1999-05-26 Ben Elliston <bje@cygnus.com>
[bpt/emacs.git] / lisp / international / mule.el
CommitLineData
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
eb9386bb 27(defconst mule-version "4.0 (HANANOEN)" "\
4ed46869
KH
28Version number and name of this version of MULE (multilingual environment).")
29
eb9386bb 30(defconst mule-version-date "1998.7.1" "\
4ed46869
KH
31Distribution 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.
35The file contents are decoded before evaluation if necessary.
4ed46869
KH
36If optional second arg NOERROR is non-nil,
37 report no error if FILE doesn't exist.
38Print messages at start and end of loading unless
39 optional third arg NOMESSAGE is non-nil.
40Return 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
RS
87;; (not default-enable-multibyte-characters)
88 ))
cfc70cdf
RS
89 (let (kill-buffer-hook kill-buffer-query-functions)
90 (kill-buffer buffer)))
4ed46869 91 (let ((hook (assoc file after-load-alist)))
db5cae4b
SM
92 (when hook
93 (mapcar (function eval) (cdr hook))))
94 (unless (or nomessage noninteractive)
95 (if source
96 (message "Loading %s (source)...done" file)
97 (message "Loading %s...done" file)))
4ed46869
KH
98 t)))
99
100;; API (Application Program Interface) for charsets.
101
2e02a76f
RS
102;; Return t if OBJ is a quoted symbol
103;; and the symbol is the name of a standard charset.
104(defsubst charset-quoted-standard-p (obj)
105 (and (listp obj) (eq (car obj) 'quote)
106 (symbolp (car-safe (cdr obj)))
107 (let ((vector (get (car-safe (cdr obj)) 'charset)))
108 (and (vectorp vector)
109 (< (aref vector 0) 160)))))
4ed46869
KH
110
111(defsubst charsetp (object)
112 "T is OBJECT is a charset."
113 (and (symbolp object) (vectorp (get object 'charset))))
114
115(defsubst charset-info (charset)
116 "Return a vector of information of CHARSET.
117The elements of the vector are:
118 CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION,
119 LEADING-CODE-BASE, LEADING-CODE-EXT,
120 ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE,
121 REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION,
122 PLIST,
123where
124CHARSET-ID (integer) is the identification number of the charset.
125DIMENSION (integer) is the number of bytes to represent a character of
126the charset: 1 or 2.
127CHARS (integer) is the number of characters in a dimension: 94 or 96.
128BYTE (integer) is the length of multi-byte form of a character in
129 the charset: one of 1, 2, 3, and 4.
130WIDTH (integer) is the number of columns a character in the charset
131 occupies on the screen: one of 0, 1, and 2.
132DIRECTION (integer) is the rendering direction of characters in the
133 charset when rendering. If 0, render from right to left, else
134 render from left to right.
135LEADING-CODE-BASE (integer) is the base leading-code for the
136 charset.
137LEADING-CODE-EXT (integer) is the extended leading-code for the
138 charset. All charsets of less than 0xA0 has the value 0.
139ISO-FINAL-CHAR (character) is the final character of the
140 corresponding ISO 2022 charset.
141ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked
142 while encoding to variants of ISO 2022 coding system, one of the
143 following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR).
144REVERSE-CHARSET (integer) is the charset which differs only in
145 LEFT-TO-RIGHT value from the charset. If there's no such a
146 charset, the value is -1.
147SHORT-NAME (string) is the short name to refer to the charset.
148LONG-NAME (string) is the long name to refer to the charset
149DESCRIPTION (string) is the description string of the charset.
150PLIST (property list) may contain any type of information a user
151 want to put and get by functions `put-charset-property' and
152 `get-charset-property' respectively."
153 (get charset 'charset))
154
155(defmacro charset-id (charset)
156 "Return charset identification number of CHARSET."
2e02a76f 157 (if (charset-quoted-standard-p charset)
4ed46869
KH
158 (aref (charset-info (nth 1 charset)) 0)
159 `(aref (charset-info ,charset) 0)))
160
161(defmacro charset-bytes (charset)
900dc6e3
KH
162 "Return bytes of CHARSET.
163See the function `charset-info' for more detail."
2e02a76f 164 (if (charset-quoted-standard-p charset)
4ed46869
KH
165 (aref (charset-info (nth 1 charset)) 1)
166 `(aref (charset-info ,charset) 1)))
167
168(defmacro charset-dimension (charset)
900dc6e3
KH
169 "Return dimension of CHARSET.
170See the function `charset-info' for more detail."
2e02a76f 171 (if (charset-quoted-standard-p charset)
4ed46869
KH
172 (aref (charset-info (nth 1 charset)) 2)
173 `(aref (charset-info ,charset) 2)))
174
175(defmacro charset-chars (charset)
900dc6e3
KH
176 "Return character numbers contained in a dimension of CHARSET.
177See the function `charset-info' for more detail."
2e02a76f 178 (if (charset-quoted-standard-p charset)
4ed46869
KH
179 (aref (charset-info (nth 1 charset)) 3)
180 `(aref (charset-info ,charset) 3)))
181
182(defmacro charset-width (charset)
900dc6e3
KH
183 "Return width (how many column occupied on a screen) of CHARSET.
184See the function `charset-info' for more detail."
2e02a76f 185 (if (charset-quoted-standard-p charset)
4ed46869
KH
186 (aref (charset-info (nth 1 charset)) 4)
187 `(aref (charset-info ,charset) 4)))
188
189(defmacro charset-direction (charset)
900dc6e3
KH
190 "Return direction of CHARSET.
191See the function `charset-info' for more detail."
2e02a76f 192 (if (charset-quoted-standard-p charset)
4ed46869
KH
193 (aref (charset-info (nth 1 charset)) 5)
194 `(aref (charset-info ,charset) 5)))
195
196(defmacro charset-iso-final-char (charset)
900dc6e3
KH
197 "Return final char of CHARSET.
198See the function `charset-info' for more detail."
2e02a76f 199 (if (charset-quoted-standard-p charset)
4ed46869
KH
200 (aref (charset-info (nth 1 charset)) 8)
201 `(aref (charset-info ,charset) 8)))
202
203(defmacro charset-iso-graphic-plane (charset)
900dc6e3
KH
204 "Return graphic plane of CHARSET.
205See the function `charset-info' for more detail."
2e02a76f 206 (if (charset-quoted-standard-p charset)
4ed46869
KH
207 (aref (charset-info (nth 1 charset)) 9)
208 `(aref (charset-info ,charset) 9)))
209
210(defmacro charset-reverse-charset (charset)
900dc6e3
KH
211 "Return reverse charset of CHARSET.
212See the function `charset-info' for more detail."
2e02a76f 213 (if (charset-quoted-standard-p charset)
4ed46869
KH
214 (aref (charset-info (nth 1 charset)) 10)
215 `(aref (charset-info ,charset) 10)))
216
217(defmacro charset-short-name (charset)
900dc6e3
KH
218 "Return short name of CHARSET.
219See the function `charset-info' for more detail."
2e02a76f 220 (if (charset-quoted-standard-p charset)
4ed46869
KH
221 (aref (charset-info (nth 1 charset)) 11)
222 `(aref (charset-info ,charset) 11)))
223
224(defmacro charset-long-name (charset)
900dc6e3
KH
225 "Return long name of CHARSET.
226See the function `charset-info' for more detail."
2e02a76f 227 (if (charset-quoted-standard-p charset)
4ed46869
KH
228 (aref (charset-info (nth 1 charset)) 12)
229 `(aref (charset-info ,charset) 12)))
230
231(defmacro charset-description (charset)
0611934b 232 "Return description of CHARSET.
900dc6e3 233See the function `charset-info' for more detail."
2e02a76f 234 (if (charset-quoted-standard-p charset)
4ed46869
KH
235 (aref (charset-info (nth 1 charset)) 13)
236 `(aref (charset-info ,charset) 13)))
237
238(defmacro charset-plist (charset)
900dc6e3
KH
239 "Return list charset property of CHARSET.
240See the function `charset-info' for more detail."
2e02a76f 241 (if (charset-quoted-standard-p charset)
f98e2797 242 `(aref ,(charset-info (nth 1 charset)) 14)
4ed46869
KH
243 `(aref (charset-info ,charset) 14)))
244
245(defun set-charset-plist (charset plist)
e8dd0160 246 "Set CHARSET's property list to PLIST, and return PLIST."
4ed46869
KH
247 (aset (charset-info charset) 14 plist))
248
a6acd8a2 249(defun make-char (charset &optional c1 c2)
a9fb0b58 250 "Return a character of CHARSET and position codes CODE1 and CODE2.
f98e2797 251CODE1 and CODE2 are optional, but if you don't supply
a9fb0b58
KH
252sufficient position codes, return a generic character which stands for
253all characters or group of characters in the character set.
a73a8c89 254A generic character can be used to index a char table (e.g. syntax-table)."
a6acd8a2
KH
255 (make-char-internal (charset-id charset) c1 c2))
256
257(put 'make-char 'byte-compile
258 (function
259 (lambda (form)
260 (let ((charset (nth 1 form)))
261 (if (charset-quoted-standard-p charset)
262 (byte-compile-normal-call
263 (cons 'make-char-internal
264 (cons (charset-id (nth 1 charset)) (nthcdr 2 form))))
265 (byte-compile-normal-call
266 (cons 'make-char-internal
267 (cons (list 'charset-id charset) (nthcdr 2 form)))))))))
4ed46869 268
d3675a42 269(defun charset-list ()
900dc6e3
KH
270 "Return list of charsets ever defined.
271
d3675a42 272This function is provided for backward compatibility.
900dc6e3 273Now we have the variable `charset-list'."
d3675a42
KH
274 charset-list)
275
13d5617d
KH
276(defsubst generic-char-p (char)
277 "Return t if and only if CHAR is a generic character.
278See also the documentation of make-char."
1d935669
KH
279 (and (>= char 0400)
280 (let ((l (split-char char)))
281 (and (or (= (nth 1 l) 0) (eq (nth 2 l) 0))
282 (not (eq (car l) 'composition))))))
69eba008 283
0269ddfb 284\f
8057896b 285;; Coding system staffs
4ed46869 286
8057896b 287;; Coding system is a symbol that has the property `coding-system'.
4ed46869 288;;
8057896b
KH
289;; The value of the property `coding-system' is a vector of the
290;; following format:
d3675a42 291;; [TYPE MNEMONIC DOC-STRING PLIST FLAGS]
8057896b 292;; We call this vector as coding-spec. See comments in src/coding.c
d3675a42 293;; for more detail.
8057896b
KH
294
295(defconst coding-spec-type-idx 0)
296(defconst coding-spec-mnemonic-idx 1)
297(defconst coding-spec-doc-string-idx 2)
0269ddfb 298(defconst coding-spec-plist-idx 3)
8057896b
KH
299(defconst coding-spec-flags-idx 4)
300
a6acd8a2
KH
301;; PLIST is a property list of a coding system. To share PLIST among
302;; alias coding systems, a coding system has PLIST in coding-spec
303;; instead of having it in normal property list of Lisp symbol.
304;; Here's a list of coding system properties currently being used.
0269ddfb
KH
305;;
306;; o coding-category
307;;
308;; The value is a coding category the coding system belongs to. The
309;; function `make-coding-system' and `define-coding-system-alias' sets
310;; this value automatically.
4ed46869 311;;
0269ddfb 312;; o alias-coding-systems
4ed46869 313;;
0269ddfb
KH
314;; The value is a list of coding systems of the same alias group. The
315;; first element is the coding system made at first, which we call as
316;; `base coding system'. The function `make-coding-system' and
317;; `define-coding-system-alias' set this value automatically.
318;;
319;; o post-read-conversion
320;;
321;; The value is a function to call after some text is inserted and
322;; decoded by the coding system itself and before any functions in
323;; `after-insert-functions' are called. The arguments to this
324;; function is the same as those of a function in
4ed46869
KH
325;; `after-insert-functions', i.e. LENGTH of a text while putting point
326;; at the head of the text to be decoded
327;;
0269ddfb
KH
328;; o pre-write-conversion
329;;
330;; The value is a function to call after all functions in
331;; `write-region-annotate-functions' and `buffer-file-format' are
332;; called, and before the text is encoded by the coding system itself.
333;; The arguments to this function is the same as those of a function
334;; in `write-region-annotate-functions', i.e. FROM and TO specifying
335;; region of a text.
336;;
f967223b 337;; o translation-table-for-decode
0269ddfb 338;;
f967223b
KH
339;; The value is a translation table to be applied on decoding. See
340;; the function `make-translation-table' for the format of translation
341;; table.
0269ddfb 342;;
f967223b 343;; o translation-table-for-encode
0269ddfb 344;;
f967223b 345;; The value is a translation table to be applied on encoding.
a6acd8a2
KH
346;;
347;; o safe-charsets
348;;
349;; The value is a list of charsets safely supported by the coding
350;; system. The value t means that all charsets Emacs handles are
351;; supported. Even if some charset is not in this list, it doesn't
352;; mean that the charset can't be encoded in the coding system,
353;; instead, it just means that some other receiver of a text encoded
354;; in the coding system won't be able to handle that charset.
b25eef20
KH
355;;
356;; o mime-charset
357;;
358;; The value is a symbol of which name is `MIME-charset' parameter of
359;; the coding system.
c76b5c99 360;;
1322c8ca
KH
361;; o charset-origin-alist
362;;
363;; The value is a list of this form:
364;; (CHARSET EXTERNAL-CHARSET-NAME ENCODING-FUNCTION).
365;; ENCODING-FUNCTION is a function to encode a character in CHARSET
366;; to the code in EXTERNAL-CHARSET-NAME. The command what-cursor-position
367;; uses this information of the buffer-file-coding-system.
caa85ad9
KH
368;; ENCODING-FUNCTION may be a translation table or a symbol whose
369;; property `translation-table' is a translation table. In these case,
370;; the translation table is used to encode the character.
1322c8ca 371;;
c76b5c99
KH
372;; o valid-codes (meaningful only for a coding system based on CCL)
373;;
374;; The value is a list to indicate valid byte ranges of the encoded
375;; file. Each element of the list is an integer or a cons of integer.
376;; In the former case, the integer value is a valid byte code. In the
e8dd0160 377;; latter case, the integers specifies the range of valid byte codes.
c76b5c99 378
0269ddfb
KH
379
380;; Return coding-spec of CODING-SYSTEM
381(defsubst coding-system-spec (coding-system)
382 (get (check-coding-system coding-system) 'coding-system))
4ed46869 383
4ed46869 384(defun coding-system-type (coding-system)
0269ddfb
KH
385 "Return the coding type of CODING-SYSTEM.
386A coding type is an integer value indicating the encoding method
387of CODING-SYSTEM. See the function `make-coding-system' for more detail."
388 (aref (coding-system-spec coding-system) coding-spec-type-idx))
4ed46869 389
4ed46869 390(defun coding-system-mnemonic (coding-system)
0269ddfb 391 "Return the mnemonic character of CODING-SYSTEM.
6e2c8840
KH
392The mnemonic character of a coding system is used in mode line
393to indicate the coding system. If the arg is nil, return ?-."
394 (let ((spec (coding-system-spec coding-system)))
395 (if spec (aref spec coding-spec-mnemonic-idx) ?-)))
4ed46869 396
8057896b 397(defun coding-system-doc-string (coding-system)
0269ddfb
KH
398 "Return the documentation string for CODING-SYSTEM."
399 (aref (coding-system-spec coding-system) coding-spec-doc-string-idx))
4ed46869 400
d3675a42 401(defun coding-system-plist (coding-system)
0269ddfb
KH
402 "Return the property list of CODING-SYSTEM."
403 (aref (coding-system-spec coding-system) coding-spec-plist-idx))
d3675a42 404
4ed46869 405(defun coding-system-flags (coding-system)
0269ddfb
KH
406 "Return `flags' of CODING-SYSTEM.
407A `flags' of a coding system is a vector of length 32 indicating detailed
408information of a coding system. See the function `make-coding-system'
409for more detail."
410 (aref (coding-system-spec coding-system) coding-spec-flags-idx))
411
412(defun coding-system-get (coding-system prop)
413 "Extract a value from CODING-SYSTEM's property list for property PROP."
414 (plist-get (coding-system-plist coding-system) prop))
415
416(defun coding-system-put (coding-system prop val)
417 "Change value in CODING-SYSTEM's property list PROP to VAL."
418 (let ((plist (coding-system-plist coding-system)))
419 (if plist
420 (plist-put plist prop val)
421 (aset (coding-system-spec coding-system) coding-spec-plist-idx
422 (list prop val)))))
423
424(defun coding-system-category (coding-system)
425 "Return the coding category of CODING-SYSTEM."
426 (coding-system-get coding-system 'coding-category))
427
428(defun coding-system-base (coding-system)
429 "Return the base coding system of CODING-SYSTEM.
d9e3229d
KH
430A base coding system is what made by `make-coding-system'.
431Any alias nor subsidiary coding systems are not base coding system."
0269ddfb
KH
432 (car (coding-system-get coding-system 'alias-coding-systems)))
433
434(defalias 'coding-system-parent 'coding-system-base)
435(make-obsolete 'coding-system-parent 'coding-system-base)
436
437;; Coding system also has a property `eol-type'.
438;;
439;; This property indicates how the coding system handles end-of-line
440;; format. The value is integer 0, 1, 2, or a vector of three coding
441;; systems. Each integer value 0, 1, and 2 indicates the format of
442;; end-of-line LF, CRLF, and CR respectively. A vector value
443;; indicates that the format of end-of-line should be detected
444;; automatically. Nth element of the vector is the subsidiary coding
445;; system whose `eol-type' property is N.
4ed46869 446
8057896b 447(defun coding-system-eol-type (coding-system)
0269ddfb
KH
448 "Return eol-type of CODING-SYSTEM.
449An eol-type is integer 0, 1, 2, or a vector of coding systems.
4ed46869 450
0269ddfb
KH
451Integer values 0, 1, and 2 indicate a format of end-of-line; LF,
452CRLF, and CR respectively.
453
454A vector value indicates that a format of end-of-line should be
455detected automatically. Nth element of the vector is the subsidiary
456coding system whose eol-type is N."
457 (get coding-system 'eol-type))
bd882697 458
857ea15c
AS
459(defun coding-system-lessp (x y)
460 (cond ((eq x 'no-conversion) t)
461 ((eq y 'no-conversion) nil)
462 ((eq x 'emacs-mule) t)
463 ((eq y 'emacs-mule) nil)
464 ((eq x 'undecided) t)
465 ((eq y 'undecided) nil)
466 (t (let ((c1 (coding-system-mnemonic x))
467 (c2 (coding-system-mnemonic y)))
468 (or (< (downcase c1) (downcase c2))
469 (and (not (> (downcase c1) (downcase c2)))
470 (< c1 c2)))))))
471
472;; Add CODING-SYSTEM to coding-system-list while keeping it sorted.
473(defun add-to-coding-system-list (coding-system)
474 (if (or (null coding-system-list)
475 (coding-system-lessp coding-system (car coding-system-list)))
476 (setq coding-system-list (cons coding-system coding-system-list))
477 (let ((len (length coding-system-list))
478 mid (tem coding-system-list))
479 (while (> len 1)
480 (setq mid (nthcdr (/ len 2) tem))
481 (if (coding-system-lessp (car mid) coding-system)
482 (setq tem mid
483 len (- len (/ len 2)))
484 (setq len (/ len 2))))
485 (setcdr tem (cons coding-system (cdr tem))))))
486
80a7463d
KH
487(defun coding-system-list (&optional base-only)
488 "Return a list of all existing coding systems.
489If optional arg BASE-ONLY is non-nil, only base coding systems are listed."
490 (let* ((codings (copy-sequence coding-system-list))
491 (tail (cons nil codings)))
492 ;; Remove subsidiary coding systems (eol variants) and alias
493 ;; coding systems (if necessary).
494 (while (cdr tail)
495 (let* ((coding (car (cdr tail)))
496 (aliases (coding-system-get coding 'alias-coding-systems)))
497 (if (or
498 ;; CODING is an eol variant if not in ALIASES.
499 (not (memq coding aliases))
500 ;; CODING is an alias if it is not car of ALIASES.
501 (and base-only (not (eq coding (car aliases)))))
502 (setcdr tail (cdr (cdr tail)))
503 (setq tail (cdr tail)))))
504 codings))
505
6e9722b0
KH
506;; Make subsidiary coding systems (eol-type variants) of CODING-SYSTEM.
507(defun make-subsidiary-coding-system (coding-system)
0269ddfb
KH
508 (let ((coding-spec (coding-system-spec coding-system))
509 (subsidiaries (vector (intern (format "%s-unix" coding-system))
8057896b
KH
510 (intern (format "%s-dos" coding-system))
511 (intern (format "%s-mac" coding-system))))
0269ddfb
KH
512 (i 0)
513 temp)
8057896b 514 (while (< i 3)
0269ddfb 515 (put (aref subsidiaries i) 'coding-system coding-spec)
8057896b 516 (put (aref subsidiaries i) 'eol-type i)
857ea15c 517 (add-to-coding-system-list (aref subsidiaries i))
0269ddfb
KH
518 (setq coding-system-alist
519 (cons (list (symbol-name (aref subsidiaries i)))
520 coding-system-alist))
8057896b
KH
521 (setq i (1+ i)))
522 subsidiaries))
4ed46869 523
8057896b 524(defun make-coding-system (coding-system type mnemonic doc-string
d9e3229d 525 &optional flags properties)
3bb1accb 526 "Define a new coding system CODING-SYSTEM (symbol).
a6acd8a2 527Remaining arguments are TYPE, MNEMONIC, DOC-STRING, FLAGS (optional),
d9e3229d 528and PROPERTIES (optional) which construct a coding-spec of CODING-SYSTEM
a6acd8a2 529in the following format:
0269ddfb 530 [TYPE MNEMONIC DOC-STRING PLIST FLAGS]
3bb1accb
KH
531
532TYPE is an integer value indicating the type of the coding system as follows:
4ed46869
KH
533 0: Emacs internal format,
534 1: Shift-JIS (or MS-Kanji) used mainly on Japanese PC,
535 2: ISO-2022 including many variants,
536 3: Big5 used mainly on Chinese PC,
cd9c3177
KH
537 4: private, CCL programs provide encoding/decoding algorithm,
538 5: Raw-text, which means that text contains random 8-bit codes.
0269ddfb 539
3bb1accb 540MNEMONIC is a character to be displayed on mode line for the coding system.
0269ddfb 541
3bb1accb 542DOC-STRING is a documentation string for the coding system.
0269ddfb 543
3bb1accb 544FLAGS specifies more detailed information of the coding system as follows:
d6d6d592 545
3bb1accb 546 If TYPE is 2 (ISO-2022), FLAGS is a list of these elements:
4ed46869
KH
547 CHARSET0, CHARSET1, CHARSET2, CHARSET3, SHORT-FORM,
548 ASCII-EOL, ASCII-CNTL, SEVEN, LOCKING-SHIFT, SINGLE-SHIFT,
d6d6d592 549 USE-ROMAN, USE-OLDJIS, NO-ISO6429, INIT-BOL, DESIGNATION-BOL,
850101ed 550 SAFE, ACCEPT-LATIN-EXTRA-CODE.
4ed46869
KH
551 CHARSETn are character sets initially designated to Gn graphic registers.
552 If CHARSETn is nil, Gn is never used.
553 If CHARSETn is t, Gn can be used but nothing designated initially.
554 If CHARSETn is a list of character sets, those character sets are
555 designated to Gn on output, but nothing designated to Gn initially.
421e3b4e 556 But, character set `ascii' can be designated only to G0.
4ed46869
KH
557 SHORT-FORM non-nil means use short designation sequence on output.
558 ASCII-EOL non-nil means designate ASCII to g0 at end of line on output.
559 ASCII-CNTL non-nil means designate ASCII to g0 before control codes and
560 SPACE on output.
561 SEVEN non-nil means use 7-bit code only on output.
562 LOCKING-SHIFT non-nil means use locking-shift.
563 SINGLE-SHIFT non-nil means use single-shift.
564 USE-ROMAN non-nil means designate JIS0201-1976-Roman instead of ASCII.
565 USE-OLDJIS non-nil means designate JIS0208-1976 instead of JIS0208-1983.
566 NO-ISO6429 non-nil means not use ISO6429's direction specification.
69eba008
KH
567 INIT-BOL non-nil means any designation state is assumed to be reset
568 to initial at each beginning of line on output.
569 DESIGNATION-BOL non-nil means designation sequences should be placed
570 at beginning of line on output.
a6acd8a2
KH
571 SAFE non-nil means convert unsafe characters to `?' on output.
572 Unsafe characters are what not specified in SAFE-CHARSET.
850101ed
RS
573 ACCEPT-LATIN-EXTRA-CODE non-nil means code-detection routine accepts
574 a code specified in `latin-extra-code-table' (which see) as a valid
575 code of the coding system.
d6d6d592 576
3bb1accb
KH
577 If TYPE is 4 (private), FLAGS should be a cons of CCL programs, for
578 decoding and encoding. CCL programs should be specified by their
579 symbols.
d9e3229d
KH
580
581PROPERTIES is an alist of properties vs the corresponding values.
582These properties are set in PLIST, a property list. This function
583also sets properties `coding-category' and `alias-coding-systems'
584automatically.
4ed46869 585
3bb1accb
KH
586Kludgy features for backward compatibility:
587
5881. If TYPE is 4 and car or cdr of FLAGS is a vector, the vector is
589treated as a compiled CCL code.
590
5912. If PROPERTIES is just a list of character sets, the list is set as
592a value of `safe-charsets' in PLIST."
0269ddfb 593 (if (memq coding-system coding-system-list)
d9e3229d 594 (error "Coding system %s already exists" coding-system))
0269ddfb
KH
595
596 ;; Set a value of `coding-system' property.
6e9722b0 597 (let ((coding-spec (make-vector 5 nil))
d9e3229d
KH
598 (no-initial-designation t)
599 (no-alternative-designation t)
6e9722b0 600 coding-category)
cd9c3177 601 (if (or (not (integerp type)) (< type 0) (> type 5))
0269ddfb 602 (error "TYPE argument must be 0..5"))
8057896b 603 (if (or (not (integerp mnemonic)) (<= mnemonic ? ) (> mnemonic 127))
e8dd0160 604 (error "MNEMONIC argument must be an ASCII printable character."))
0269ddfb
KH
605 (aset coding-spec coding-spec-type-idx type)
606 (aset coding-spec coding-spec-mnemonic-idx mnemonic)
607 (aset coding-spec coding-spec-doc-string-idx
608 (if (stringp doc-string) doc-string ""))
6e9722b0
KH
609 (cond ((= type 0)
610 (setq coding-category 'coding-category-emacs-mule))
611 ((= type 1)
612 (setq coding-category 'coding-category-sjis))
613 ((= type 2) ; ISO2022
4ed46869 614 (let ((i 0)
6e9722b0 615 (vec (make-vector 32 nil))
05134257
KH
616 (g1-designation nil)
617 (fl flags))
4ed46869 618 (while (< i 4)
05134257 619 (let ((charset (car fl)))
6e9722b0
KH
620 (if (and no-initial-designation
621 (> i 0)
622 (or (charsetp charset)
623 (and (consp charset)
624 (charsetp (car charset)))))
625 (setq no-initial-designation nil))
626 (if (charsetp charset)
627 (if (= i 1) (setq g1-designation charset))
628 (if (consp charset)
629 (let ((tail charset)
630 elt)
631 (while tail
632 (setq elt (car tail))
d9e3229d
KH
633 (if (eq elt t)
634 (setq no-alternative-designation nil)
635 (if (and elt (not (charsetp elt)))
636 (error "Invalid charset: %s" elt)))
6e9722b0
KH
637 (setq tail (cdr tail)))
638 (setq g1-designation (car charset)))
d9e3229d
KH
639 (if charset
640 (if (eq charset t)
641 (setq no-alternative-designation nil)
642 (error "Invalid charset: %s" charset)))))
4ed46869 643 (aset vec i charset))
05134257
KH
644 (setq fl (cdr fl) i (1+ i)))
645 (while (and (< i 32) fl)
646 (aset vec i (car fl))
647 (setq fl (cdr fl) i (1+ i)))
6e9722b0 648 (aset coding-spec 4 vec)
6e9722b0
KH
649 (setq coding-category
650 (if (aref vec 8) ; Use locking-shift.
dc64cd19
KH
651 (or (and (aref vec 7) 'coding-category-iso-7-else)
652 'coding-category-iso-8-else)
6e9722b0
KH
653 (if (aref vec 7) ; 7-bit only.
654 (if (aref vec 9) ; Use single-shift.
dc64cd19 655 'coding-category-iso-7-else
d9e3229d
KH
656 (if no-alternative-designation
657 'coding-category-iso-7-tight
658 'coding-category-iso-7))
3ad911d8
KH
659 (if (or no-initial-designation
660 (not no-alternative-designation))
dc64cd19 661 'coding-category-iso-8-else
6e9722b0
KH
662 (if (and (charsetp g1-designation)
663 (= (charset-dimension g1-designation) 2))
664 'coding-category-iso-8-2
665 'coding-category-iso-8-1)))))))
666 ((= type 3)
667 (setq coding-category 'coding-category-big5))
668 ((= type 4) ; private
c76b5c99 669 (setq coding-category 'coding-category-ccl)
3bb1accb
KH
670 (if (not (consp flags))
671 (error "Invalid FLAGS argument for TYPE 4 (CCL)")
672 (let ((decoder (check-ccl-program
673 (car flags)
674 (intern (format "%s-decoder" coding-system))))
675 (encoder (check-ccl-program
676 (cdr flags)
677 (intern (format "%s-encoder" coding-system)))))
678 (if (and decoder encoder)
679 (aset coding-spec 4 (cons decoder encoder))
680 (error "Invalid FLAGS argument for TYPE 4 (CCL)")))))
cd9c3177
KH
681 (t ; i.e. (= type 5)
682 (setq coding-category 'coding-category-raw-text)))
0269ddfb
KH
683
684 (let ((plist (list 'coding-category coding-category
d9e3229d 685 'alias-coding-systems (list coding-system))))
0269ddfb 686 (if no-initial-designation
d9e3229d
KH
687 (plist-put plist 'no-initial-designation t))
688 (if (and properties
689 (or (eq properties t)
690 (not (consp (car properties)))))
691 ;; In the old version, the arg PROPERTIES is a list to be
692 ;; set in PLIST as a value of property `safe-charsets'.
693 (plist-put plist 'safe-charsets properties)
05134257
KH
694 (let ((l properties))
695 (while l
696 (plist-put plist (car (car l)) (cdr (car l)))
697 (setq l (cdr l)))))
0269ddfb 698 (aset coding-spec coding-spec-plist-idx plist))
6e9722b0 699 (put coding-system 'coding-system coding-spec)
6e9722b0
KH
700 (put coding-category 'coding-systems
701 (cons coding-system (get coding-category 'coding-systems))))
4ed46869
KH
702
703 ;; Next, set a value of `eol-type' property. The value is a vector
6e9722b0 704 ;; of subsidiary coding systems, each corresponds to a coding system
4ed46869 705 ;; for the detected end-of-line format.
8057896b 706 (put coding-system 'eol-type
cd9c3177 707 (if (or (<= type 3) (= type 5))
6e9722b0 708 (make-subsidiary-coding-system coding-system)
0269ddfb
KH
709 0))
710
711 ;; At last, register CODING-SYSTEM in `coding-system-list' and
712 ;; `coding-system-alist'.
857ea15c 713 (add-to-coding-system-list coding-system)
0269ddfb 714 (setq coding-system-alist (cons (list (symbol-name coding-system))
d9e3229d 715 coding-system-alist))
05134257
KH
716
717 ;; For a coding system of cateogory iso-8-1 and iso-8-2, create
718 ;; XXX-with-esc variants.
719 (let ((coding-category (coding-system-category coding-system)))
720 (if (or (eq coding-category 'coding-category-iso-8-1)
721 (eq coding-category 'coding-category-iso-8-2))
722 (let ((esc (intern (concat (symbol-name coding-system) "-with-esc")))
723 (doc (format "Same as %s but can handle any charsets by ISO's escape sequences." coding-system)))
724 (make-coding-system esc type mnemonic doc
725 (if (listp (car flags))
726 (cons (append (car flags) '(t)) (cdr flags))
727 (cons (list (car flags) t) (cdr flags)))
728 properties)
d00742a1 729 (coding-system-put esc 'mime-charset nil)
05134257
KH
730 (coding-system-put esc 'safe-charsets t))))
731
d9e3229d 732 coding-system)
8057896b 733
a42763dc 734(defun define-coding-system-alias (alias coding-system)
358d28fb 735 "Define ALIAS as an alias for coding system CODING-SYSTEM."
0269ddfb
KH
736 (put alias 'coding-system (coding-system-spec coding-system))
737 (nconc (coding-system-get alias 'alias-coding-systems) (list alias))
857ea15c 738 (add-to-coding-system-list alias)
0269ddfb
KH
739 (setq coding-system-alist (cons (list (symbol-name alias))
740 coding-system-alist))
741 (let ((eol-type (coding-system-eol-type coding-system)))
742 (if (vectorp eol-type)
743 (put alias 'eol-type (make-subsidiary-coding-system alias))
744 (put alias 'eol-type eol-type))))
4ed46869
KH
745
746(defun set-buffer-file-coding-system (coding-system &optional force)
358d28fb
RS
747 "Set the file coding-system of the current buffer to CODING-SYSTEM.
748This means that when you save the buffer, it will be converted
749according to CODING-SYSTEM. For a list of possible values of CODING-SYSTEM,
750use \\[list-coding-systems].
751
752If the buffer's previous file coding-system value specifies end-of-line
753conversion, and CODING-SYSTEM does not specify one, CODING-SYSTEM is
754merged with the already-specified end-of-line conversion.
755However, if the optional prefix argument FORCE is non-nil,
0269ddfb 756then CODING-SYSTEM is used exactly as specified."
d9e3229d 757 (interactive "zCoding system for visited file (default, nil): \nP")
4ed46869
KH
758 (check-coding-system coding-system)
759 (if (null force)
8057896b
KH
760 (let ((x (coding-system-eol-type buffer-file-coding-system))
761 (y (coding-system-eol-type coding-system)))
4ed46869
KH
762 (if (and (numberp x) (>= x 0) (<= x 2) (vectorp y))
763 (setq coding-system (aref y x)))))
764 (setq buffer-file-coding-system coding-system)
765 (set-buffer-modified-p t)
766 (force-mode-line-update))
767
358d28fb
RS
768(defvar default-terminal-coding-system nil
769 "Default value for the terminal coding system.
770This is normally set according to the selected language environment.
771See also the command `set-terminal-coding-system'.")
772
df100398
KH
773(defun set-terminal-coding-system (coding-system)
774 "Set coding system of your terminal to CODING-SYSTEM.
358d28fb
RS
775All text output to the terminal will be encoded
776with the specified coding system.
777For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems].
778The default is determined by the selected language environment
779or by the previous use of this command."
780 (interactive
2e02a76f
RS
781 (list (let ((default (if (and (not (terminal-coding-system))
782 default-terminal-coding-system)
783 default-terminal-coding-system)))
784 (read-coding-system
785 (format "Coding system for terminal display (default, %s): "
786 default)
787 default))))
358d28fb
RS
788 (if (and (not coding-system)
789 (not (terminal-coding-system)))
790 (setq coding-system default-terminal-coding-system))
791 (if coding-system
792 (setq default-terminal-coding-system coding-system))
df100398
KH
793 (set-terminal-coding-system-internal coding-system)
794 (redraw-frame (selected-frame)))
795
358d28fb
RS
796(defvar default-keyboard-coding-system nil
797 "Default value of the keyboard coding system.
798This is normally set according to the selected language environment.
799See also the command `set-keyboard-coding-system'.")
800
df100398 801(defun set-keyboard-coding-system (coding-system)
358d28fb
RS
802 "Set coding system for keyboard input to CODING-SYSTEM.
803In addition, this command enables Encoded-kbd minor mode.
26471f06 804\(If CODING-SYSTEM is nil, Encoded-kbd mode is turned off.)
358d28fb
RS
805For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems].
806The default is determined by the selected language environment
807or by the previous use of this command."
808 (interactive
2e02a76f
RS
809 (list (let ((default (if (and (not (keyboard-coding-system))
810 default-keyboard-coding-system)
811 default-keyboard-coding-system)))
812 (read-coding-system
813 (format "Coding system for keyboard input (default, %s): "
814 default)
815 default))))
358d28fb
RS
816 (if (and (not coding-system)
817 (not (keyboard-coding-system)))
818 (setq coding-system default-keyboard-coding-system))
819 (if coding-system
820 (setq default-keyboard-coding-system coding-system))
df100398
KH
821 (set-keyboard-coding-system-internal coding-system)
822 (encoded-kbd-mode (if coding-system 1 0)))
823
824(defun set-buffer-process-coding-system (decoding encoding)
358d28fb 825 "Set coding systems for the process associated with the current buffer.
df100398 826DECODING is the coding system to be used to decode input from the process,
358d28fb
RS
827ENCODING is the coding system to be used to encode output to the process.
828
829For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]."
4ed46869
KH
830 (interactive
831 "zCoding-system for process input: \nzCoding-system for process output: ")
832 (let ((proc (get-buffer-process (current-buffer))))
833 (if (null proc)
834 (error "no process")
df100398
KH
835 (check-coding-system decoding)
836 (check-coding-system encoding)
837 (set-process-coding-system proc decoding encoding)))
4ed46869
KH
838 (force-mode-line-update))
839
d0b99881
RS
840(defalias 'set-clipboard-coding-system 'set-selection-coding-system)
841
14915c37 842(defun set-selection-coding-system (coding-system)
b25eef20
KH
843 "Make CODING-SYSTEM used for communicating with other X clients .
844When sending or receiving text via cut_buffer, selection, and clipboard,
845the text is encoded or decoded by CODING-SYSTEM."
a03b3ce1 846 (interactive "zCoding system for X selection: ")
b25eef20 847 (check-coding-system coding-system)
14915c37 848 (setq selection-coding-system coding-system))
b25eef20 849
e8dd0160 850;; Coding system lastly specified by the command
a03b3ce1
KH
851;; set-next-selection-coding-system.
852(defvar last-next-selection-coding-system nil)
853
854(defun set-next-selection-coding-system (coding-system)
855 "Make CODING-SYSTEM used for the next communication with other X clients.
856This setting is effective for the next communication only."
857 (interactive
858 (list (read-coding-system
859 (if last-next-selection-coding-system
860 (format "Coding system for the next X selection (default, %S): "
861 last-next-selection-coding-system)
862 "Coding system for the next X selection: ")
863 last-next-selection-coding-system)))
864 (if coding-system
865 (setq last-next-selection-coding-system coding-system)
866 (setq coding-system last-next-selection-coding-system))
867 (check-coding-system coding-system)
868
869 (setq next-selection-coding-system coding-system))
870
4ed46869 871(defun set-coding-priority (arg)
d9e3229d
KH
872 "Set priority of coding categories according to LIST.
873LIST is a list of coding categories ordered by priority."
874 (let ((l arg)
875 (current-list (copy-sequence coding-category-list)))
e8dd0160 876 ;; Check the validity of ARG while deleting coding categories in
d9e3229d
KH
877 ;; ARG from CURRENT-LIST. We assume that CODING-CATEGORY-LIST
878 ;; contains all coding categories.
879 (while l
880 (if (or (null (get (car l) 'coding-category-index))
881 (null (memq (car l) current-list)))
882 (error "Invalid or duplicated element in argument: %s" arg))
883 (setq current-list (delq (car l) current-list))
884 (setq l (cdr l)))
4ed46869 885 ;; Update `coding-category-list' and return it.
2feaf204
KH
886 (setq coding-category-list (append arg current-list))
887 (set-coding-priority-internal)))
4ed46869
KH
888
889;;; FILE I/O
890
835f49b8 891(defvar auto-coding-alist
0735296c
DL
892 '(("\\.\\(arc\\|zip\\|lzh\\|zoo\\|jar\\|tar\\|tgz\\)\\'" . no-conversion)
893 ("\\.\\(ARC\\|ZIP\\|LZH\\|ZOO\\|JAR\\|TAR\\|TGZ\\)\\'" . no-conversion))
835f49b8
KH
894 "Alist of filename patterns vs corresponding coding systems.
895Each element looks like (REGEXP . CODING-SYSTEM).
558b0c86 896A file whose name matches REGEXP is decoded by CODING-SYSTEM on reading.
835f49b8 897
7fed493a
RS
898The settings in this alist take priority over `coding:' tags
899in the file (see the function `set-auto-coding')
900and the contents of `file-coding-system-alist'.")
835f49b8 901
1c4cc63a
KH
902(defvar set-auto-coding-for-load nil
903 "Non-nil means look for `load-coding' property instead of `coding'.
904This is used for loading and byte-compiling Emacs Lisp files.")
905
8a592131
RS
906(defun auto-coding-alist-lookup (filename)
907 "Return the coding system specified by `auto-coding-alist' for FILENAME."
908 (let ((alist auto-coding-alist)
ca128d75 909 (case-fold-search (memq system-type '(vax-vms windows-nt ms-dos)))
8a592131
RS
910 coding-system)
911 (while (and alist (not coding-system))
912 (if (string-match (car (car alist)) filename)
913 (setq coding-system (cdr (car alist)))
914 (setq alist (cdr alist))))
915 coding-system))
916
835f49b8
KH
917(defun set-auto-coding (filename size)
918 "Return coding system for a file FILENAME of which SIZE bytes follow point.
1c4cc63a
KH
919These bytes should include at least the first 1k of the file
920and the last 3k of the file, but the middle may be omitted.
63561304 921
835f49b8
KH
922It checks FILENAME against the variable `auto-coding-alist'.
923If FILENAME doesn't match any entries in the variable,
924it checks for a `coding:' tag in the first one or two lines following
e8dd0160 925point. If no `coding:' tag is found, it checks for local variables
1c4cc63a 926list in the last 3K bytes out of the SIZE bytes.
63561304
KH
927
928The return value is the specified coding system,
929or nil if nothing specified.
87aba788 930
ba74e833 931The variable `set-auto-coding-function' (which see) is set to this
87aba788 932function by default."
8a592131 933 (let ((coding-system (auto-coding-alist-lookup filename)))
835f49b8
KH
934
935 (or coding-system
936 (let* ((case-fold-search t)
937 (head-start (point))
938 (head-end (+ head-start (min size 1024)))
939 (tail-start (+ head-start (max (- size 3072) 0)))
940 (tail-end (+ head-start size))
941 coding-system head-found tail-found pos)
942 ;; Try a short cut by searching for the string "coding:"
e8dd0160 943 ;; and for "unibyte:" at the head and tail of SIZE bytes.
835f49b8
KH
944 (setq head-found (or (search-forward "coding:" head-end t)
945 (search-forward "unibyte:" head-end t)))
946 (if (and head-found (> head-found tail-start))
947 ;; Head and tail are overlapped.
948 (setq tail-found head-found)
949 (goto-char tail-start)
950 (setq tail-found (or (search-forward "coding:" tail-end t)
951 (search-forward "unibyte:" tail-end t))))
952
953 ;; At first check the head.
954 (when head-found
955 (goto-char head-start)
956 (setq pos (re-search-forward "[\n\r]" head-end t))
957 (if (and pos
958 (= (char-after head-start) ?#)
959 (= (char-after (1+ head-start)) ?!))
960 ;; If the file begins with "#!" (exec interpreter magic),
961 ;; look for coding frobs in the first two lines. You cannot
962 ;; necessarily put them in the first line of such a file
963 ;; without screwing up the interpreter invocation.
964 (setq pos (search-forward "\n" head-end t)))
965 (if pos (setq head-end pos))
966 (when (< head-found head-end)
967 (goto-char head-start)
968 (when (and set-auto-coding-for-load
969 (re-search-forward
970 "-\\*-\\(.*;\\)?[ \t]*unibyte:[ \t]*\\([^ ;]+\\)"
971 head-end t))
972 (setq coding-system 'raw-text))
973 (when (and (not coding-system)
974 (re-search-forward
975 "-\\*-\\(.*;\\)?[ \t]*coding:[ \t]*\\([^ ;]+\\)"
976 head-end t))
977 (setq coding-system (intern (match-string 2)))
978 (or (coding-system-p coding-system)
979 (setq coding-system nil)))))
980
981 ;; If no coding: tag in the head, check the tail.
982 (when (and tail-found (not coding-system))
983 (goto-char tail-start)
984 (search-forward "\n\^L" nil t)
985 (if (re-search-forward
986 "^\\(.*\\)[ \t]*Local Variables:[ \t]*\\(.*\\)$" tail-end t)
987 ;; The prefix is what comes before "local variables:" in its
988 ;; line. The suffix is what comes after "local variables:"
989 ;; in its line.
990 (let* ((prefix (regexp-quote (match-string 1)))
991 (suffix (regexp-quote (match-string 2)))
992 (re-coding (concat
993 "^" prefix
994 "coding[ \t]*:[ \t]*\\([^ \t]+\\)[ \t]*"
995 suffix "$"))
996 (re-unibyte (concat
997 "^" prefix
998 "unibyte[ \t]*:[ \t]*\\([^ \t]+\\)[ \t]*"
999 suffix "$"))
1000 (re-end (concat
1001 "^" prefix "end *:[ \t]*" suffix "$"))
1002 (pos (point)))
1003 (re-search-forward re-end tail-end 'move)
1004 (setq tail-end (point))
1005 (goto-char pos)
1006 (when (and set-auto-coding-for-load
1007 (re-search-forward re-unibyte tail-end t))
1008 (setq coding-system 'raw-text))
1009 (when (and (not coding-system)
1010 (re-search-forward re-coding tail-end t))
1011 (setq coding-system (intern (match-string 1)))
1012 (or (coding-system-p coding-system)
1013 (setq coding-system nil))))))
1014 coding-system))))
63561304
KH
1015
1016(setq set-auto-coding-function 'set-auto-coding)
87aba788 1017
4ed46869
KH
1018;; Set buffer-file-coding-system of the current buffer after some text
1019;; is inserted.
1020(defun after-insert-file-set-buffer-file-coding-system (inserted)
1021 (if last-coding-system-used
1022 (let ((coding-system
1023 (find-new-buffer-file-coding-system last-coding-system-used))
1024 (modified-p (buffer-modified-p)))
0269ddfb
KH
1025 (when coding-system
1026 (set-buffer-file-coding-system coding-system)
d0c26c63
KH
1027 (if (and enable-multibyte-characters
1028 (or (eq coding-system 'no-conversion)
ccb77b4e
RS
1029 (eq (coding-system-type coding-system) 5))
1030 ;; If buffer was unmodified, we must be visiting it.
1031 (not modified-p))
1032 ;; For coding systems no-conversion and raw-text...,
1033 ;; edit the buffer as unibyte.
d0c26c63
KH
1034 (let ((pos-byte (position-bytes (+ (point) inserted))))
1035 (set-buffer-multibyte nil)
1036 (setq inserted (- pos-byte (position-bytes (point))))))
0269ddfb 1037 (set-buffer-modified-p modified-p))))
d0c26c63 1038 inserted)
4ed46869 1039
84c9d215
KH
1040(add-hook 'after-insert-file-functions
1041 'after-insert-file-set-buffer-file-coding-system)
4ed46869 1042
8057896b 1043;; The coding-spec and eol-type of coding-system returned is decided
4ed46869
KH
1044;; independently in the following order.
1045;; 1. That of buffer-file-coding-system locally bound.
1046;; 2. That of CODING.
1047
1048(defun find-new-buffer-file-coding-system (coding)
1049 "Return a coding system for a buffer when a file of CODING is inserted.
a73a8c89
KH
1050The local variable `buffer-file-coding-system' of the current buffer
1051is set to the returned value.
509064c5 1052Return nil if there's no need to set `buffer-file-coding-system'."
4ed46869 1053 (let (local-coding local-eol
b685f8d6 1054 found-coding found-eol
4ed46869
KH
1055 new-coding new-eol)
1056 (if (null coding)
1057 ;; Nothing found about coding.
1058 nil
1059
b685f8d6
RS
1060 ;; Get information of `buffer-file-coding-system' in LOCAL-EOL
1061 ;; and LOCAL-CODING.
1062 (setq local-eol (coding-system-eol-type buffer-file-coding-system))
1063 (if (null (numberp local-eol))
1064 ;; But eol-type is not yet set.
1065 (setq local-eol nil))
0269ddfb
KH
1066 (if (and buffer-file-coding-system
1067 (not (eq (coding-system-type buffer-file-coding-system) t)))
1068 ;; This is not `undecided'.
1069 (setq local-coding (coding-system-base buffer-file-coding-system)))
b685f8d6
RS
1070
1071 (if (and (local-variable-p 'buffer-file-coding-system)
1072 local-eol local-coding)
4ed46869
KH
1073 ;; The current buffer has already set full coding-system, we
1074 ;; had better not change it.
1075 nil
1076
8057896b 1077 (setq found-eol (coding-system-eol-type coding))
4ed46869 1078 (if (null (numberp found-eol))
be02cd54
EZ
1079 ;; But eol-type is not found.
1080 ;; If EOL conversions are inhibited, force unix eol-type.
1081 (setq found-eol (if inhibit-eol-conversion 0)))
c76b5c99
KH
1082 (if (eq (coding-system-type coding) t)
1083 (setq found-coding 'undecided)
1084 (setq found-coding (coding-system-base coding)))
1085
1086 (if (and (not found-eol) (eq found-coding 'undecided))
1087 ;; No valid coding information found.
1088 nil
1089
1090 ;; Some coding information (eol or text) found.
1091
1092 ;; The local setting takes precedence over the found one.
1093 (setq new-coding (if (local-variable-p 'buffer-file-coding-system)
1094 (or local-coding found-coding)
1095 (or found-coding local-coding)))
1096 (setq new-eol (if (local-variable-p 'buffer-file-coding-system)
1097 (or local-eol found-eol)
1098 (or found-eol local-eol)))
1099
1100 (let ((eol-type (coding-system-eol-type new-coding)))
1101 (if (and (numberp new-eol) (vectorp eol-type))
1102 (aref eol-type new-eol)
1103 new-coding)))))))
4ed46869 1104
fe831d33
GV
1105(defun modify-coding-system-alist (target-type regexp coding-system)
1106 "Modify one of look up tables for finding a coding system on I/O operation.
8c453b46
RS
1107There are three of such tables, `file-coding-system-alist',
1108`process-coding-system-alist', and `network-coding-system-alist'.
fe831d33
GV
1109
1110TARGET-TYPE specifies which of them to modify.
8c453b46
RS
1111If it is `file', it affects `file-coding-system-alist' (which see).
1112If it is `process', it affects `process-coding-system-alist' (which see).
e8dd0160 1113If it is `network', it affects `network-coding-system-alist' (which see).
fe831d33
GV
1114
1115REGEXP is a regular expression matching a target of I/O operation.
1116The target is a file name if TARGET-TYPE is `file', a program name if
1117TARGET-TYPE is `process', or a network service name or a port number
1118to connect to if TARGET-TYPE is `network'.
1119
1120CODING-SYSTEM is a coding system to perform code conversion on the I/O
8c453b46
RS
1121operation, or a cons cell (DECODING . ENCODING) specifying the coding systems
1122for decoding and encoding respectively,
1123or a function symbol which, when called, returns such a cons cell."
fe831d33
GV
1124 (or (memq target-type '(file process network))
1125 (error "Invalid target type: %s" target-type))
1126 (or (stringp regexp)
1127 (and (eq target-type 'network) (integerp regexp))
1128 (error "Invalid regular expression: %s" regexp))
1129 (if (symbolp coding-system)
1130 (if (not (fboundp coding-system))
1131 (progn
1132 (check-coding-system coding-system)
1133 (setq coding-system (cons coding-system coding-system))))
1134 (check-coding-system (car coding-system))
1135 (check-coding-system (cdr coding-system)))
1136 (cond ((eq target-type 'file)
1137 (let ((slot (assoc regexp file-coding-system-alist)))
1138 (if slot
1139 (setcdr slot coding-system)
1140 (setq file-coding-system-alist
1141 (cons (cons regexp coding-system)
1142 file-coding-system-alist)))))
1143 ((eq target-type 'process)
1144 (let ((slot (assoc regexp process-coding-system-alist)))
1145 (if slot
1146 (setcdr slot coding-system)
1147 (setq process-coding-system-alist
1148 (cons (cons regexp coding-system)
1149 process-coding-system-alist)))))
1150 (t
1151 (let ((slot (assoc regexp network-coding-system-alist)))
1152 (if slot
1153 (setcdr slot coding-system)
1154 (setq network-coding-system-alist
1155 (cons (cons regexp coding-system)
1156 network-coding-system-alist)))))))
1157
b25eef20 1158(defun make-translation-table (&rest args)
f967223b 1159 "Make a translation table (char table) from arguments.
13d5617d 1160Each argument is a list of the form (FROM . TO),
b25eef20 1161where FROM is a character to be translated to TO.
13d5617d 1162
452fdb31
RS
1163FROM can be a generic character (see `make-char'). In this case, TO is
1164a generic character containing the same number of characters, or a
1165ordinary character. If FROM and TO are both generic characters, all
b25eef20 1166characters belonging to FROM are translated to characters belonging to TO
13d5617d 1167without changing their position code(s)."
f967223b 1168 (let ((table (make-char-table 'translation-table))
a73a8c89
KH
1169 revlist)
1170 (while args
1171 (let ((elts (car args)))
1172 (while elts
13d5617d
KH
1173 (let* ((from (car (car elts)))
1174 (from-i 0) ; degree of freedom of FROM
1175 (from-rev (nreverse (split-char from)))
1176 (to (cdr (car elts)))
1177 (to-i 0) ; degree of freedom of TO
1178 (to-rev (nreverse (split-char to))))
1179 ;; Check numbers of heading 0s in FROM-REV and TO-REV.
1180 (while (eq (car from-rev) 0)
1181 (setq from-i (1+ from-i) from-rev (cdr from-rev)))
1182 (while (eq (car to-rev) 0)
1183 (setq to-i (1+ to-i) to-rev (cdr to-rev)))
1184 (if (and (/= from-i to-i) (/= to-i 0))
1185 (error "Invalid character pair (%d . %d)" from to))
b25eef20
KH
1186 ;; If we have already translated TO to TO-ALT, FROM should
1187 ;; also be translated to TO-ALT. But, this is only if TO
1188 ;; is a generic character or TO-ALT is not a generic
13d5617d
KH
1189 ;; character.
1190 (let ((to-alt (aref table to)))
1191 (if (and to-alt
1192 (or (> to-i 0) (not (generic-char-p to-alt))))
1193 (setq to to-alt)))
1194 (if (> from-i 0)
1195 (set-char-table-default table from to)
1196 (aset table from to))
b25eef20
KH
1197 ;; If we have already translated some chars to FROM, they
1198 ;; should also be translated to TO.
a73a8c89
KH
1199 (let ((l (assq from revlist)))
1200 (if l
1201 (let ((ch (car l)))
1202 (setcar l to)
1203 (setq l (cdr l))
1204 (while l
1205 (aset table ch to)
1206 (setq l (cdr l)) ))))
1207 ;; Now update REVLIST.
1208 (let ((l (assq to revlist)))
1209 (if l
1210 (setcdr l (cons from (cdr l)))
1211 (setq revlist (cons (list to from) revlist)))))
1212 (setq elts (cdr elts))))
1213 (setq args (cdr args)))
1214 ;; Return TABLE just created.
1215 table))
1216
c76b5c99
KH
1217(defun make-translation-table-from-vector (vec)
1218 "Make translation table from decoding vector VEC.
1219VEC is an array of 256 elements to map unibyte codes to multibyte characters.
1220See also the variable `nonascii-translation-table'."
1221 (let ((table (make-char-table 'translation-table))
1222 (rev-table (make-char-table 'translation-table))
1223 (i 0)
1224 ch)
1225 (while (< i 256)
1226 (setq ch (aref vec i))
1227 (aset table i ch)
1228 (if (>= ch 256)
1229 (aset rev-table ch i))
1230 (setq i (1+ i)))
1231 (set-char-table-extra-slot table 0 rev-table)
1232 table))
1233
f967223b 1234(defun define-translation-table (symbol &rest args)
e8dd0160 1235 "Define SYMBOL as a name of translation table made by ARGS.
b25eef20 1236
007c79c8
KH
1237If the first element of ARGS is a char-table of which purpose is
1238translation-table, just define SYMBOL as the name of it.
1239
1240In the other case, ARGS are the same as arguments to the function
1241`make-translation-table' (which see).
b25eef20 1242
452fdb31
RS
1243This function sets properties `translation-table' and
1244`translation-table-id' of SYMBOL to the created table itself and
f967223b 1245identification number of the table respectively."
007c79c8
KH
1246 (let ((table (if (and (char-table-p (car args))
1247 (eq (char-table-subtype (car args))
1248 'translation-table))
1249 (car args)
1250 (apply 'make-translation-table args)))
f967223b 1251 (len (length translation-table-vector))
d9e3229d 1252 (id 0)
b25eef20 1253 (done nil))
f967223b 1254 (put symbol 'translation-table table)
b25eef20
KH
1255 (while (not done)
1256 (if (>= id len)
f967223b
KH
1257 (setq translation-table-vector
1258 (vconcat translation-table-vector (make-vector len nil))))
1259 (let ((slot (aref translation-table-vector id)))
b25eef20
KH
1260 (if (or (not slot)
1261 (eq (car slot) symbol))
1262 (progn
f967223b 1263 (aset translation-table-vector id (cons symbol table))
007c79c8
KH
1264 (setq done t))
1265 (setq id (1+ id)))))
f967223b 1266 (put symbol 'translation-table-id id)
d9e3229d
KH
1267 id))
1268
69eba008
KH
1269;;; Initialize some variables.
1270
1271(put 'use-default-ascent 'char-table-extra-slots 0)
1272(setq use-default-ascent (make-char-table 'use-default-ascent))
d6d6d592
KH
1273(put 'ignore-relative-composition 'char-table-extra-slots 0)
1274(setq ignore-relative-composition
1275 (make-char-table 'ignore-relative-composition))
69eba008
KH
1276
1277;;;
4ed46869
KH
1278(provide 'mule)
1279
1280;;; mule.el ends here