assq 'safe-charsets properties)) | |
899 | (mime-charset (assq 'mime-charset properties))) | |
900 | (if safe-charsets | |
901 | (setcdr safe-charsets t) | |
902 | (setq properties (cons (cons 'safe-charsets t) properties))) | |
903 | (if mime-charset | |
904 | (setcdr mime-charset nil)) | |
05134257 KH |
905 | (make-coding-system esc type mnemonic doc |
906 | (if (listp (car flags)) | |
907 | (cons (append (car flags) '(t)) (cdr flags)) | |
908 | (cons (list (car flags) t) (cdr flags))) | |
c11a8f77 | 909 | properties)))) |
05134257 | 910 | |
d9e3229d | 911 | coding-system) |
8057896b | 912 | |
a42763dc | 913 | (defun define-coding-system-alias (alias coding-system) |
358d28fb | 914 | "Define ALIAS as an alias for coding system CODING-SYSTEM." |
0269ddfb KH |
915 | (put alias 'coding-system (coding-system-spec coding-system)) |
916 | (nconc (coding-system-get alias 'alias-coding-systems) (list alias)) | |
857ea15c | 917 | (add-to-coding-system-list alias) |
0269ddfb KH |
918 | (setq coding-system-alist (cons (list (symbol-name alias)) |
919 | coding-system-alist)) | |
920 | (let ((eol-type (coding-system-eol-type coding-system))) | |
921 | (if (vectorp eol-type) | |
922 | (put alias 'eol-type (make-subsidiary-coding-system alias)) | |
923 | (put alias 'eol-type eol-type)))) | |
4ed46869 KH |
924 | |
925 | (defun set-buffer-file-coding-system (coding-system &optional force) | |
358d28fb RS |
926 | "Set the file coding-system of the current buffer to CODING-SYSTEM. |
927 | This means that when you save the buffer, it will be converted | |
928 | according to CODING-SYSTEM. For a list of possible values of CODING-SYSTEM, | |
929 | use \\[list-coding-systems]. | |
930 | ||
931 | If the buffer's previous file coding-system value specifies end-of-line | |
932 | conversion, and CODING-SYSTEM does not specify one, CODING-SYSTEM is | |
933 | merged with the already-specified end-of-line conversion. | |
b839fdcc KH |
934 | |
935 | If the buffer's previous file coding-system value specifies text | |
936 | conversion, and CODING-SYSTEM does not specify one, CODING-SYSTEM is | |
937 | merged with the already-specified text conversion. | |
938 | ||
939 | However, if the optional prefix argument FORCE is non-nil, then | |
940 | CODING-SYSTEM is used exactly as specified. | |
aeef8f07 KH |
941 | |
942 | This marks the buffer modified so that the succeeding \\[save-buffer] | |
943 | surely saves the buffer with CODING-SYSTEM. From a program, if you | |
944 | don't want to mark the buffer modified, just set the variable | |
945 | `buffer-file-coding-system' directly." | |
d9e3229d | 946 | (interactive "zCoding system for visited file (default, nil): \nP") |
4ed46869 | 947 | (check-coding-system coding-system) |
36d455c4 | 948 | (if (and coding-system buffer-file-coding-system (null force)) |
8dd735c1 KH |
949 | (let ((base (coding-system-base buffer-file-coding-system)) |
950 | (eol (coding-system-eol-type buffer-file-coding-system))) | |
951 | ;; If CODING-SYSTEM doesn't specify text conversion, merge | |
952 | ;; with that of buffer-file-coding-system. | |
953 | (if (eq (coding-system-base coding-system) 'undecided) | |
954 | (setq coding-system (coding-system-change-text-conversion | |
955 | coding-system base))) | |
956 | ;; If CODING-SYSTEM doesn't specify eol conversion, merge with | |
957 | ;; that of buffer-file-coding-system. | |
958 | (if (and (vectorp (coding-system-eol-type coding-system)) | |
959 | (numberp eol) (>= eol 0) (<= eol 2)) | |
960 | (setq coding-system (coding-system-change-eol-conversion | |
961 | coding-system eol))))) | |
4ed46869 KH |
962 | (setq buffer-file-coding-system coding-system) |
963 | (set-buffer-modified-p t) | |
964 | (force-mode-line-update)) | |
965 | ||
358d28fb RS |
966 | (defvar default-terminal-coding-system nil |
967 | "Default value for the terminal coding system. | |
968 | This is normally set according to the selected language environment. | |
969 | See also the command `set-terminal-coding-system'.") | |
970 | ||
df100398 KH |
971 | (defun set-terminal-coding-system (coding-system) |
972 | "Set coding system of your terminal to CODING-SYSTEM. | |
358d28fb RS |
973 | All text output to the terminal will be encoded |
974 | with the specified coding system. | |
975 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. | |
976 | The default is determined by the selected language environment | |
977 | or by the previous use of this command." | |
978 | (interactive | |
2e02a76f RS |
979 | (list (let ((default (if (and (not (terminal-coding-system)) |
980 | default-terminal-coding-system) | |
981 | default-terminal-coding-system))) | |
982 | (read-coding-system | |
983 | (format "Coding system for terminal display (default, %s): " | |
984 | default) | |
985 | default)))) | |
358d28fb RS |
986 | (if (and (not coding-system) |
987 | (not (terminal-coding-system))) | |
988 | (setq coding-system default-terminal-coding-system)) | |
989 | (if coding-system | |
990 | (setq default-terminal-coding-system coding-system)) | |
df100398 KH |
991 | (set-terminal-coding-system-internal coding-system) |
992 | (redraw-frame (selected-frame))) | |
993 | ||
358d28fb RS |
994 | (defvar default-keyboard-coding-system nil |
995 | "Default value of the keyboard coding system. | |
996 | This is normally set according to the selected language environment. | |
997 | See also the command `set-keyboard-coding-system'.") | |
998 | ||
df100398 | 999 | (defun set-keyboard-coding-system (coding-system) |
358d28fb RS |
1000 | "Set coding system for keyboard input to CODING-SYSTEM. |
1001 | In addition, this command enables Encoded-kbd minor mode. | |
6d34f495 DL |
1002 | \(If CODING-SYSTEM is nil, Encoded-kbd mode is turned off -- see |
1003 | `encoded-kbd-mode'.) | |
358d28fb RS |
1004 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]. |
1005 | The default is determined by the selected language environment | |
1006 | or by the previous use of this command." | |
1007 | (interactive | |
2e02a76f RS |
1008 | (list (let ((default (if (and (not (keyboard-coding-system)) |
1009 | default-keyboard-coding-system) | |
1010 | default-keyboard-coding-system))) | |
1011 | (read-coding-system | |
1012 | (format "Coding system for keyboard input (default, %s): " | |
1013 | default) | |
1014 | default)))) | |
358d28fb RS |
1015 | (if (and (not coding-system) |
1016 | (not (keyboard-coding-system))) | |
1017 | (setq coding-system default-keyboard-coding-system)) | |
1018 | (if coding-system | |
1019 | (setq default-keyboard-coding-system coding-system)) | |
df100398 KH |
1020 | (set-keyboard-coding-system-internal coding-system) |
1021 | (encoded-kbd-mode (if coding-system 1 0))) | |
1022 | ||
6d34f495 DL |
1023 | (defcustom keyboard-coding-system nil |
1024 | "Specify coding system for keyboard input. | |
1025 | If you set this on a terminal which can't distinguish Meta keys from | |
1026 | 8-bit characters, you will have to use ESC to type Meta characters. | |
1027 | See Info node `Specify Coding' and Info node `Single-Byte Character Support'. | |
1028 | ||
1029 | Setting this variable directly does not take effect; | |
1030 | use either M-x customize or \\[set-keyboard-coding-system]." | |
1031 | :type '(coding-system :tag "Coding system") | |
1032 | :link '(info-link "(emacs)Specify Coding") | |
1033 | :link '(info-link "(emacs)Single-Byte Character Support") | |
1034 | :set (lambda (symbol value) | |
1035 | ;; Don't load encoded-kbd-mode unnecessarily. | |
1036 | (if (or value (boundp 'encoded-kbd-mode)) | |
1037 | (set-keyboard-coding-system value) | |
1038 | (set-default 'keyboard-coding-system nil))) ; must initialize | |
1039 | :version "21.1" | |
1040 | :group 'keyboard | |
1041 | :group 'mule) | |
1042 | ||
df100398 | 1043 | (defun set-buffer-process-coding-system (decoding encoding) |
358d28fb | 1044 | "Set coding systems for the process associated with the current buffer. |
df100398 | 1045 | DECODING is the coding system to be used to decode input from the process, |
358d28fb RS |
1046 | ENCODING is the coding system to be used to encode output to the process. |
1047 | ||
1048 | For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]." | |
4ed46869 | 1049 | (interactive |
83911021 | 1050 | "zCoding-system for output from the process: \nzCoding-system for input to the process: ") |
4ed46869 KH |
1051 | (let ((proc (get-buffer-process (current-buffer)))) |
1052 | (if (null proc) | |
1053 | (error "no process") | |
df100398 KH |
1054 | (check-coding-system decoding) |
1055 | (check-coding-system encoding) | |
1056 | (set-process-coding-system proc decoding encoding))) | |
4ed46869 KH |
1057 | (force-mode-line-update)) |
1058 | ||
d0b99881 RS |
1059 | (defalias 'set-clipboard-coding-system 'set-selection-coding-system) |
1060 | ||
14915c37 | 1061 | (defun set-selection-coding-system (coding-system) |
b25eef20 KH |
1062 | "Make CODING-SYSTEM used for communicating with other X clients . |
1063 | When sending or receiving text via cut_buffer, selection, and clipboard, | |
1064 | the text is encoded or decoded by CODING-SYSTEM." | |
a03b3ce1 | 1065 | (interactive "zCoding system for X selection: ") |
b25eef20 | 1066 | (check-coding-system coding-system) |
14915c37 | 1067 | (setq selection-coding-system coding-system)) |
b25eef20 | 1068 | |
e8dd0160 | 1069 | ;; Coding system lastly specified by the command |
a03b3ce1 KH |
1070 | ;; set-next-selection-coding-system. |
1071 | (defvar last-next-selection-coding-system nil) | |
1072 | ||
1073 | (defun set-next-selection-coding-system (coding-system) | |
1074 | "Make CODING-SYSTEM used for the next communication with other X clients. | |
1075 | This setting is effective for the next communication only." | |
1076 | (interactive | |
1077 | (list (read-coding-system | |
1078 | (if last-next-selection-coding-system | |
1079 | (format "Coding system for the next X selection (default, %S): " | |
1080 | last-next-selection-coding-system) | |
1081 | "Coding system for the next X selection: ") | |
1082 | last-next-selection-coding-system))) | |
1083 | (if coding-system | |
1084 | (setq last-next-selection-coding-system coding-system) | |
1085 | (setq coding-system last-next-selection-coding-system)) | |
1086 | (check-coding-system coding-system) | |
1087 | ||
1088 | (setq next-selection-coding-system coding-system)) | |
1089 | ||
4ed46869 | 1090 | (defun set-coding-priority (arg) |
d9e3229d KH |
1091 | "Set priority of coding categories according to LIST. |
1092 | LIST is a list of coding categories ordered by priority." | |
1093 | (let ((l arg) | |
1094 | (current-list (copy-sequence coding-category-list))) | |
e8dd0160 | 1095 | ;; Check the validity of ARG while deleting coding categories in |
d9e3229d KH |
1096 | ;; ARG from CURRENT-LIST. We assume that CODING-CATEGORY-LIST |
1097 | ;; contains all coding categories. | |
1098 | (while l | |
1099 | (if (or (null (get (car l) 'coding-category-index)) | |
1100 | (null (memq (car l) current-list))) | |
1101 | (error "Invalid or duplicated element in argument: %s" arg)) | |
1102 | (setq current-list (delq (car l) current-list)) | |
1103 | (setq l (cdr l))) | |
4ed46869 | 1104 | ;; Update `coding-category-list' and return it. |
2feaf204 KH |
1105 | (setq coding-category-list (append arg current-list)) |
1106 | (set-coding-priority-internal))) | |
4ed46869 KH |
1107 | |
1108 | ;;; FILE I/O | |
1109 | ||
e76938e7 | 1110 | (defcustom auto-coding-alist |
0735296c DL |
1111 | '(("\\.\\(arc\\|zip\\|lzh\\|zoo\\|jar\\|tar\\|tgz\\)\\'" . no-conversion) |
1112 | ("\\.\\(ARC\\|ZIP\\|LZH\\|ZOO\\|JAR\\|TAR\\|TGZ\\)\\'" . no-conversion)) | |
835f49b8 KH |
1113 | "Alist of filename patterns vs corresponding coding systems. |
1114 | Each element looks like (REGEXP . CODING-SYSTEM). | |
558b0c86 | 1115 | A file whose name matches REGEXP is decoded by CODING-SYSTEM on reading. |
835f49b8 | 1116 | |
7fed493a RS |
1117 | The settings in this alist take priority over `coding:' tags |
1118 | in the file (see the function `set-auto-coding') | |
e76938e7 DL |
1119 | and the contents of `file-coding-system-alist'." |
1120 | :group 'files | |
1121 | :group 'mule | |
1122 | :type '(repeat (cons (regexp :tag "File name regexp") | |
1123 | (symbol :tag "Coding system")))) | |
835f49b8 | 1124 | |
1c4cc63a KH |
1125 | (defvar set-auto-coding-for-load nil |
1126 | "Non-nil means look for `load-coding' property instead of `coding'. | |
1127 | This is used for loading and byte-compiling Emacs Lisp files.") | |
1128 | ||
8a592131 RS |
1129 | (defun auto-coding-alist-lookup (filename) |
1130 | "Return the coding system specified by `auto-coding-alist' for FILENAME." | |
1131 | (let ((alist auto-coding-alist) | |
ca128d75 | 1132 | (case-fold-search (memq system-type '(vax-vms windows-nt ms-dos))) |
8a592131 RS |
1133 | coding-system) |
1134 | (while (and alist (not coding-system)) | |
1135 | (if (string-match (car (car alist)) filename) | |
1136 | (setq coding-system (cdr (car alist))) | |
1137 | (setq alist (cdr alist)))) | |
1138 | coding-system)) | |
1139 | ||
835f49b8 KH |
1140 | (defun set-auto-coding (filename size) |
1141 | "Return coding system for a file FILENAME of which SIZE bytes follow point. | |
1c4cc63a KH |
1142 | These bytes should include at least the first 1k of the file |
1143 | and the last 3k of the file, but the middle may be omitted. | |
63561304 | 1144 | |
835f49b8 KH |
1145 | It checks FILENAME against the variable `auto-coding-alist'. |
1146 | If FILENAME doesn't match any entries in the variable, | |
1147 | it checks for a `coding:' tag in the first one or two lines following | |
e8dd0160 | 1148 | point. If no `coding:' tag is found, it checks for local variables |
1c4cc63a | 1149 | list in the last 3K bytes out of the SIZE bytes. |
63561304 KH |
1150 | |
1151 | The return value is the specified coding system, | |
1152 | or nil if nothing specified. | |
87aba788 | 1153 | |
ba74e833 | 1154 | The variable `set-auto-coding-function' (which see) is set to this |
87aba788 | 1155 | function by default." |
8a592131 | 1156 | (let ((coding-system (auto-coding-alist-lookup filename))) |
835f49b8 KH |
1157 | |
1158 | (or coding-system | |
1159 | (let* ((case-fold-search t) | |
1160 | (head-start (point)) | |
1161 | (head-end (+ head-start (min size 1024))) | |
1162 | (tail-start (+ head-start (max (- size 3072) 0))) | |
1163 | (tail-end (+ head-start size)) | |
1164 | coding-system head-found tail-found pos) | |
1165 | ;; Try a short cut by searching for the string "coding:" | |
e8dd0160 | 1166 | ;; and for "unibyte:" at the head and tail of SIZE bytes. |
835f49b8 KH |
1167 | (setq head-found (or (search-forward "coding:" head-end t) |
1168 | (search-forward "unibyte:" head-end t))) | |
1169 | (if (and head-found (> head-found tail-start)) | |
1170 | ;; Head and tail are overlapped. | |
1171 | (setq tail-found head-found) | |
1172 | (goto-char tail-start) | |
1173 | (setq tail-found (or (search-forward "coding:" tail-end t) | |
1174 | (search-forward "unibyte:" tail-end t)))) | |
1175 | ||
1176 | ;; At first check the head. | |
1177 | (when head-found | |
1178 | (goto-char head-start) | |
1179 | (setq pos (re-search-forward "[\n\r]" head-end t)) | |
1180 | (if (and pos | |
1181 | (= (char-after head-start) ?#) | |
1182 | (= (char-after (1+ head-start)) ?!)) | |
1183 | ;; If the file begins with "#!" (exec interpreter magic), | |
1184 | ;; look for coding frobs in the first two lines. You cannot | |
1185 | ;; necessarily put them in the first line of such a file | |
1186 | ;; without screwing up the interpreter invocation. | |
1187 | (setq pos (search-forward "\n" head-end t))) | |
1188 | (if pos (setq head-end pos)) | |
1189 | (when (< head-found head-end) | |
1190 | (goto-char head-start) | |
1191 | (when (and set-auto-coding-for-load | |
1192 | (re-search-forward | |
1193 | "-\\*-\\(.*;\\)?[ \t]*unibyte:[ \t]*\\([^ ;]+\\)" | |
1194 | head-end t)) | |
1195 | (setq coding-system 'raw-text)) | |
1196 | (when (and (not coding-system) | |
1197 | (re-search-forward | |
1198 | "-\\*-\\(.*;\\)?[ \t]*coding:[ \t]*\\([^ ;]+\\)" | |
1199 | head-end t)) | |
1200 | (setq coding-system (intern (match-string 2))) | |
1201 | (or (coding-system-p coding-system) | |
1202 | (setq coding-system nil))))) | |
1203 | ||
1204 | ;; If no coding: tag in the head, check the tail. | |
1205 | (when (and tail-found (not coding-system)) | |
1206 | (goto-char tail-start) | |
1207 | (search-forward "\n\^L" nil t) | |
1208 | (if (re-search-forward | |
1209 | "^\\(.*\\)[ \t]*Local Variables:[ \t]*\\(.*\\)$" tail-end t) | |
1210 | ;; The prefix is what comes before "local variables:" in its | |
1211 | ;; line. The suffix is what comes after "local variables:" | |
1212 | ;; in its line. | |
1213 | (let* ((prefix (regexp-quote (match-string 1))) | |
1214 | (suffix (regexp-quote (match-string 2))) | |
d318fcf4 EZ |
1215 | (re-coding |
1216 | (concat | |
1217 | "^" prefix | |
1218 | "[ \t]*coding[ \t]*:[ \t]*\\([^ \t]+\\)[ \t]*" | |
1219 | suffix "$")) | |
1220 | (re-unibyte | |
1221 | (concat | |
1222 | "^" prefix | |
1223 | "[ \t]*unibyte[ \t]*:[ \t]*\\([^ \t]+\\)[ \t]*" | |
1224 | suffix "$")) | |
1225 | (re-end | |
1226 | (concat "^" prefix "[ \t]*end *:[ \t]*" suffix "$")) | |
835f49b8 KH |
1227 | (pos (point))) |
1228 | (re-search-forward re-end tail-end 'move) | |
1229 | (setq tail-end (point)) | |
1230 | (goto-char pos) | |
1231 | (when (and set-auto-coding-for-load | |
1232 | (re-search-forward re-unibyte tail-end t)) | |
1233 | (setq coding-system 'raw-text)) | |
1234 | (when (and (not coding-system) | |
1235 | (re-search-forward re-coding tail-end t)) | |
1236 | (setq coding-system (intern (match-string 1))) | |
1237 | (or (coding-system-p coding-system) | |
1238 | (setq coding-system nil)))))) | |
1239 | coding-system)))) | |
63561304 KH |
1240 | |
1241 | (setq set-auto-coding-function 'set-auto-coding) | |
87aba788 | 1242 | |
4ed46869 KH |
1243 | ;; Set buffer-file-coding-system of the current buffer after some text |
1244 | ;; is inserted. | |
1245 | (defun after-insert-file-set-buffer-file-coding-system (inserted) | |
1246 | (if last-coding-system-used | |
1247 | (let ((coding-system | |
1248 | (find-new-buffer-file-coding-system last-coding-system-used)) | |
1249 | (modified-p (buffer-modified-p))) | |
0269ddfb | 1250 | (when coding-system |
71983219 | 1251 | (set-buffer-file-coding-system coding-system t) |
d0c26c63 KH |
1252 | (if (and enable-multibyte-characters |
1253 | (or (eq coding-system 'no-conversion) | |
ccb77b4e | 1254 | (eq (coding-system-type coding-system) 5)) |
136e48e4 KH |
1255 | ;; If buffer was unmodified and the size is the |
1256 | ;; same as INSERTED, we must be visiting it. | |
1257 | (not modified-p) | |
1258 | (= (buffer-size) inserted)) | |
ccb77b4e RS |
1259 | ;; For coding systems no-conversion and raw-text..., |
1260 | ;; edit the buffer as unibyte. | |
d0c26c63 KH |
1261 | (let ((pos-byte (position-bytes (+ (point) inserted)))) |
1262 | (set-buffer-multibyte nil) | |
1263 | (setq inserted (- pos-byte (position-bytes (point)))))) | |
0269ddfb | 1264 | (set-buffer-modified-p modified-p)))) |
d0c26c63 | 1265 | inserted) |
4ed46869 | 1266 | |
84c9d215 KH |
1267 | (add-hook 'after-insert-file-functions |
1268 | 'after-insert-file-set-buffer-file-coding-system) | |
4ed46869 | 1269 | |
8057896b | 1270 | ;; The coding-spec and eol-type of coding-system returned is decided |
4ed46869 KH |
1271 | ;; independently in the following order. |
1272 | ;; 1. That of buffer-file-coding-system locally bound. | |
1273 | ;; 2. That of CODING. | |
1274 | ||
1275 | (defun find-new-buffer-file-coding-system (coding) | |
1276 | "Return a coding system for a buffer when a file of CODING is inserted. | |
a73a8c89 KH |
1277 | The local variable `buffer-file-coding-system' of the current buffer |
1278 | is set to the returned value. | |
509064c5 | 1279 | Return nil if there's no need to set `buffer-file-coding-system'." |
4ed46869 | 1280 | (let (local-coding local-eol |
b685f8d6 | 1281 | found-coding found-eol |
4ed46869 KH |
1282 | new-coding new-eol) |
1283 | (if (null coding) | |
1284 | ;; Nothing found about coding. | |
1285 | nil | |
1286 | ||
b685f8d6 RS |
1287 | ;; Get information of `buffer-file-coding-system' in LOCAL-EOL |
1288 | ;; and LOCAL-CODING. | |
1289 | (setq local-eol (coding-system-eol-type buffer-file-coding-system)) | |
1290 | (if (null (numberp local-eol)) | |
1291 | ;; But eol-type is not yet set. | |
1292 | (setq local-eol nil)) | |
0269ddfb KH |
1293 | (if (and buffer-file-coding-system |
1294 | (not (eq (coding-system-type buffer-file-coding-system) t))) | |
1295 | ;; This is not `undecided'. | |
1296 | (setq local-coding (coding-system-base buffer-file-coding-system))) | |
b685f8d6 RS |
1297 | |
1298 | (if (and (local-variable-p 'buffer-file-coding-system) | |
1299 | local-eol local-coding) | |
4ed46869 KH |
1300 | ;; The current buffer has already set full coding-system, we |
1301 | ;; had better not change it. | |
1302 | nil | |
1303 | ||
8057896b | 1304 | (setq found-eol (coding-system-eol-type coding)) |
4ed46869 | 1305 | (if (null (numberp found-eol)) |
be02cd54 EZ |
1306 | ;; But eol-type is not found. |
1307 | ;; If EOL conversions are inhibited, force unix eol-type. | |
1308 | (setq found-eol (if inhibit-eol-conversion 0))) | |
c76b5c99 KH |
1309 | (if (eq (coding-system-type coding) t) |
1310 | (setq found-coding 'undecided) | |
1311 | (setq found-coding (coding-system-base coding))) | |
1312 | ||
1313 | (if (and (not found-eol) (eq found-coding 'undecided)) | |
1314 | ;; No valid coding information found. | |
1315 | nil | |
1316 | ||
1317 | ;; Some coding information (eol or text) found. | |
1318 | ||
1319 | ;; The local setting takes precedence over the found one. | |
1320 | (setq new-coding (if (local-variable-p 'buffer-file-coding-system) | |
1321 | (or local-coding found-coding) | |
1322 | (or found-coding local-coding))) | |
1323 | (setq new-eol (if (local-variable-p 'buffer-file-coding-system) | |
1324 | (or local-eol found-eol) | |
1325 | (or found-eol local-eol))) | |
1326 | ||
1327 | (let ((eol-type (coding-system-eol-type new-coding))) | |
1328 | (if (and (numberp new-eol) (vectorp eol-type)) | |
1329 | (aref eol-type new-eol) | |
1330 | new-coding))))))) | |
4ed46869 | 1331 | |
fe831d33 GV |
1332 | (defun modify-coding-system-alist (target-type regexp coding-system) |
1333 | "Modify one of look up tables for finding a coding system on I/O operation. | |
8c453b46 RS |
1334 | There are three of such tables, `file-coding-system-alist', |
1335 | `process-coding-system-alist', and `network-coding-system-alist'. | |
fe831d33 GV |
1336 | |
1337 | TARGET-TYPE specifies which of them to modify. | |
8c453b46 RS |
1338 | If it is `file', it affects `file-coding-system-alist' (which see). |
1339 | If it is `process', it affects `process-coding-system-alist' (which see). | |
e8dd0160 | 1340 | If it is `network', it affects `network-coding-system-alist' (which see). |
fe831d33 GV |
1341 | |
1342 | REGEXP is a regular expression matching a target of I/O operation. | |
1343 | The target is a file name if TARGET-TYPE is `file', a program name if | |
1344 | TARGET-TYPE is `process', or a network service name or a port number | |
1345 | to connect to if TARGET-TYPE is `network'. | |
1346 | ||
1347 | CODING-SYSTEM is a coding system to perform code conversion on the I/O | |
8c453b46 RS |
1348 | operation, or a cons cell (DECODING . ENCODING) specifying the coding systems |
1349 | for decoding and encoding respectively, | |
1350 | or a function symbol which, when called, returns such a cons cell." | |
fe831d33 GV |
1351 | (or (memq target-type '(file process network)) |
1352 | (error "Invalid target type: %s" target-type)) | |
1353 | (or (stringp regexp) | |
1354 | (and (eq target-type 'network) (integerp regexp)) | |
1355 | (error "Invalid regular expression: %s" regexp)) | |
1356 | (if (symbolp coding-system) | |
1357 | (if (not (fboundp coding-system)) | |
1358 | (progn | |
1359 | (check-coding-system coding-system) | |
1360 | (setq coding-system (cons coding-system coding-system)))) | |
1361 | (check-coding-system (car coding-system)) | |
1362 | (check-coding-system (cdr coding-system))) | |
1363 | (cond ((eq target-type 'file) | |
1364 | (let ((slot (assoc regexp file-coding-system-alist))) | |
1365 | (if slot | |
1366 | (setcdr slot coding-system) | |
1367 | (setq file-coding-system-alist | |
1368 | (cons (cons regexp coding-system) | |
1369 | file-coding-system-alist))))) | |
1370 | ((eq target-type 'process) | |
1371 | (let ((slot (assoc regexp process-coding-system-alist))) | |
1372 | (if slot | |
1373 | (setcdr slot coding-system) | |
1374 | (setq process-coding-system-alist | |
1375 | (cons (cons regexp coding-system) | |
1376 | process-coding-system-alist))))) | |
1377 | (t | |
1378 | (let ((slot (assoc regexp network-coding-system-alist))) | |
1379 | (if slot | |
1380 | (setcdr slot coding-system) | |
1381 | (setq network-coding-system-alist | |
1382 | (cons (cons regexp coding-system) | |
1383 | network-coding-system-alist))))))) | |
1384 | ||
b25eef20 | 1385 | (defun make-translation-table (&rest args) |
f967223b | 1386 | "Make a translation table (char table) from arguments. |
13d5617d | 1387 | Each argument is a list of the form (FROM . TO), |
b25eef20 | 1388 | where FROM is a character to be translated to TO. |
13d5617d | 1389 | |
452fdb31 RS |
1390 | FROM can be a generic character (see `make-char'). In this case, TO is |
1391 | a generic character containing the same number of characters, or a | |
1392 | ordinary character. If FROM and TO are both generic characters, all | |
b25eef20 | 1393 | characters belonging to FROM are translated to characters belonging to TO |
13d5617d | 1394 | without changing their position code(s)." |
f967223b | 1395 | (let ((table (make-char-table 'translation-table)) |
a73a8c89 KH |
1396 | revlist) |
1397 | (while args | |
1398 | (let ((elts (car args))) | |
1399 | (while elts | |
13d5617d KH |
1400 | (let* ((from (car (car elts))) |
1401 | (from-i 0) ; degree of freedom of FROM | |
1402 | (from-rev (nreverse (split-char from))) | |
1403 | (to (cdr (car elts))) | |
1404 | (to-i 0) ; degree of freedom of TO | |
1405 | (to-rev (nreverse (split-char to)))) | |
1406 | ;; Check numbers of heading 0s in FROM-REV and TO-REV. | |
1407 | (while (eq (car from-rev) 0) | |
1408 | (setq from-i (1+ from-i) from-rev (cdr from-rev))) | |
1409 | (while (eq (car to-rev) 0) | |
1410 | (setq to-i (1+ to-i) to-rev (cdr to-rev))) | |
1411 | (if (and (/= from-i to-i) (/= to-i 0)) | |
1412 | (error "Invalid character pair (%d . %d)" from to)) | |
b25eef20 KH |
1413 | ;; If we have already translated TO to TO-ALT, FROM should |
1414 | ;; also be translated to TO-ALT. But, this is only if TO | |
1415 | ;; is a generic character or TO-ALT is not a generic | |
13d5617d KH |
1416 | ;; character. |
1417 | (let ((to-alt (aref table to))) | |
1418 | (if (and to-alt | |
1419 | (or (> to-i 0) (not (generic-char-p to-alt)))) | |
1420 | (setq to to-alt))) | |
1421 | (if (> from-i 0) | |
1422 | (set-char-table-default table from to) | |
1423 | (aset table from to)) | |
b25eef20 KH |
1424 | ;; If we have already translated some chars to FROM, they |
1425 | ;; should also be translated to TO. | |
a73a8c89 KH |
1426 | (let ((l (assq from revlist))) |
1427 | (if l | |
1428 | (let ((ch (car l))) | |
1429 | (setcar l to) | |
1430 | (setq l (cdr l)) | |
1431 | (while l | |
1432 | (aset table ch to) | |
1433 | (setq l (cdr l)) )))) | |
1434 | ;; Now update REVLIST. | |
1435 | (let ((l (assq to revlist))) | |
1436 | (if l | |
1437 | (setcdr l (cons from (cdr l))) | |
1438 | (setq revlist (cons (list to from) revlist))))) | |
1439 | (setq elts (cdr elts)))) | |
1440 | (setq args (cdr args))) | |
1441 | ;; Return TABLE just created. | |
1442 | table)) | |
1443 | ||
c76b5c99 KH |
1444 | (defun make-translation-table-from-vector (vec) |
1445 | "Make translation table from decoding vector VEC. | |
1446 | VEC is an array of 256 elements to map unibyte codes to multibyte characters. | |
1447 | See also the variable `nonascii-translation-table'." | |
1448 | (let ((table (make-char-table 'translation-table)) | |
1449 | (rev-table (make-char-table 'translation-table)) | |
1450 | (i 0) | |
1451 | ch) | |
1452 | (while (< i 256) | |
1453 | (setq ch (aref vec i)) | |
1454 | (aset table i ch) | |
1455 | (if (>= ch 256) | |
1456 | (aset rev-table ch i)) | |
1457 | (setq i (1+ i))) | |
1458 | (set-char-table-extra-slot table 0 rev-table) | |
1459 | table)) | |
1460 | ||
f967223b | 1461 | (defun define-translation-table (symbol &rest args) |
e8dd0160 | 1462 | "Define SYMBOL as a name of translation table made by ARGS. |
b25eef20 | 1463 | |
007c79c8 KH |
1464 | If the first element of ARGS is a char-table of which purpose is |
1465 | translation-table, just define SYMBOL as the name of it. | |
1466 | ||
1467 | In the other case, ARGS are the same as arguments to the function | |
1468 | `make-translation-table' (which see). | |
b25eef20 | 1469 | |
452fdb31 RS |
1470 | This function sets properties `translation-table' and |
1471 | `translation-table-id' of SYMBOL to the created table itself and | |
f967223b | 1472 | identification number of the table respectively." |
007c79c8 KH |
1473 | (let ((table (if (and (char-table-p (car args)) |
1474 | (eq (char-table-subtype (car args)) | |
1475 | 'translation-table)) | |
1476 | (car args) | |
1477 | (apply 'make-translation-table args))) | |
f967223b | 1478 | (len (length translation-table-vector)) |
d9e3229d | 1479 | (id 0) |
b25eef20 | 1480 | (done nil)) |
f967223b | 1481 | (put symbol 'translation-table table) |
b25eef20 KH |
1482 | (while (not done) |
1483 | (if (>= id len) | |
f967223b KH |
1484 | (setq translation-table-vector |
1485 | (vconcat translation-table-vector (make-vector len nil)))) | |
1486 | (let ((slot (aref translation-table-vector id))) | |
b25eef20 KH |
1487 | (if (or (not slot) |
1488 | (eq (car slot) symbol)) | |
1489 | (progn | |
f967223b | 1490 | (aset translation-table-vector id (cons symbol table)) |
007c79c8 KH |
1491 | (setq done t)) |
1492 | (setq id (1+ id))))) | |
f967223b | 1493 | (put symbol 'translation-table-id id) |
d9e3229d KH |
1494 | id)) |
1495 | ||
35554641 KH |
1496 | (put 'with-category-table 'lisp-indent-function 1) |
1497 | ||
1498 | (defmacro with-category-table (category-table &rest body) | |
1499 | `(let ((current-category-table (category-table))) | |
1500 | (set-category-table ,category-table) | |
1501 | (unwind-protect | |
1502 | (progn ,@body) | |
1503 | (set-category-table current-category-table)))) | |
1504 | ||
69eba008 KH |
1505 | ;;; Initialize some variables. |
1506 | ||
1507 | (put 'use-default-ascent 'char-table-extra-slots 0) | |
1508 | (setq use-default-ascent (make-char-table 'use-default-ascent)) | |
d6d6d592 KH |
1509 | (put 'ignore-relative-composition 'char-table-extra-slots 0) |
1510 | (setq ignore-relative-composition | |
1511 | (make-char-table 'ignore-relative-composition)) | |
69eba008 KH |
1512 | |
1513 | ;;; | |
4ed46869 KH |
1514 | (provide 'mule) |
1515 | ||
1516 | ;;; mule.el ends here |