1 ;;; mule.el --- basic commands for multilingual environment
3 ;; Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
4 ;; Free Software Foundation, Inc.
5 ;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
6 ;; 2005, 2006, 2007, 2008
7 ;; National Institute of Advanced Industrial Science and Technology (AIST)
8 ;; Registration Number H14PRO021
10 ;; National Institute of Advanced Industrial Science and Technology (AIST)
11 ;; Registration Number H13PRO009
13 ;; Keywords: mule, multilingual, character set, coding system
15 ;; This file is part of GNU Emacs.
17 ;; GNU Emacs is free software: you can redistribute it and/or modify
18 ;; it under the terms of the GNU General Public License as published by
19 ;; the Free Software Foundation, either version 3 of the License, or
20 ;; (at your option) any later version.
22 ;; GNU Emacs is distributed in the hope that it will be useful,
23 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
24 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
25 ;; GNU General Public License for more details.
27 ;; You should have received a copy of the GNU General Public License
28 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
34 (defconst mule-version
"6.0 (HANACHIRUSATO)" "\
35 Version number and name of this version of MULE (multilingual environment).")
37 (defconst mule-version-date
"2003.9.1" "\
38 Distribution date of this version of MULE (multilingual environment).")
43 ;; Backward compatibility code for handling emacs-mule charsets.
44 (defvar private-char-area-1-min
#xF0000
)
45 (defvar private-char-area-1-max
#xFFFFE
)
46 (defvar private-char-area-2-min
#x100000
)
47 (defvar private-char-area-2-max
#x10FFFE
)
49 ;; Table of emacs-mule charsets indexed by their emacs-mule ID.
50 (defvar emacs-mule-charset-table
(make-vector 256 nil
))
51 (aset emacs-mule-charset-table
0 'ascii
)
53 ;; Convert the argument of old-style calll of define-charset to a
54 ;; property list used by the new-style.
55 ;; INFO-VECTOR is a vector of the format:
56 ;; [DIMENSION CHARS WIDTH DIRECTION ISO-FINAL-CHAR ISO-GRAPHIC-PLANE
57 ;; SHORT-NAME LONG-NAME DESCRIPTION]
59 (defun convert-define-charset-argument (emacs-mule-id info-vector
)
60 (let* ((dim (aref info-vector
0))
61 (chars (aref info-vector
1))
62 (total (if (= dim
1) chars
(* chars chars
)))
63 (code-space (if (= dim
1) (if (= chars
96) [32 127] [33 126])
64 (if (= chars
96) [32 127 32 127] [33 126 33 126])))
66 (if (integerp emacs-mule-id
)
67 (or (= emacs-mule-id
0)
68 (and (>= emacs-mule-id
129) (< emacs-mule-id
256))
69 (error "Invalid CHARSET-ID: %d" emacs-mule-id
))
71 (if (= dim
1) (setq from-id
160 to-id
224)
72 (setq from-id
224 to-id
255))
73 (while (and (< from-id to-id
)
74 (not (aref emacs-mule-charset-table from-id
)))
75 (setq from-id
(1+ from-id
)))
77 (error "No more room for the new Emacs-mule charset"))
78 (setq emacs-mule-id from-id
)))
79 (if (> (- private-char-area-1-max private-char-area-1-min
) total
)
80 (setq code-offset private-char-area-1-min
81 private-char-area-1-min
(+ private-char-area-1-min total
))
82 (if (> (- private-char-area-2-max private-char-area-2-min
) total
)
83 (setq code-offset private-char-area-2-min
84 private-char-area-2-min
(+ private-char-area-2-min total
))
85 (error "No more space for a new charset")))
87 :code-space code-space
88 :iso-final-char
(aref info-vector
4)
89 :code-offset code-offset
90 :emacs-mule-id emacs-mule-id
)))
92 (defun define-charset (name docstring
&rest props
)
93 "Define NAME (symbol) as a charset with DOCSTRING.
94 The remaining arguments must come in pairs ATTRIBUTE VALUE. ATTRIBUTE
95 may be any symbol. The following have special meanings, and one of
96 `:code-offset', `:map', `:subset', `:superset' must be specified.
100 VALUE must be a short string to identify the charset. If omitted,
105 VALUE must be a string longer than `:short-name' to identify the
106 charset. If omitted, the value of the `:short-name' attribute is used.
110 VALUE must be an integer 0, 1, 2, or 3, specifying the dimension of
111 code-points of the charsets. If omitted, it is calculated from the
112 value of the `:code-space' attribute.
116 VALUE must be a vector of length at most 8 specifying the byte code
117 range of each dimension in this format:
118 [ MIN-1 MAX-1 MIN-2 MAX-2 ... ]
119 where MIN-N is the minimum byte value of Nth dimension of code-point,
120 MAX-N is the maximum byte value of that.
124 VALUE must be an integer specifying the mininum code point of the
125 charset. If omitted, it is calculated from `:code-space'. VALUE may
126 be a cons (HIGH . LOW), where HIGH is the most significant 16 bits of
127 the code point and LOW is the least significant 16 bits.
131 VALUE must be an integer specifying the maxinum code point of the
132 charset. If omitted, it is calculated from `:code-space'. VALUE may
133 be a cons (HIGH . LOW), where HIGH is the most significant 16 bits of
134 the code point and LOW is the least significant 16 bits.
138 VALUE must be a character in the range 32 to 127 (inclusive)
139 specifying the final char of the charset for ISO-2022 encoding. If
140 omitted, the charset can't be encoded by ISO-2022 based
143 `:iso-revision-number'
145 VALUE must be an integer in the range 0..63, specifying the revision
146 number of the charset for ISO-2022 encoding.
150 VALUE must be an integer of 0, 129..255. If omitted, the charset
151 can't be encoded by coding-systems of type `emacs-mule'.
153 `:ascii-compatible-p'
155 VALUE must be nil or t (default nil). If VALUE is t, the charset is
156 compatible with ASCII, i.e. the first 128 code points map to ASCII.
160 VALUE must be nil or t. If the VALUE is t, the charset is
161 supplementary, which means it is used only as a parent or a
162 subset of some other charset, or it is provided just for backward
167 VALUE must be a nonnegative integer that can be used as an invalid
168 code point of the charset. If the minimum code is 0 and the maximum
169 code is greater than Emacs' maximum integer value, `:invalid-code'
170 should not be omitted.
174 VALUE must be an integer added to the index number of a character to
175 get the corresponding character code.
179 VALUE must be vector or string.
181 If it is a vector, the format is [ CODE-1 CHAR-1 CODE-2 CHAR-2 ... ],
182 where CODE-n is a code-point of the charset, and CHAR-n is the
183 corresponding character code.
185 If it is a string, it is a name of file that contains the above
186 information. Each line of the file must be this format:
188 where XXX is a hexadecimal representation of CODE-n and YYY is a
189 hexadecimal representation of CHAR-n. A line starting with `#' is a
194 VALUE must be a list:
195 ( PARENT MIN-CODE MAX-CODE OFFSET )
196 PARENT is a parent charset. MIN-CODE and MAX-CODE specify the range
197 of characters inherited from the parent. OFFSET is an integer value
198 to add to a code point of the parent charset to get the corresponding
199 code point of this charset.
203 VALUE must be a list of parent charsets. The charset inherits
204 characters from them. Each element of the list may be a cons (PARENT
205 . OFFSET), where PARENT is a parent charset, and OFFSET is an offset
206 value to add to a code point of PARENT to get the corresponding code
207 point of this charset.
211 VALUE must be vector or string.
213 If it is a vector, the format is [ CODE-1 CHAR-1 CODE-2 CHAR-2 ... ],
214 where CODE-n is a code-point of the charset, and CHAR-n is the
215 corresponding Unicode character code.
217 If it is a string, it is a name of file that contains the above
218 information. The file format is the same as what described for `:map'
220 (when (vectorp (car props
))
222 ;; (define-charset CHARSET-ID CHARSET-SYMBOL INFO-VECTOR)
223 ;; Convert the argument to make it fit with the current style.
224 (let ((vec (car props
)))
225 (setq props
(convert-define-charset-argument name vec
)
227 docstring
(aref vec
8))))
228 (let ((attrs (mapcar 'list
'(:dimension
245 ;; If :dimension is omitted, get the dimension from :code-space.
246 (let ((dimension (plist-get props
:dimension
)))
248 (let ((code-space (plist-get props
:code-space
)))
249 (setq dimension
(if code-space
(/ (length code-space
) 2) 4))
250 (setq props
(plist-put props
:dimension dimension
)))))
252 (let ((code-space (plist-get props
:code-space
)))
254 (let ((dimension (plist-get props
:dimension
)))
255 (setq code-space
(make-vector 8 0))
256 (dotimes (i dimension
)
257 (aset code-space
(1+ (* i
2)) #xFF
))
258 (setq props
(plist-put props
:code-space code-space
)))))
260 ;; If :emacs-mule-id is specified, update emacs-mule-charset-table.
261 (let ((emacs-mule-id (plist-get props
:emacs-mule-id
)))
262 (if (integerp emacs-mule-id
)
263 (aset emacs-mule-charset-table emacs-mule-id name
)))
266 (setcdr slot
(plist-get props
(car slot
))))
268 ;; Make sure that the value of :code-space is a vector of 8
270 (let* ((slot (assq :code-space attrs
))
275 (vconcat val
(make-vector (- 8 len
) 0)))))
277 ;; Add :name and :docstring properties to PROPS.
279 (cons :name
(cons name
(cons :docstring
(cons docstring props
)))))
280 (or (plist-get props
:short-name
)
281 (plist-put props
:short-name
(symbol-name name
)))
282 (or (plist-get props
:long-name
)
283 (plist-put props
:long-name
(plist-get props
:short-name
)))
284 ;; We can probably get a worthwhile amount in purespace.
286 (mapcar (lambda (elt)
291 (setcdr (assq :plist attrs
) props
)
293 (apply 'define-charset-internal name
(mapcar 'cdr attrs
))))
296 (defun load-with-code-conversion (fullname file
&optional noerror nomessage
)
297 "Execute a file of Lisp code named FILE whose absolute name is FULLNAME.
298 The file contents are decoded before evaluation if necessary.
299 If optional third arg NOERROR is non-nil,
300 report no error if FILE doesn't exist.
301 Print messages at start and end of loading unless
302 optional fourth arg NOMESSAGE is non-nil.
303 Return t if file exists."
304 (if (null (file-readable-p fullname
))
306 (signal 'file-error
(list "Cannot open load file" file
)))
307 ;; Read file with code conversion, and then eval.
309 ;; To avoid any autoloading, set default-major-mode to
311 (let ((default-major-mode 'fundamental-mode
))
312 ;; We can't use `generate-new-buffer' because files.el
313 ;; is not yet loaded.
314 (get-buffer-create (generate-new-buffer-name " *load*"))))
316 (source (save-match-data (string-match "\\.el\\'" fullname
))))
319 (message "Loading %s (source)..." file
)
320 (message "Loading %s..." file
)))
322 (push file preloaded-file-list
))
324 (let ((load-file-name fullname
)
325 (set-auto-coding-for-load t
)
326 (inhibit-file-name-operation nil
))
327 (with-current-buffer buffer
328 ;; So that we don't get completely screwed if the
329 ;; file is encoded in some complicated character set,
330 ;; read it with real decoding, as a multibyte buffer,
331 ;; even if this is a --unibyte Emacs session.
332 (set-buffer-multibyte t
)
333 ;; Don't let deactivate-mark remain set.
334 (let (deactivate-mark)
335 (insert-file-contents fullname
))
336 ;; If the loaded file was inserted with no-conversion or
337 ;; raw-text coding system, make the buffer unibyte.
338 ;; Otherwise, eval-buffer might try to interpret random
339 ;; binary junk as multibyte characters.
340 (if (and enable-multibyte-characters
341 (or (eq (coding-system-type last-coding-system-used
)
343 (set-buffer-multibyte nil
))
344 ;; Make `kill-buffer' quiet.
345 (set-buffer-modified-p nil
))
346 ;; Have the original buffer current while we eval.
347 (eval-buffer buffer nil
348 ;; This is compatible with what `load' does.
349 (if purify-flag file fullname
)
350 ;; If this Emacs is running with --unibyte,
351 ;; convert multibyte strings to unibyte
352 ;; after reading them.
353 ;; (not default-enable-multibyte-characters)
356 (let (kill-buffer-hook kill-buffer-query-functions
)
357 (kill-buffer buffer
)))
359 (do-after-load-evaluation fullname
))
361 (unless (or nomessage noninteractive
)
363 (message "Loading %s (source)...done" file
)
364 (message "Loading %s...done" file
)))
367 (defun charset-info (charset)
368 "Return a vector of information of CHARSET.
369 This function is provided for backward compatibility.
371 The elements of the vector are:
372 CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION,
373 LEADING-CODE-BASE, LEADING-CODE-EXT,
374 ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE,
375 REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION,
378 CHARSET-ID is always 0.
380 DIMENSION is the number of bytes of a code-point of the charset:
382 CHARS is the number of characters in a dimension:
385 DIRECTION is always 0.
386 LEADING-CODE-BASE is always 0.
387 LEADING-CODE-EXT is always 0.
388 ISO-FINAL-CHAR (character) is the final character of the
389 corresponding ISO 2022 charset. If the charset is not assigned
390 any final character, the value is -1.
391 ISO-GRAPHIC-PLANE is always 0.
392 REVERSE-CHARSET is always -1.
393 SHORT-NAME (string) is the short name to refer to the charset.
394 LONG-NAME (string) is the long name to refer to the charset
395 DESCRIPTION (string) is the description string of the charset.
396 PLIST (property list) may contain any type of information a user
397 want to put and get by functions `put-charset-property' and
398 `get-charset-property' respectively."
401 (charset-dimension charset
)
402 (charset-chars charset
)
407 (charset-iso-final-char charset
)
410 (get-charset-property charset
:short-name
)
411 (get-charset-property charset
:short-name
)
412 (charset-description charset
)
413 (charset-plist charset
)))
415 ;; It is better not to use backquote in this file,
416 ;; because that makes a bootstrapping problem
417 ;; if you need to recompile all the Lisp files using interpreted code.
419 (defun charset-id (charset)
420 "Always return 0. This is provided for backward compatibility."
423 (defmacro charset-bytes
(charset)
424 "Always return 0. This is provided for backward compatibility."
427 (defun get-charset-property (charset propname
)
428 "Return the value of CHARSET's PROPNAME property.
429 This is the last value stored with
430 (put-charset-property CHARSET PROPNAME VALUE)."
431 (plist-get (charset-plist charset
) propname
))
433 (defun put-charset-property (charset propname value
)
434 "Set CHARSETS's PROPNAME property to value VALUE.
435 It can be retrieved with `(get-charset-property CHARSET PROPNAME)'."
436 (set-charset-plist charset
437 (plist-put (charset-plist charset
) propname value
)))
439 (defun charset-description (charset)
440 "Return description string of CHARSET."
441 (plist-get (charset-plist charset
) :docstring
))
443 (defun charset-dimension (charset)
444 "Return dimension of CHARSET."
445 (plist-get (charset-plist charset
) :dimension
))
447 (defun charset-chars (charset &optional dimension
)
448 "Return number of characters contained in DIMENSION of CHARSET.
449 DIMENSION defaults to the first dimension."
450 (unless dimension
(setq dimension
1))
451 (let ((code-space (plist-get (charset-plist charset
) :code-space
)))
452 (1+ (- (aref code-space
(1- (* 2 dimension
)))
453 (aref code-space
(- (* 2 dimension
) 2))))))
455 (defun charset-iso-final-char (charset)
456 "Return ISO-2022 final character of CHARSET.
457 Return -1 if charset isn't an ISO 2022 one."
458 (or (plist-get (charset-plist charset
) :iso-final-char
)
461 (defmacro charset-short-name
(charset)
462 "Return short name of CHARSET."
463 (plist-get (charset-plist charset
) :short-name
))
465 (defmacro charset-long-name
(charset)
466 "Return long name of CHARSET."
467 (plist-get (charset-plist charset
) :long-name
))
469 (defun charset-list ()
470 "Return list of all charsets ever defined.
472 This function is provided for backward compatibility.
473 Now we have the variable `charset-list'."
475 (make-obsolete 'charset-list
"use variable `charset-list'." "23.1")
479 (defalias 'char-valid-p
'characterp
)
480 (make-obsolete 'char-valid-p
'characterp
"23.1")
482 (defun generic-char-p (char)
483 "Always return nil. This is provided for backward compatibility."
485 (make-obsolete 'generic-char-p
"generic characters no longer exist." "23.1")
487 (defun make-char-internal (charset-id &optional code1 code2
)
488 (let ((charset (aref emacs-mule-charset-table charset-id
)))
490 (error "Invalid Emacs-mule charset ID: %d" charset-id
))
491 (make-char charset code1 code2
)))
493 ;; Save the ASCII case table in case we need it later. Some locales
494 ;; (such as Turkish) modify the case behavior of ASCII characters,
495 ;; which can interfere with networking code that uses ASCII strings.
497 (defvar ascii-case-table
498 ;; Code copied from copy-case-table to avoid requiring case-table.el
499 (let ((tbl (copy-sequence (standard-case-table)))
500 (up (char-table-extra-slot (standard-case-table) 0)))
501 (if up
(set-char-table-extra-slot tbl
0 (copy-sequence up
)))
502 (set-char-table-extra-slot tbl
1 nil
)
503 (set-char-table-extra-slot tbl
2 nil
)
505 "Case table for the ASCII character set.")
507 ;; Coding system stuff
509 ;; Coding system is a symbol that has been defined by the function
510 ;; `define-coding-system'.
512 (defconst coding-system-iso-2022-flags
530 "List of symbols that control ISO-2022 encoder/decoder.
532 The value of the `:flags' attribute in the argument of the function
533 `define-coding-system' must be one of them.
535 If `long-form' is specified, use a long designation sequence on
536 encoding for the charsets `japanese-jisx0208-1978', `chinese-gb2312',
537 and `japanese-jisx0208'. The long designation sequence doesn't
538 conform to ISO 2022, but is used by such coding systems as
541 If `ascii-at-eol' is specified, designate ASCII to g0 at end of line
544 If `ascii-at-cntl' is specified, designate ASCII to g0 before control
545 codes and SPC on encoding.
547 If `7-bit' is specified, use 7-bit code only on encoding.
549 If `locking-shift' is specified, decode locking-shift code correctly
550 on decoding, and use locking-shift to invoke a graphic element on
553 If `single-shift' is specified, decode single-shift code correctly on
554 decoding, and use single-shift to invoke a graphic element on encoding.
556 If `designation' is specified, decode designation code correctly on
557 decoding, and use designation to designate a charset to a graphic
560 If `revision' is specified, produce an escape sequence to specify
561 revision number of a charset on encoding. Such an escape sequence is
562 always correctly decoded on decoding.
564 If `direction' is specified, decode ISO6429's code for specifying
565 direction correctly, and produce the code on encoding.
567 If `init-at-bol' is specified, on encoding, it is assumed that
568 invocation and designation statuses are reset at each beginning of
569 line even if `ascii-at-eol' is not specified; thus no codes for
570 resetting them are produced.
572 If `safe' is specified, on encoding, characters not supported by a
573 coding are replaced with `?'.
575 If `latin-extra' is specified, the code-detection routine assumes that a
576 code specified in `latin-extra-code-table' (which see) is valid.
578 If `composition' is specified, an escape sequence to specify
579 composition sequence is correctly decoded on decoding, and is produced
582 If `euc-tw-shift' is specified, the EUC-TW specific shifting code is
583 correctly decoded on decoding, and is produced on encoding.
585 If `use-roman' is specified, JIS0201-1976-Roman is designated instead
588 If `use-oldjis' is specified, JIS0208-1976 is designated instead of
591 (defun define-coding-system (name docstring
&rest props
)
592 "Define NAME (a symbol) as a coding system with DOCSTRING and attributes.
593 The remaining arguments must come in pairs ATTRIBUTE VALUE. ATTRIBUTE
596 The following attributes have special meanings. Those labeled as
597 \"(required)\", should not be omitted.
599 `:mnemonic' (required)
601 VALUE is a character to display on mode line for the coding system.
603 `:coding-type' (required)
605 VALUE must be one of `charset', `utf-8', `utf-16', `iso-2022',
606 `emacs-mule', `shift-jis', `ccl', `raw-text', `undecided'.
610 VALUE is the EOL (end-of-line) format of the coding system. It must be
611 one of `unix', `dos', `mac'. The symbol `unix' means Unix-like EOL
612 \(i.e. single LF), `dos' means DOS-like EOL \(i.e. sequence of CR LF),
613 and `mac' means MAC-like EOL \(i.e. single CR). If omitted, on
614 decoding by the coding system, Emacs automatically detects the EOL
615 format of the source text.
619 VALUE must be a list of charsets supported by the coding system. On
620 encoding by the coding system, if a character belongs to multiple
621 charsets in the list, a charset that comes earlier in the list is
622 selected. If `:coding-type' is `iso-2022', VALUE may be `iso-2022',
623 which indicates that the coding system supports all ISO-2022 based
624 charsets. If `:coding-type' is `emacs-mule', VALUE may be
625 `emacs-mule', which indicates that the coding system supports all
626 charsets that have the `:emacs-mule-id' property.
628 `:ascii-compatible-p'
630 If VALUE is non-nil, the coding system decodes all 7-bit bytes into
631 the corresponding ASCII characters, and encodes all ASCII characters
632 back to the corresponding 7-bit bytes. VALUE defaults to nil.
634 `:decode-translation-table'
636 VALUE must be a translation table to use on decoding.
638 `:encode-translation-table'
640 VALUE must be a translation table to use on encoding.
642 `:post-read-conversion'
644 VALUE must be a function to call after some text is inserted and
645 decoded by the coding system itself and before any functions in
646 `after-insert-functions' are called. The arguments to this function
647 are the same as those of a function in `after-insert-file-functions',
648 i.e. LENGTH of the text to be decoded with point at the head of it,
649 and the function should leave point unchanged.
651 `:pre-write-conversion'
653 VALUE must be a function to call after all functions in
654 `write-region-annotate-functions' and `buffer-file-format' are called,
655 and before the text is encoded by the coding system itself. The
656 arguments to this function are the same as those of a function in
657 `write-region-annotate-functions'.
661 VALUE must be a character. On encoding, a character not supported by
662 the coding system is replaced with VALUE.
666 VALUE non-nil means that visiting a file with the coding system
667 results in a unibyte buffer.
671 VALUE must be `unix', `dos', `mac'. The symbol `unix' means Unix-like
672 EOL (LF), `dos' means DOS-like EOL (CRLF), and `mac' means MAC-like
673 EOL (CR). If omitted, on decoding, the coding system detects EOL
674 format automatically, and on encoding, uses Unix-like EOL.
678 VALUE must be a symbol whose name is that of a MIME charset converted
681 `:mime-text-unsuitable'
683 VALUE non-nil means the `:mime-charset' property names a charset which
684 is unsuitable for the top-level media type \"text\".
688 VALUE must be a list of symbols that control the ISO-2022 converter.
689 Each must be a member of the list `coding-system-iso-2022-flags'
690 \(which see). This attribute has a meaning only when `:coding-type'
695 VALUE must be a vector [G0-USAGE G1-USAGE G2-USAGE G3-USAGE].
696 GN-USAGE specifies the usage of graphic register GN as follows.
698 If it is nil, no charset can be designated to GN.
700 If it is a charset, the charset is initially designated to GN, and
701 never used by the other charsets.
703 If it is a list, the elements must be charsets, nil, 94, or 96. GN
704 can be used by all the listed charsets. If the list contains 94, any
705 iso-2022 charset whose code-space ranges are 94 long can be designated
706 to GN. If the list contains 96, any charsets whose whose ranges are
707 96 long can be designated to GN. If the first element is a charset,
708 that charset is initially designated to GN.
710 This attribute has a meaning only when `:coding-type' is `iso-2022'.
714 This attributes specifies whether the coding system uses a `byte order
715 mark'. VALUE must nil, t, or cons of coding systems whose
716 `:coding-type' is `utf-16'.
718 If the value is nil, on decoding, don't treat the first two-byte as
719 BOM, and on encoding, don't produce BOM bytes.
721 If the value is t, on decoding, skip the first two-byte as BOM, and on
722 encoding, produce BOM bytes accoding to the value of `:endian'.
724 If the value is cons, on decoding, check the first two-byte. If theyq
725 are 0xFE 0xFF, use the car part coding system of the value. If they
726 are 0xFF 0xFE, use the car part coding system of the value.
727 Otherwise, treat them as bytes for a normal character. On encoding,
728 produce BOM bytes accoding to the value of `:endian'.
730 This attribute has a meaning only when `:coding-type' is `utf-16'.
734 VALUE must be `big' or `little' specifying big-endian and
735 little-endian respectively. The default value is `big'.
737 This attribute has a meaning only when `:coding-type' is `utf-16'.
741 VALUE is a symbol representing the registered CCL program used for
742 decoding. This attribute has a meaning only when `:coding-type' is
747 VALUE is a symbol representing the registered CCL program used for
748 encoding. This attribute has a meaning only when `:coding-type' is
750 (let* ((common-attrs (mapcar 'list
755 :decode-translation-table
756 :encode-translation-table
757 :post-read-conversion
758 :pre-write-conversion
763 (coding-type (plist-get props
:coding-type
))
764 (spec-attrs (mapcar 'list
765 (cond ((eq coding-type
'iso-2022
)
770 ((eq coding-type
'utf-8
)
772 ((eq coding-type
'utf-16
)
775 ((eq coding-type
'ccl
)
780 (dolist (slot common-attrs
)
781 (setcdr slot
(plist-get props
(car slot
))))
783 (dolist (slot spec-attrs
)
784 (setcdr slot
(plist-get props
(car slot
))))
786 (if (eq coding-type
'iso-2022
)
787 (let ((designation (plist-get props
:designation
))
788 (flags (plist-get props
:flags
))
789 (initial (make-vector 4 nil
))
790 (reg-usage (cons 4 4))
793 (setq elt
(aref designation i
))
794 (cond ((charsetp elt
)
796 (setq request
(cons (cons elt i
) request
)))
798 (aset initial i
(car elt
))
799 (if (charsetp (car elt
))
800 (setq request
(cons (cons (car elt
) i
) request
)))
801 (dolist (e (cdr elt
))
803 (setq request
(cons (cons e i
) request
)))
805 (setcar reg-usage i
))
807 (setcdr reg-usage i
))
810 (setcdr reg-usage i
)))))))
811 (setcdr (assq :initial spec-attrs
) initial
)
812 (setcdr (assq :reg-usage spec-attrs
) reg-usage
)
813 (setcdr (assq :request spec-attrs
) request
)
815 ;; Change :flags value from a list to a bit-mask.
818 (dolist (elt coding-system-iso-2022-flags
)
820 (setq bits
(logior bits
(lsh 1 i
))))
822 (setcdr (assq :flags spec-attrs
) bits
))))
824 ;; Add :name and :docstring properties to PROPS.
826 (cons :name
(cons name
(cons :docstring
(cons (purecopy docstring
)
828 (setcdr (assq :plist common-attrs
) props
)
829 (apply 'define-coding-system-internal
830 name
(mapcar 'cdr
(append common-attrs spec-attrs
)))))
832 (defun coding-system-doc-string (coding-system)
833 "Return the documentation string for CODING-SYSTEM."
834 (plist-get (coding-system-plist coding-system
) :docstring
))
836 (defun coding-system-mnemonic (coding-system)
837 "Return the mnemonic character of CODING-SYSTEM.
838 The mnemonic character of a coding system is used in mode line to
839 indicate the coding system. If CODING-SYSTEM is nil, return ?=."
840 (plist-get (coding-system-plist coding-system
) :mnemonic
))
842 (defun coding-system-type (coding-system)
843 "Return the coding type of CODING-SYSTEM.
844 A coding type is a symbol indicating the encoding method of CODING-SYSTEM.
845 See the function `define-coding-system' for more detail."
846 (plist-get (coding-system-plist coding-system
) :coding-type
))
848 (defun coding-system-charset-list (coding-system)
849 "Return list of charsets supported by CODING-SYSTEM.
850 If CODING-SYSTEM supports all ISO-2022 charsets, return `iso-2022'.
851 If CODING-SYSTEM supports all emacs-mule charsets, return `emacs-mule'."
852 (plist-get (coding-system-plist coding-system
) :charset-list
))
854 (defun coding-system-category (coding-system)
855 "Return a category symbol of CODING-SYSTEM."
856 (plist-get (coding-system-plist coding-system
) :category
))
858 (defun coding-system-get (coding-system prop
)
859 "Extract a value from CODING-SYSTEM's property list for property PROP.
860 For compatibility with Emacs 20/21, this accepts old-style symbols
861 like `mime-charset' as well as the current style like `:mime-charset'."
862 (or (plist-get (coding-system-plist coding-system
) prop
)
863 (if (not (keywordp prop
))
864 ;; For backward compatiblity.
865 (if (eq prop
'ascii-incompatible
)
866 (not (plist-get (coding-system-plist coding-system
)
867 :ascii-compatible-p
))
868 (plist-get (coding-system-plist coding-system
)
869 (intern (concat ":" (symbol-name prop
))))))))
871 (defun coding-system-eol-type-mnemonic (coding-system)
872 "Return the string indicating end-of-line format of CODING-SYSTEM."
873 (let* ((eol-type (coding-system-eol-type coding-system
))
874 (val (cond ((eq eol-type
0) eol-mnemonic-unix
)
875 ((eq eol-type
1) eol-mnemonic-dos
)
876 ((eq eol-type
2) eol-mnemonic-mac
)
877 (t eol-mnemonic-undecided
))))
880 (char-to-string val
))))
882 (defun coding-system-lessp (x y
)
883 (cond ((eq x
'no-conversion
) t
)
884 ((eq y
'no-conversion
) nil
)
885 ((eq x
'emacs-mule
) t
)
886 ((eq y
'emacs-mule
) nil
)
887 ((eq x
'undecided
) t
)
888 ((eq y
'undecided
) nil
)
889 (t (let ((c1 (coding-system-mnemonic x
))
890 (c2 (coding-system-mnemonic y
)))
891 (or (< (downcase c1
) (downcase c2
))
892 (and (not (> (downcase c1
) (downcase c2
)))
895 (defun coding-system-equal (coding-system-1 coding-system-2
)
896 "Return t if and only if CODING-SYSTEM-1 and CODING-SYSTEM-2 are identical.
897 Two coding systems are identical if two symbols are equal
898 or one is an alias of the other."
899 (or (eq coding-system-1 coding-system-2
)
900 (and (equal (coding-system-plist coding-system-1
)
901 (coding-system-plist coding-system-2
))
902 (let ((eol-type-1 (coding-system-eol-type coding-system-1
))
903 (eol-type-2 (coding-system-eol-type coding-system-2
)))
904 (or (eq eol-type-1 eol-type-2
)
905 (and (vectorp eol-type-1
) (vectorp eol-type-2
)))))))
907 (defun add-to-coding-system-list (coding-system)
908 "Add CODING-SYSTEM to `coding-system-list' while keeping it sorted."
909 (if (or (null coding-system-list
)
910 (coding-system-lessp coding-system
(car coding-system-list
)))
911 (setq coding-system-list
(cons coding-system coding-system-list
))
912 (let ((len (length coding-system-list
))
913 mid
(tem coding-system-list
))
915 (setq mid
(nthcdr (/ len
2) tem
))
916 (if (coding-system-lessp (car mid
) coding-system
)
918 len
(- len
(/ len
2)))
919 (setq len
(/ len
2))))
920 (setcdr tem
(cons coding-system
(cdr tem
))))))
922 (defun coding-system-list (&optional base-only
)
923 "Return a list of all existing non-subsidiary coding systems.
924 If optional arg BASE-ONLY is non-nil, only base coding systems are
925 listed. The value doesn't include subsidiary coding systems which are
926 made from bases and aliases automatically for various end-of-line
927 formats (e.g. iso-latin-1-unix, koi8-r-dos)."
929 (dolist (coding coding-system-list
)
930 (if (eq (coding-system-base coding
) coding
)
932 (setq codings
(cons coding codings
))
933 (dolist (alias (coding-system-aliases coding
))
934 (setq codings
(cons alias codings
))))))
937 (defconst char-coding-system-table nil
938 "It exists just for backward compatibility, and the value is always nil.")
939 (make-obsolete-variable 'char-coding-system-table nil
"23.1")
941 (defun transform-make-coding-system-args (name type
&optional doc-string props
)
942 "For internal use only.
943 Transform XEmacs style args for `make-coding-system' to Emacs style.
944 Value is a list of transformed arguments."
945 (let ((mnemonic (string-to-char (or (plist-get props
'mnemonic
) "?")))
946 (eol-type (plist-get props
'eol-type
))
949 ((eq eol-type
'lf
) (setq eol-type
'unix
))
950 ((eq eol-type
'crlf
) (setq eol-type
'dos
))
951 ((eq eol-type
'cr
) (setq eol-type
'mac
)))
952 (if (setq tmp
(plist-get props
'post-read-conversion
))
953 (setq properties
(plist-put properties
'post-read-conversion tmp
)))
954 (if (setq tmp
(plist-get props
'pre-write-conversion
))
955 (setq properties
(plist-put properties
'pre-write-conversion tmp
)))
957 ((eq type
'shift-jis
)
958 `(,name
1 ,mnemonic
,doc-string
() ,properties
,eol-type
))
959 ((eq type
'iso2022
) ; This is not perfect.
960 (if (plist-get props
'escape-quoted
)
961 (error "escape-quoted is not supported: %S"
962 `(,name
,type
,doc-string
,props
)))
963 (let ((g0 (plist-get props
'charset-g0
))
964 (g1 (plist-get props
'charset-g1
))
965 (g2 (plist-get props
'charset-g2
))
966 (g3 (plist-get props
'charset-g3
))
969 (eq (cadr (assoc 'latin-jisx0201
970 (plist-get props
'input-charset-conversion
)))
972 (eq (cadr (assoc 'ascii
973 (plist-get props
'output-charset-conversion
)))
977 (eq (cadr (assoc 'japanese-jisx0208-1978
978 (plist-get props
'input-charset-conversion
)))
980 (eq (cadr (assoc 'japanese-jisx0208
981 (plist-get props
'output-charset-conversion
)))
982 'japanese-jisx0208-1978
))))
984 (if (plist-get props
'force-g0-on-output
)
988 (if (plist-get props
'force-g1-on-output
)
992 (if (plist-get props
'force-g2-on-output
)
996 (if (plist-get props
'force-g3-on-output
)
999 `(,name
2 ,mnemonic
,doc-string
1001 ,(plist-get props
'short
)
1002 ,(not (plist-get props
'no-ascii-eol
))
1003 ,(not (plist-get props
'no-ascii-cntl
))
1004 ,(plist-get props
'seven
)
1006 ,(not (plist-get props
'lock-shift
))
1009 ,(plist-get props
'no-iso6429
)
1011 ,properties
,eol-type
)))
1013 `(,name
3 ,mnemonic
,doc-string
() ,properties
,eol-type
))
1015 `(,name
4 ,mnemonic
,doc-string
1016 (,(plist-get props
'decode
) .
,(plist-get props
'encode
))
1017 ,properties
,eol-type
))
1019 (error "unsupported XEmacs style make-coding-style arguments: %S"
1020 `(,name
,type
,doc-string
,props
))))))
1022 (defun make-coding-system (coding-system type mnemonic doc-string
1027 "Define a new coding system CODING-SYSTEM (symbol).
1028 This function is provided for backward compatibility.
1029 Use `define-coding-system' instead."
1030 ;; For compatiblity with XEmacs, we check the type of TYPE. If it
1031 ;; is a symbol, perhaps, this function is called with XEmacs-style
1032 ;; arguments. Here, try to transform that kind of arguments to
1035 (let ((args (transform-make-coding-system-args coding-system type
1036 mnemonic doc-string
)))
1037 (setq coding-system
(car args
)
1039 mnemonic
(nth 2 args
)
1040 doc-string
(nth 3 args
)
1042 properties
(nth 5 args
)
1043 eol-type
(nth 6 args
))))
1046 (cond ((eq type
0) 'emacs-mule
)
1047 ((eq type
1) 'shift-jis
)
1048 ((eq type
2) 'iso2022
)
1051 ((eq type
5) 'raw-text
)
1053 (error "Invalid coding system type: %s" type
))))
1056 (let ((plist nil
) key
)
1057 (dolist (elt properties
)
1058 (setq key
(car elt
))
1059 (cond ((eq key
'post-read-conversion
)
1060 (setq key
:post-read-conversion
))
1061 ((eq key
'pre-write-conversion
)
1062 (setq key
:pre-write-conversion
))
1063 ((eq key
'translation-table-for-decode
)
1064 (setq key
:decode-translation-table
))
1065 ((eq key
'translation-table-for-encode
)
1066 (setq key
:encode-translation-table
))
1067 ((eq key
'safe-charsets
)
1068 (setq key
:charset-list
))
1069 ((eq key
'mime-charset
)
1070 (setq key
:mime-charset
))
1071 ((eq key
'valid-codes
)
1072 (setq key
:valids
)))
1073 (setq plist
(plist-put plist key
(cdr elt
))))
1075 (setq properties
(plist-put properties
:mnemonic mnemonic
))
1076 (plist-put properties
:coding-type type
)
1077 (cond ((eq eol-type
0) (setq eol-type
'unix
))
1078 ((eq eol-type
1) (setq eol-type
'dos
))
1079 ((eq eol-type
2) (setq eol-type
'mac
))
1080 ((vectorp eol-type
) (setq eol-type nil
)))
1081 (plist-put properties
:eol-type eol-type
)
1085 (plist-put properties
:flags
1086 (list (and (or (consp (nth 0 flags
))
1087 (consp (nth 1 flags
))
1088 (consp (nth 2 flags
))
1089 (consp (nth 3 flags
))) 'designation
)
1090 (or (nth 4 flags
) 'long-form
)
1091 (and (nth 5 flags
) 'ascii-at-eol
)
1092 (and (nth 6 flags
) 'ascii-at-cntl
)
1093 (and (nth 7 flags
) '7-bit
)
1094 (and (nth 8 flags
) 'locking-shift
)
1095 (and (nth 9 flags
) 'single-shift
)
1096 (and (nth 10 flags
) 'use-roman
)
1097 (and (nth 11 flags
) 'use-oldjis
)
1098 (or (nth 12 flags
) 'direction
)
1099 (and (nth 13 flags
) 'init-at-bol
)
1100 (and (nth 14 flags
) 'designate-at-bol
)
1101 (and (nth 15 flags
) 'safe
)
1102 (and (nth 16 flags
) 'latin-extra
)))
1103 (plist-put properties
:designation
1104 (let ((vec (make-vector 4 nil
)))
1106 (let ((spec (nth i flags
)))
1108 (aset vec i
'(94 96))
1112 (setq spec
(append (delq t spec
) '(94 96))))
1113 (aset vec i spec
))))))
1117 (plist-put properties
:ccl-decoder
(car flags
))
1118 (plist-put properties
:ccl-encoder
(cdr flags
))))
1120 (apply 'define-coding-system coding-system doc-string properties
))
1122 (defun merge-coding-systems (first second
)
1123 "Fill in any unspecified aspects of coding system FIRST from SECOND.
1124 Return the resulting coding system."
1125 (let ((base (coding-system-base second
))
1126 (eol (coding-system-eol-type second
)))
1127 ;; If FIRST doesn't specify text conversion, merge with that of SECOND.
1128 (if (eq (coding-system-base first
) 'undecided
)
1129 (setq first
(coding-system-change-text-conversion first base
)))
1130 ;; If FIRST doesn't specify eol conversion, merge with that of SECOND.
1131 (if (and (vectorp (coding-system-eol-type first
))
1132 (numberp eol
) (>= eol
0) (<= eol
2))
1133 (setq first
(coding-system-change-eol-conversion
1137 (defun autoload-coding-system (symbol form
)
1138 "Define SYMBOL as a coding-system that is defined on demand.
1140 FROM is a form to evaluate to define the coding-system."
1141 (put symbol
'coding-system-define-form form
)
1142 (setq coding-system-alist
(cons (list (symbol-name symbol
))
1143 coding-system-alist
))
1144 (dolist (elt '("-unix" "-dos" "-mac"))
1145 (let ((name (concat (symbol-name symbol
) elt
)))
1146 (put (intern name
) 'coding-system-define-form form
)
1147 (setq coding-system-alist
(cons (list name
) coding-system-alist
)))))
1149 (defun set-buffer-file-coding-system (coding-system &optional force nomodify
)
1150 "Set the file coding-system of the current buffer to CODING-SYSTEM.
1151 This means that when you save the buffer, it will be converted
1152 according to CODING-SYSTEM. For a list of possible values of CODING-SYSTEM,
1153 use \\[list-coding-systems].
1155 If CODING-SYSTEM leaves the text conversion unspecified, or if it
1156 leaves the end-of-line conversion unspecified, FORCE controls what to
1157 do. If FORCE is nil, get the unspecified aspect (or aspects) from the
1158 buffer's previous `buffer-file-coding-system' value (if it is
1159 specified there). Otherwise, leave it unspecified.
1161 This marks the buffer modified so that the succeeding \\[save-buffer]
1162 surely saves the buffer with CODING-SYSTEM. From a program, if you
1163 don't want to mark the buffer modified, specify t for NOMODIFY.
1164 If you know exactly what coding system you want to use,
1165 just set the variable `buffer-file-coding-system' directly."
1166 (interactive "zCoding system for saving file (default nil): \nP")
1167 (check-coding-system coding-system
)
1168 (if (and coding-system buffer-file-coding-system
(null force
))
1170 (merge-coding-systems coding-system buffer-file-coding-system
)))
1171 (setq buffer-file-coding-system coding-system
)
1172 ;; This is in case of an explicit call. Normally, `normal-mode' and
1173 ;; `set-buffer-major-mode-hook' take care of setting the table.
1174 (if (fboundp 'ucs-set-table-for-input
) ; don't lose when building
1175 (ucs-set-table-for-input))
1177 (set-buffer-modified-p t
))
1178 (force-mode-line-update))
1180 (defun revert-buffer-with-coding-system (coding-system &optional force
)
1181 "Visit the current buffer's file again using coding system CODING-SYSTEM.
1182 For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems].
1184 If CODING-SYSTEM leaves the text conversion unspecified, or if it
1185 leaves the end-of-line conversion unspecified, FORCE controls what to
1186 do. If FORCE is nil, get the unspecified aspect (or aspects) from the
1187 buffer's previous `buffer-file-coding-system' value (if it is
1188 specified there). Otherwise, determine it from the file contents as
1189 usual for visiting a file."
1190 (interactive "zCoding system for visited file (default nil): \nP")
1191 (check-coding-system coding-system
)
1192 (if (and coding-system buffer-file-coding-system
(null force
))
1194 (merge-coding-systems coding-system buffer-file-coding-system
)))
1195 (let ((coding-system-for-read coding-system
))
1198 (defun set-file-name-coding-system (coding-system)
1199 "Set coding system for decoding and encoding file names to CODING-SYSTEM.
1200 It actually just set the variable `file-name-coding-system' (which
1201 see) to CODING-SYSTEM."
1202 (interactive "zCoding system for file names (default nil): ")
1203 (check-coding-system coding-system
)
1204 (if (and coding-system
1205 (not (coding-system-get coding-system
:ascii-compatible-p
))
1206 (not (coding-system-get coding-system
:suitable-for-file-name
)))
1207 (error "%s is not suitable for file names" coding-system
))
1208 (setq file-name-coding-system coding-system
))
1210 (defvar default-terminal-coding-system nil
1211 "Default value for the terminal coding system.
1212 This is normally set according to the selected language environment.
1213 See also the command `set-terminal-coding-system'.")
1215 (defun set-terminal-coding-system (coding-system &optional display
)
1216 "Set coding system of terminal output to CODING-SYSTEM.
1217 All text output to DISPLAY will be encoded
1218 with the specified coding system.
1220 For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems].
1221 The default is determined by the selected language environment
1222 or by the previous use of this command.
1224 DISPLAY may be a display id, a frame, or nil for the selected frame's display.
1225 The setting has no effect on graphical displays."
1227 (list (let ((default (if (and (not (terminal-coding-system))
1228 default-terminal-coding-system
)
1229 default-terminal-coding-system
)))
1231 (format "Coding system for terminal display (default %s): "
1234 (if (and (not coding-system
)
1235 (not (terminal-coding-system)))
1236 (setq coding-system default-terminal-coding-system
))
1238 (setq default-terminal-coding-system coding-system
))
1239 (set-terminal-coding-system-internal coding-system display
)
1240 (redraw-frame (selected-frame)))
1242 (defvar default-keyboard-coding-system nil
1243 "Default value of the keyboard coding system.
1244 This is normally set according to the selected language environment.
1245 See also the command `set-keyboard-coding-system'.")
1247 (defun set-keyboard-coding-system (coding-system &optional display
)
1248 "Set coding system for keyboard input on DISPLAY to CODING-SYSTEM.
1249 In addition, this command calls `encoded-kbd-setup-display' to set up the
1250 translation of keyboard input events to the specified coding system.
1252 For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems].
1253 The default is determined by the selected language environment
1254 or by the previous use of this command.
1256 DISPLAY may be a display id, a frame, or nil for the selected frame's display.
1257 The setting has no effect on graphical displays."
1259 (list (let ((default (if (and (not (keyboard-coding-system))
1260 default-keyboard-coding-system
)
1261 default-keyboard-coding-system
)))
1263 (format "Coding system for keyboard input (default %s): "
1266 (if (and (not coding-system
)
1267 (not (keyboard-coding-system)))
1268 (setq coding-system default-keyboard-coding-system
))
1270 (setq default-keyboard-coding-system coding-system
))
1271 (if (and coding-system
1272 (not (coding-system-get coding-system
:ascii-compatible-p
))
1273 (not (coding-system-get coding-system
:suitable-for-keyboard
)))
1274 (error "%s is not suitable for keyboard" coding-system
))
1275 (set-keyboard-coding-system-internal coding-system display
)
1276 (setq keyboard-coding-system coding-system
)
1277 (encoded-kbd-setup-display display
))
1279 (defcustom keyboard-coding-system nil
1280 "Specify coding system for keyboard input.
1281 If you set this on a terminal which can't distinguish Meta keys from
1282 8-bit characters, you will have to use ESC to type Meta characters.
1283 See Info node `Terminal Coding' and Info node `Unibyte Mode'.
1285 On non-windowing terminals, this is set from the locale by default.
1287 Setting this variable directly does not take effect;
1288 use either \\[customize] or \\[set-keyboard-coding-system]."
1289 :type
'(coding-system :tag
"Coding system")
1290 :link
'(info-link "(emacs)Terminal Coding")
1291 :link
'(info-link "(emacs)Unibyte Mode")
1292 :set
(lambda (symbol value
)
1293 ;; Don't load encoded-kb unnecessarily.
1294 (if (or value
(boundp 'encoded-kbd-setup-display
))
1295 (set-keyboard-coding-system value
)
1296 (set-default 'keyboard-coding-system nil
))) ; must initialize
1301 (defun set-buffer-process-coding-system (decoding encoding
)
1302 "Set coding systems for the process associated with the current buffer.
1303 DECODING is the coding system to be used to decode input from the process,
1304 ENCODING is the coding system to be used to encode output to the process.
1306 For a list of possible values of CODING-SYSTEM, use \\[list-coding-systems]."
1308 "zCoding-system for output from the process: \nzCoding-system for input to the process: ")
1309 (let ((proc (get-buffer-process (current-buffer))))
1311 (error "No process")
1312 (check-coding-system decoding
)
1313 (check-coding-system encoding
)
1314 (set-process-coding-system proc decoding encoding
)))
1315 (force-mode-line-update))
1317 (defalias 'set-clipboard-coding-system
'set-selection-coding-system
)
1319 (defun set-selection-coding-system (coding-system)
1320 "Make CODING-SYSTEM used for communicating with other X clients.
1321 When sending or receiving text via cut_buffer, selection, and clipboard,
1322 the text is encoded or decoded by CODING-SYSTEM."
1323 (interactive "zCoding system for X selection: ")
1324 (check-coding-system coding-system
)
1325 (setq selection-coding-system coding-system
))
1327 ;; Coding system lastly specified by the command
1328 ;; set-next-selection-coding-system.
1329 (defvar last-next-selection-coding-system nil
)
1331 (defun set-next-selection-coding-system (coding-system)
1332 "Use CODING-SYSTEM for next communication with other window system clients.
1333 This setting is effective for the next communication only."
1335 (list (read-coding-system
1336 (if last-next-selection-coding-system
1337 (format "Coding system for the next selection (default %S): "
1338 last-next-selection-coding-system
)
1339 "Coding system for the next selection: ")
1340 last-next-selection-coding-system
)))
1342 (setq last-next-selection-coding-system coding-system
)
1343 (setq coding-system last-next-selection-coding-system
))
1344 (check-coding-system coding-system
)
1346 (setq next-selection-coding-system coding-system
))
1348 (defun set-coding-priority (arg)
1349 "Set priority of coding categories according to ARG.
1350 ARG is a list of coding categories ordered by priority.
1352 This function is provided for backward compatibility.
1353 Now we have more convenient function `set-coding-system-priority'."
1354 (apply 'set-coding-system-priority
1355 (mapcar #'(lambda (x) (symbol-value x
)) arg
)))
1356 (make-obsolete 'set-coding-priority
'set-coding-system-priority
"23.1")
1360 (defvar ctext-non-standard-encodings-alist
1361 '(("big5-0" big5
2 big5
)
1362 ("ISO8859-14" iso-8859-14
1 latin-iso8859-14
)
1363 ("ISO8859-15" iso-8859-15
1 latin-iso8859-15
)
1364 ("gbk-0" gbk
2 chinese-gbk
))
1365 "Alist of non-standard encoding names vs the corresponding usages in CTEXT.
1367 It controls how extended segments of a compound text are handled
1368 by the coding system `compound-text-with-extensions'.
1370 Each element has the form (ENCODING-NAME CODING-SYSTEM N-OCTET CHARSET).
1372 ENCODING-NAME is an encoding name of an \"extended segments\".
1374 CODING-SYSTEM is the coding-system to encode (or decode) the
1375 characters into (or from) the extended segment.
1377 N-OCTET is the number of octets (bytes) that encodes a character
1378 in the segment. It can be 0 (meaning the number of octets per
1379 character is variable), 1, 2, 3, or 4.
1381 CHARSET is a charater set containing characters that are encoded
1382 in the segment. It can be a list of character sets.
1384 On decoding CTEXT, all encoding names listed here are recognized.
1386 On encoding CTEXT, encoding names in the variable
1387 `ctext-non-standard-encodings' (which see) and in the information
1388 listed for the current language environment under the key
1389 `ctext-non-standard-encodings' are used.")
1391 (defvar ctext-non-standard-encodings nil
1392 "List of non-standard encoding names used in extended segments of CTEXT.
1393 Each element must be one of the names listed in the variable
1394 `ctext-non-standard-encodings-alist' (which see).")
1396 (defvar ctext-non-standard-encodings-regexp
1397 (string-to-multibyte
1399 ;; For non-standard encodings.
1400 "\\(\e%/[0-4][\200-\377][\200-\377]\\([^\002]+\\)\002\\)"
1402 ;; For UTF-8 encoding.
1403 "\\(\e%G[^\e]*\e%@\\)")))
1405 ;; Functions to support "Non-Standard Character Set Encodings" defined
1406 ;; by the COMPOUND-TEXT spec. They also support "The UTF-8 encoding"
1407 ;; described in the section 7 of the documentation of COMPOUND-TEXT
1408 ;; distributed with XFree86.
1410 (defun ctext-post-read-conversion (len)
1411 "Decode LEN characters encoded as Compound Text with Extended Segments."
1412 ;; We don't need the following because it is expected that this
1413 ;; function is mainly used for decoding X selection which is not
1415 ;;(buffer-disable-undo) ; minimize consing due to insertions and deletions
1418 (narrow-to-region (point) (+ (point) len
))
1419 (let ((case-fold-search nil
)
1420 last-coding-system-used
1422 (decode-coding-region (point-min) (point-max) 'ctext
)
1423 (while (re-search-forward ctext-non-standard-encodings-regexp
1425 (setq pos
(match-beginning 0))
1426 (if (match-beginning 1)
1427 ;; ESC % / [0-4] M L --ENCODING-NAME-- \002 --BYTES--
1428 (let* ((M (multibyte-char-to-unibyte (char-after (+ pos
4))))
1429 (L (multibyte-char-to-unibyte (char-after (+ pos
5))))
1430 (encoding (match-string 2))
1431 (encoding-info (assoc-string
1433 ctext-non-standard-encodings-alist t
))
1434 (coding (if encoding-info
1435 (nth 1 encoding-info
)
1436 (setq encoding
(intern (downcase encoding
)))
1437 (and (coding-system-p encoding
)
1439 (setq bytes
(- (+ (* (- M
128) 128) (- L
128))
1440 (- (point) (+ pos
6))))
1442 (delete-region pos
(point))
1443 (forward-char bytes
)
1444 (decode-coding-region (- (point) bytes
) (point) coding
)))
1445 ;; ESC % G --UTF-8-BYTES-- ESC % @
1447 (delete-region pos
(+ pos
3))
1448 (decode-coding-region pos
(point) 'utf-8
))))
1449 (goto-char (point-min))
1450 (- (point-max) (point)))))
1452 ;; Return an alist of CHARSET vs CTEXT-USAGE-INFO generated from
1453 ;; `ctext-non-standard-encodings' and a list specified by the key
1454 ;; `ctext-non-standard-encodings' for the currrent language
1455 ;; environment. CTEXT-USAGE-INFO is one of the element of
1456 ;; `ctext-non-standard-encodings-alist' or nil. In the former case, a
1457 ;; character in CHARSET is encoded using extended segment. In the
1458 ;; latter case, a character in CHARSET is encoded using normal ISO2022
1459 ;; designation sequence. If a character is not in any of CHARSETs, it
1460 ;; is encoded using UTF-8 encoding extention.
1462 (defun ctext-non-standard-encodings-table ()
1464 ;; Setup charsets specified by the key
1465 ;; `ctext-non-standard-encodings' for the current language
1466 ;; environment and in `ctext-non-standard-encodings'.
1467 (dolist (encoding (append
1468 (get-language-info current-language-environment
1469 'ctext-non-standard-encodings
)
1470 ctext-non-standard-encodings
))
1471 (let* ((slot (assoc encoding ctext-non-standard-encodings-alist
))
1472 (charset (nth 3 slot
)))
1473 (if (charsetp charset
)
1474 (push (cons charset slot
) table
)
1475 (dolist (cs charset
)
1476 (push (cons cs slot
) table
)))))
1478 ;; Next prepend charsets for ISO2022 designation sequence.
1479 (dolist (charset charset-list
)
1480 (let ((final (plist-get (charset-plist charset
) :iso-final-char
)))
1481 (if (and (integerp final
)
1482 (>= final
#x40
) (<= final
#x7e
)
1483 ;; Exclude ascii and chinese-cns11643-X.
1484 (not (eq charset
'ascii
))
1485 (not (string-match "cns11643" (symbol-name charset
))))
1486 (push (cons charset nil
) table
))))
1488 ;; Returned reversed list so that the charsets specified by the
1489 ;; key `ctext-non-standard-encodings' for the current language
1490 ;; have the highest priority.
1493 (defun ctext-pre-write-conversion (from to
)
1494 "Encode characters between FROM and TO as Compound Text w/Extended Segments.
1496 If FROM is a string, or if the current buffer is not the one set up for us
1497 by `encode-coding-string', generate a new temp buffer, insert the
1498 text, and convert it in the temporary buffer. Otherwise, convert in-place."
1500 ;; Setup a working buffer if necessary.
1501 (when (stringp from
)
1502 (set-buffer (generate-new-buffer " *temp"))
1503 (set-buffer-multibyte (multibyte-string-p from
))
1506 ;; Now we can encode the whole buffer.
1507 (let ((encoding-table (ctext-non-standard-encodings-table))
1508 last-coding-system-used
1509 last-pos last-encoding-info
1510 encoding-info end-pos ch
)
1511 (goto-char (setq last-pos
(point-min)))
1512 (setq end-pos
(point-marker))
1513 (while (re-search-forward "[^\000-\177]+" nil t
)
1514 ;; Found a sequence of non-ASCII characters.
1515 (setq last-pos
(match-beginning 0)
1516 ch
(char-after last-pos
)
1517 last-encoding-info
(catch 'tag
1518 (dolist (elt encoding-table
)
1519 (if (encode-char ch
(car elt
))
1520 (throw 'tag
(cdr elt
))))
1522 (set-marker end-pos
(match-end 0))
1523 (goto-char (1+ last-pos
))
1527 (if (< (point) end-pos
)
1529 (setq ch
(following-char))
1530 (dolist (elt encoding-table
)
1531 (if (encode-char ch
(car elt
))
1532 (throw 'tag
(cdr elt
))))
1534 (unless (eq last-encoding-info encoding-info
)
1535 (cond ((consp last-encoding-info
)
1536 ;; Encode the previous range using an extended
1538 (let ((encoding-name (car last-encoding-info
))
1539 (coding-system (nth 1 last-encoding-info
))
1540 (noctets (nth 2 last-encoding-info
))
1542 (encode-coding-region last-pos
(point) coding-system
)
1543 (setq len
(+ (length encoding-name
) 1
1544 (- (point) last-pos
)))
1545 ;; According to the spec of CTEXT, it is not
1546 ;; necessary to produce this extra designation
1547 ;; sequence, but some buggy application
1548 ;; (e.g. crxvt-gb) requires it.
1551 (goto-char last-pos
)
1552 (insert (format "\e%%/%d" noctets
))
1553 (insert-byte (+ (/ len
128) 128) 1)
1554 (insert-byte (+ (% len
128) 128) 1)
1555 (insert encoding-name
)
1557 ((eq last-encoding-info
'utf-8
)
1558 ;; Encode the previous range using UTF-8 encoding
1560 (encode-coding-region last-pos
(point) 'mule-utf-8
)
1562 (goto-char last-pos
)
1565 (setq last-pos
(point)
1566 last-encoding-info encoding-info
))
1567 (if (< (point) end-pos
)
1569 (throw 'tag nil
)))))
1570 (set-marker end-pos nil
)
1571 (goto-char (point-min))))
1572 ;; Must return nil, as build_annotations_2 expects that.
1577 (defcustom auto-coding-alist
1578 ;; .exe and .EXE are added to support archive-mode looking at DOS
1579 ;; self-extracting exe archives.
1581 arc\\|zip\\|lzh\\|lha\\|zoo\\|[jew]ar\\|xpi\\|rar\\|\
1582 ARC\\|ZIP\\|LZH\\|LHA\\|ZOO\\|[JEW]AR\\|XPI\\|RAR\\)\\'"
1583 . no-conversion-multibyte
)
1584 ("\\.\\(exe\\|EXE\\)\\'" . no-conversion
)
1585 ("\\.\\(sx[dmicw]\\|odt\\|tar\\|tgz\\)\\'" . no-conversion
)
1586 ("\\.\\(gz\\|Z\\|bz\\|bz2\\|gpg\\)\\'" . no-conversion
)
1587 ("\\.\\(jpe?g\\|png\\|gif\\|tiff?\\|p[bpgn]m\\)\\'" . no-conversion
)
1588 ("\\.pdf\\'" . no-conversion
)
1589 ("/#[^/]+#\\'" . emacs-mule
))
1590 "Alist of filename patterns vs corresponding coding systems.
1591 Each element looks like (REGEXP . CODING-SYSTEM).
1592 A file whose name matches REGEXP is decoded by CODING-SYSTEM on reading.
1594 The settings in this alist take priority over `coding:' tags
1595 in the file (see the function `set-auto-coding')
1596 and the contents of `file-coding-system-alist'."
1599 :type
'(repeat (cons (regexp :tag
"File name regexp")
1600 (symbol :tag
"Coding system"))))
1602 (defcustom auto-coding-regexp-alist
1603 '(("^BABYL OPTIONS:[ \t]*-\\*-[ \t]*rmail[ \t]*-\\*-" . no-conversion
)
1604 ("\\`\xFE\xFF" . utf-16be-with-signature
)
1605 ("\\`\xFF\xFE" . utf-16le-with-signature
)
1606 ("\\`\xEF\xBB\xBF" . utf-8
)
1607 ("\\`;ELC\024\0\0\0" . emacs-mule
)) ; Emacs 20-compiled
1608 "Alist of patterns vs corresponding coding systems.
1609 Each element looks like (REGEXP . CODING-SYSTEM).
1610 A file whose first bytes match REGEXP is decoded by CODING-SYSTEM on reading.
1612 The settings in this alist take priority over `coding:' tags
1613 in the file (see the function `set-auto-coding')
1614 and the contents of `file-coding-system-alist'."
1617 :type
'(repeat (cons (regexp :tag
"Regexp")
1618 (symbol :tag
"Coding system"))))
1620 (defun auto-coding-regexp-alist-lookup (from to
)
1621 "Lookup `auto-coding-regexp-alist' for the contents of the current buffer.
1622 The value is a coding system is specified for the region FROM and TO,
1626 (let ((alist auto-coding-regexp-alist
)
1628 (while (and alist
(not coding-system
))
1629 (let ((regexp (car (car alist
))))
1630 (if enable-multibyte-characters
1631 (setq regexp
(string-to-multibyte regexp
)))
1632 (if (re-search-forward regexp to t
)
1633 (setq coding-system
(cdr (car alist
)))
1634 (setq alist
(cdr alist
)))))
1637 ;; See the bottom of this file for built-in auto coding functions.
1638 (defcustom auto-coding-functions
'(sgml-xml-auto-coding-function
1639 sgml-html-meta-auto-coding-function
)
1640 "A list of functions which attempt to determine a coding system.
1642 Each function in this list should be written to operate on the
1643 current buffer, but should not modify it in any way. The buffer
1644 will contain undecoded text of parts of the file. Each function
1645 should take one argument, SIZE, which says how many
1646 characters (starting from point) it should look at.
1648 If one of these functions succeeds in determining a coding
1649 system, it should return that coding system. Otherwise, it
1652 If a file has a `coding:' tag, that takes precedence over these
1653 functions, so they won't be called at all."
1656 :type
'(repeat function
))
1658 (defvar set-auto-coding-for-load nil
1659 "Non-nil means look for `load-coding' property instead of `coding'.
1660 This is used for loading and byte-compiling Emacs Lisp files.")
1662 (defun auto-coding-alist-lookup (filename)
1663 "Return the coding system specified by `auto-coding-alist' for FILENAME."
1664 (let ((alist auto-coding-alist
)
1665 (case-fold-search (memq system-type
'(windows-nt ms-dos cygwin
)))
1667 (while (and alist
(not coding-system
))
1668 (if (string-match (car (car alist
)) filename
)
1669 (setq coding-system
(cdr (car alist
)))
1670 (setq alist
(cdr alist
))))
1673 (put 'enable-character-translation
'permanent-local t
)
1674 (put 'enable-character-translation
'safe-local-variable
'booleanp
)
1676 (defun find-auto-coding (filename size
)
1677 "Find a coding system for a file FILENAME of which SIZE bytes follow point.
1678 These bytes should include at least the first 1k of the file
1679 and the last 3k of the file, but the middle may be omitted.
1681 The function checks FILENAME against the variable `auto-coding-alist'.
1682 If FILENAME doesn't match any entries in the variable, it checks the
1683 contents of the current buffer following point against
1684 `auto-coding-regexp-alist'. If no match is found, it checks for a
1685 `coding:' tag in the first one or two lines following point. If no
1686 `coding:' tag is found, it checks any local variables list in the last
1687 3K bytes out of the SIZE bytes. Finally, if none of these methods
1688 succeed, it checks to see if any function in `auto-coding-functions'
1691 If a coding system is specifed, the return value is a
1692 cons (CODING . SOURCE), where CODING is the specified coding
1693 system and SOURCE is a symbol `auto-coding-alist',
1694 `auto-coding-regexp-alist', `coding:', or `auto-coding-functions'
1695 indicating by what CODING is specified. Note that the validity
1696 of CODING is not checked; it's the caller's responsibility to check it.
1698 If nothing is specified, the return value is nil."
1699 (or (let ((coding-system (auto-coding-alist-lookup filename
)))
1701 (cons coding-system
'auto-coding-alist
)))
1702 ;; Try using `auto-coding-regexp-alist'.
1703 (let ((coding-system (auto-coding-regexp-alist-lookup (point)
1706 (cons coding-system
'auto-coding-regexp-alist
)))
1707 (let* ((case-fold-search t
)
1708 (head-start (point))
1709 (head-end (+ head-start
(min size
1024)))
1710 (tail-start (+ head-start
(max (- size
3072) 0)))
1711 (tail-end (+ head-start size
))
1712 coding-system head-found tail-found pos char-trans
)
1713 ;; Try a short cut by searching for the string "coding:"
1714 ;; and for "unibyte:" at the head and tail of SIZE bytes.
1715 (setq head-found
(or (search-forward "coding:" head-end t
)
1716 (search-forward "unibyte:" head-end t
)
1717 (search-forward "enable-character-translation:"
1719 (if (and head-found
(> head-found tail-start
))
1720 ;; Head and tail are overlapped.
1721 (setq tail-found head-found
)
1722 (goto-char tail-start
)
1723 (setq tail-found
(or (search-forward "coding:" tail-end t
)
1724 (search-forward "unibyte:" tail-end t
)
1725 (search-forward "enable-character-translation:"
1728 ;; At first check the head.
1730 (goto-char head-start
)
1731 (setq head-end
(set-auto-mode-1))
1732 (setq head-start
(point))
1733 (when (and head-end
(< head-found head-end
))
1734 (goto-char head-start
)
1735 (when (and set-auto-coding-for-load
1737 "\\(.*;\\)?[ \t]*unibyte:[ \t]*\\([^ ;]+\\)"
1739 (setq coding-system
'raw-text
))
1740 (when (and (not coding-system
)
1742 "\\(.*;\\)?[ \t]*coding:[ \t]*\\([^ ;]+\\)"
1744 (setq coding-system
(intern (match-string 2))))
1745 (when (re-search-forward
1746 "\\(.*;\\)?[ \t]*enable-character-translation:[ \t]*\\([^ ;]+\\)"
1748 (setq char-trans
(match-string 2)))))
1750 ;; If no coding: tag in the head, check the tail.
1751 ;; Here we must pay attention to the case that the end-of-line
1752 ;; is just "\r" and we can't use "^" nor "$" in regexp.
1753 (when (and tail-found
(or (not coding-system
) (not char-trans
)))
1754 (goto-char tail-start
)
1755 (re-search-forward "[\r\n]\^L" tail-end t
)
1756 (if (re-search-forward
1757 "[\r\n]\\([^[\r\n]*\\)[ \t]*Local Variables:[ \t]*\\([^\r\n]*\\)[\r\n]"
1759 ;; The prefix is what comes before "local variables:" in its
1760 ;; line. The suffix is what comes after "local variables:"
1762 (let* ((prefix (regexp-quote (match-string 1)))
1763 (suffix (regexp-quote (match-string 2)))
1767 ;; N.B. without the \n below, the regexp can
1769 "[ \t]*coding[ \t]*:[ \t]*\\([^ \t\r\n]+\\)[ \t]*"
1774 "[ \t]*unibyte[ \t]*:[ \t]*\\([^ \t\r\n]+\\)[ \t]*"
1779 "[ \t]*enable-character-translation[ \t]*:[ \t]*\\([^ \t\r\n]+\\)[ \t]*"
1782 (concat "[\r\n]" prefix
"[ \t]*End *:[ \t]*" suffix
1785 (forward-char -
1) ; skip back \r or \n.
1786 (re-search-forward re-end tail-end
'move
)
1787 (setq tail-end
(point))
1789 (when (and set-auto-coding-for-load
1790 (re-search-forward re-unibyte tail-end t
))
1791 (setq coding-system
'raw-text
))
1792 (when (and (not coding-system
)
1793 (re-search-forward re-coding tail-end t
))
1794 (setq coding-system
(intern (match-string 1))))
1795 (when (and (not char-trans
)
1796 (re-search-forward re-char-trans tail-end t
))
1797 (setq char-trans
(match-string 1))))))
1799 ;; If the coding-system name ends with "!", remove it and
1800 ;; set char-trans to "nil".
1801 (let ((name (symbol-name coding-system
)))
1802 (if (= (aref name
(1- (length name
))) ?
!)
1803 (setq coding-system
(intern (substring name
0 -
1))
1804 char-trans
"nil"))))
1805 (when (and char-trans
1806 (not (setq char-trans
(intern char-trans
))))
1807 (make-local-variable 'enable-character-translation
)
1808 (setq enable-character-translation nil
))
1810 (cons coding-system
:coding
)))
1811 ;; Finally, try all the `auto-coding-functions'.
1812 (let ((funcs auto-coding-functions
)
1813 (coding-system nil
))
1814 (while (and funcs
(not coding-system
))
1815 (setq coding-system
(condition-case e
1817 (goto-char (point-min))
1818 (funcall (pop funcs
) size
))
1821 (cons coding-system
'auto-coding-functions
)))))
1823 (defun set-auto-coding (filename size
)
1824 "Return coding system for a file FILENAME of which SIZE bytes follow point.
1825 See `find-auto-coding' for how the coding system is found.
1826 Return nil if an invalid coding system is found.
1828 The variable `set-auto-coding-function' (which see) is set to this
1829 function by default."
1830 (let ((found (find-auto-coding filename size
)))
1831 (if (and found
(coding-system-p (car found
)))
1834 (setq set-auto-coding-function
'set-auto-coding
)
1836 ;; This variable is set in these two cases:
1837 ;; (1) A file is read by a coding system specified explicitly.
1838 ;; after-insert-file-set-coding sets this value to
1839 ;; coding-system-for-read.
1840 ;; (2) A buffer is saved.
1841 ;; After writing, basic-save-buffer-1 sets this value to
1842 ;; last-coding-system-used.
1843 ;; This variable is used for decoding in revert-buffer.
1844 (defvar buffer-file-coding-system-explicit nil
1845 "The file coding system explicitly specified for the current buffer.
1846 Internal use only.")
1847 (make-variable-buffer-local 'buffer-file-coding-system-explicit
)
1848 (put 'buffer-file-coding-system-explicit
'permanent-local t
)
1850 (defun after-insert-file-set-coding (inserted &optional visit
)
1851 "Set `buffer-file-coding-system' of current buffer after text is inserted.
1852 INSERTED is the number of characters that were inserted, as figured
1853 in the situation before this function. Return the number of characters
1854 inserted, as figured in the situation after. The two numbers can be
1855 different if the buffer has become unibyte.
1856 The optional second arg VISIT non-nil means that we are visiting a file."
1858 coding-system-for-read
1859 (not (eq coding-system-for-read
'auto-save-coding
)))
1860 (setq buffer-file-coding-system-explicit coding-system-for-read
))
1861 (if last-coding-system-used
1862 (let ((coding-system
1863 (find-new-buffer-file-coding-system last-coding-system-used
)))
1865 ;; Tell set-buffer-file-coding-system not to mark the file
1866 ;; as modified; we just read it, and it's supposed to be unmodified.
1867 ;; Marking it modified would try to lock it, which would
1868 ;; check the modtime, and we don't want to do that again now.
1869 (set-buffer-file-coding-system coding-system t t
))))
1872 ;; The coding-spec and eol-type of coding-system returned is decided
1873 ;; independently in the following order.
1874 ;; 1. That of buffer-file-coding-system locally bound.
1875 ;; 2. That of CODING.
1877 (defun find-new-buffer-file-coding-system (coding)
1878 "Return a coding system for a buffer when a file of CODING is inserted.
1879 The local variable `buffer-file-coding-system' of the current buffer
1880 is set to the returned value.
1881 Return nil if there's no need to set `buffer-file-coding-system'."
1882 (let (local-coding local-eol
1883 found-coding found-eol
1886 ;; Nothing found about coding.
1889 ;; Get information of `buffer-file-coding-system' in LOCAL-EOL
1890 ;; and LOCAL-CODING.
1891 (setq local-eol
(coding-system-eol-type buffer-file-coding-system
))
1892 (if (null (numberp local-eol
))
1893 ;; But eol-type is not yet set.
1894 (setq local-eol nil
))
1895 (if (and buffer-file-coding-system
1896 (not (eq (coding-system-type buffer-file-coding-system
)
1898 (setq local-coding
(coding-system-base buffer-file-coding-system
)))
1900 (if (and (local-variable-p 'buffer-file-coding-system
)
1901 local-eol local-coding
)
1902 ;; The current buffer has already set full coding-system, we
1903 ;; had better not change it.
1906 (setq found-eol
(coding-system-eol-type coding
))
1907 (if (null (numberp found-eol
))
1908 ;; But eol-type is not found.
1909 ;; If EOL conversions are inhibited, force unix eol-type.
1910 (setq found-eol
(if inhibit-eol-conversion
0)))
1911 (setq found-coding
(coding-system-base coding
))
1913 (if (and (not found-eol
) (eq found-coding
'undecided
))
1914 ;; No valid coding information found.
1917 ;; Some coding information (eol or text) found.
1919 ;; The local setting takes precedence over the found one.
1920 (setq new-coding
(if (local-variable-p 'buffer-file-coding-system
)
1921 (or local-coding found-coding
)
1922 (or found-coding local-coding
)))
1923 (setq new-eol
(if (local-variable-p 'buffer-file-coding-system
)
1924 (or local-eol found-eol
)
1925 (or found-eol local-eol
)))
1927 (let ((eol-type (coding-system-eol-type new-coding
)))
1928 (if (and (numberp new-eol
) (vectorp eol-type
))
1929 (aref eol-type new-eol
)
1932 (defun modify-coding-system-alist (target-type regexp coding-system
)
1933 "Modify one of look up tables for finding a coding system on I/O operation.
1934 There are three of such tables, `file-coding-system-alist',
1935 `process-coding-system-alist', and `network-coding-system-alist'.
1937 TARGET-TYPE specifies which of them to modify.
1938 If it is `file', it affects `file-coding-system-alist' (which see).
1939 If it is `process', it affects `process-coding-system-alist' (which see).
1940 If it is `network', it affects `network-coding-system-alist' (which see).
1942 REGEXP is a regular expression matching a target of I/O operation.
1943 The target is a file name if TARGET-TYPE is `file', a program name if
1944 TARGET-TYPE is `process', or a network service name or a port number
1945 to connect to if TARGET-TYPE is `network'.
1947 CODING-SYSTEM is a coding system to perform code conversion on the I/O
1948 operation, or a cons cell (DECODING . ENCODING) specifying the coding systems
1949 for decoding and encoding respectively,
1950 or a function symbol which, when called, returns such a cons cell."
1951 (or (memq target-type
'(file process network
))
1952 (error "Invalid target type: %s" target-type
))
1953 (or (stringp regexp
)
1954 (and (eq target-type
'network
) (integerp regexp
))
1955 (error "Invalid regular expression: %s" regexp
))
1956 (if (symbolp coding-system
)
1957 (if (not (fboundp coding-system
))
1959 (check-coding-system coding-system
)
1960 (setq coding-system
(cons coding-system coding-system
))))
1961 (check-coding-system (car coding-system
))
1962 (check-coding-system (cdr coding-system
)))
1963 (cond ((eq target-type
'file
)
1964 (let ((slot (assoc regexp file-coding-system-alist
)))
1966 (setcdr slot coding-system
)
1967 (setq file-coding-system-alist
1968 (cons (cons regexp coding-system
)
1969 file-coding-system-alist
)))))
1970 ((eq target-type
'process
)
1971 (let ((slot (assoc regexp process-coding-system-alist
)))
1973 (setcdr slot coding-system
)
1974 (setq process-coding-system-alist
1975 (cons (cons regexp coding-system
)
1976 process-coding-system-alist
)))))
1978 (let ((slot (assoc regexp network-coding-system-alist
)))
1980 (setcdr slot coding-system
)
1981 (setq network-coding-system-alist
1982 (cons (cons regexp coding-system
)
1983 network-coding-system-alist
)))))))
1985 (defun decode-coding-inserted-region (from to filename
1986 &optional visit beg end replace
)
1987 "Decode the region between FROM and TO as if it is read from file FILENAME.
1988 The idea is that the text between FROM and TO was just inserted somehow.
1989 Optional arguments VISIT, BEG, END, and REPLACE are the same as those
1990 of the function `insert-file-contents'.
1991 Part of the job of this function is setting `buffer-undo-list' appropriately."
1994 (let ((coding coding-system-for-read
)
1997 ;; Temporarily turn off undo recording, if we're decoding the
1998 ;; text of a visited file.
1999 (setq buffer-undo-list t
)
2000 ;; Otherwise, if we can recognize the undo elt for the insertion,
2001 ;; remove it and get ready to replace it later.
2002 ;; In the mean time, turn off undo recording.
2003 (let ((last (car-safe buffer-undo-list
)))
2004 (if (and (consp last
) (eql (car last
) from
) (eql (cdr last
) to
))
2005 (setq undo-list-saved
(cdr buffer-undo-list
)
2006 buffer-undo-list t
))))
2007 (narrow-to-region from to
)
2008 (goto-char (point-min))
2010 (setq coding
(funcall set-auto-coding-function
2011 filename
(- (point-max) (point-min)))))
2013 (setq coding
(car (find-operation-coding-system
2014 'insert-file-contents
2015 (cons filename
(current-buffer))
2016 visit beg end replace
))))
2017 (if (coding-system-p coding
)
2018 (or enable-multibyte-characters
2020 (coding-system-change-text-conversion coding
'raw-text
)))
2023 (decode-coding-region (point-min) (point-max) coding
)
2024 (setq last-coding-system-used coding
))
2025 ;; If we're decoding the text of a visited file,
2026 ;; the undo list should start out empty.
2028 (setq buffer-undo-list nil
)
2029 ;; If we decided to replace the undo entry for the insertion,
2032 (setq buffer-undo-list
2033 (cons (cons from
(point-max)) undo-list-saved
))))))))
2035 (defun recode-region (start end new-coding coding
)
2036 "Re-decode the region (previously decoded by CODING) by NEW-CODING."
2038 (list (region-beginning) (region-end)
2039 (read-coding-system "Text was really in: ")
2040 (let ((coding (or buffer-file-coding-system last-coding-system-used
)))
2042 (concat "But was interpreted as"
2043 (if coding
(format " (default %S): " coding
) ": "))
2045 (or (and new-coding coding
)
2046 (error "Coding system not specified"))
2047 ;; Check it before we encode the region.
2048 (check-coding-system new-coding
)
2050 (narrow-to-region start end
)
2051 (encode-coding-region (point-min) (point-max) coding
)
2052 (decode-coding-region (point-min) (point-max) new-coding
))
2053 (if (region-active-p)
2056 (defun make-translation-table (&rest args
)
2057 "Make a translation table from arguments.
2058 A translation table is a char table intended for character
2059 translation in CCL programs.
2061 Each argument is a list of elements of the form (FROM . TO), where FROM
2062 is a character to be translated to TO.
2064 The arguments and forms in each argument are processed in the given
2065 order, and if a previous form already translates TO to some other
2066 character, say TO-ALT, FROM is also translated to TO-ALT."
2067 (let ((table (make-char-table 'translation-table
))
2071 (let ((from (car elt
))
2073 to-alt rev-from rev-to
)
2074 ;; If we have already translated TO to TO-ALT, FROM should
2075 ;; also be translated to TO-ALT.
2076 (if (setq to-alt
(aref table to
))
2078 (aset table from to
)
2079 ;; If we have already translated some chars to FROM, they
2080 ;; should also be translated to TO.
2081 (when (setq rev-from
(assq from revlist
))
2082 (dolist (elt (cdr rev-from
))
2083 (aset table elt to
))
2084 (setq revlist
(delq rev-from revlist
)
2085 rev-from
(cdr rev-from
)))
2086 ;; Now update REVLIST.
2087 (setq rev-to
(assq to revlist
))
2089 (setcdr rev-to
(cons from
(cdr rev-to
)))
2090 (setq rev-to
(list to from
)
2091 revlist
(cons rev-to revlist
)))
2093 (setcdr rev-to
(append rev-from
(cdr rev-to
)))))))
2094 ;; Return TABLE just created.
2095 (set-char-table-extra-slot table
1 1)
2098 (defun make-translation-table-from-vector (vec)
2099 "Make translation table from decoding vector VEC.
2100 VEC is an array of 256 elements to map unibyte codes to multibyte
2101 characters. Elements may be nil for undefined code points.
2102 See also the variable `nonascii-translation-table'."
2103 (let ((table (make-char-table 'translation-table
))
2104 (rev-table (make-char-table 'translation-table
))
2107 (setq ch
(aref vec i
))
2111 (aset rev-table ch i
))))
2112 (set-char-table-extra-slot table
0 rev-table
)
2113 (set-char-table-extra-slot table
1 1)
2114 (set-char-table-extra-slot rev-table
1 1)
2117 (defun make-translation-table-from-alist (alist)
2118 "Make translation table from N<->M mapping in ALIST.
2119 ALIST is an alist, each element has the form (FROM . TO).
2120 FROM and TO are a character or a vector of characters.
2121 If FROM is a character, that character is translated to TO.
2122 If FROM is a vector of characters, that sequence is translated to TO.
2123 The first extra-slot of the value is a translation table for reverse mapping."
2124 (let ((tables (vector (make-char-table 'translation-table
)
2125 (make-char-table 'translation-table
)))
2126 table max-lookup from to idx val
)
2128 (setq table
(aref tables i
))
2132 (setq from
(car elt
) to
(cdr elt
))
2133 (setq from
(cdr elt
) to
(car elt
)))
2134 (if (characterp from
)
2136 (setq idx
(aref from
0)
2137 max-lookup
(max max-lookup
(length from
))))
2138 (setq val
(aref table idx
))
2142 (setq val
(list (cons (vector idx
) val
))))
2143 (if (characterp from
)
2144 (setq from
(vector from
)))
2145 (setq val
(nconc val
(list (cons from to
)))))
2146 (if (characterp from
)
2148 (setq val
(list (cons from to
)))))
2149 (aset table idx val
))
2150 (set-char-table-extra-slot table
1 max-lookup
))
2151 (set-char-table-extra-slot (aref tables
0) 0 (aref tables
1))
2154 (defun define-translation-table (symbol &rest args
)
2155 "Define SYMBOL as the name of translation table made by ARGS.
2156 This sets up information so that the table can be used for
2157 translations in a CCL program.
2159 If the first element of ARGS is a char-table whose purpose is
2160 `translation-table', just define SYMBOL to name it. (Note that this
2161 function does not bind SYMBOL.)
2163 Any other ARGS should be suitable as arguments of the function
2164 `make-translation-table' (which see).
2166 This function sets properties `translation-table' and
2167 `translation-table-id' of SYMBOL to the created table itself and the
2168 identification number of the table respectively. It also registers
2169 the table in `translation-table-vector'."
2170 (let ((table (if (and (char-table-p (car args
))
2171 (eq (char-table-subtype (car args
))
2172 'translation-table
))
2174 (apply 'make-translation-table args
)))
2175 (len (length translation-table-vector
))
2178 (put symbol
'translation-table table
)
2181 (setq translation-table-vector
2182 (vconcat translation-table-vector
(make-vector len nil
))))
2183 (let ((slot (aref translation-table-vector id
)))
2185 (eq (car slot
) symbol
))
2187 (aset translation-table-vector id
(cons symbol table
))
2189 (setq id
(1+ id
)))))
2190 (put symbol
'translation-table-id id
)
2193 (defun translate-region (start end table
)
2194 "From START to END, translate characters according to TABLE.
2195 TABLE is a string or a char-table.
2196 If TABLE is a string, the Nth character in it is the mapping
2197 for the character with code N.
2198 If TABLE is a char-table, the element for character N is the mapping
2199 for the character with code N.
2200 It returns the number of characters changed."
2202 (list (region-beginning)
2205 (dotimes (i (length translation-table-vector
))
2206 (if (consp (aref translation-table-vector i
))
2207 (push (list (symbol-name
2208 (car (aref translation-table-vector i
)))) l
)))
2210 (error "No translation table defined"))
2212 (setq table
(completing-read "Translation table: " l nil t
)))
2215 (let ((val (get table
'translation-table
)))
2216 (or (char-table-p val
)
2217 (error "Invalid translation table name: %s" table
))
2219 (translate-region-internal start end table
))
2221 (put 'with-category-table
'lisp-indent-function
1)
2223 (defmacro with-category-table
(table &rest body
)
2224 "Execute BODY like `progn' with CATEGORY-TABLE the current category table.
2225 The category table of the current buffer is saved, BODY is evaluated,
2226 then the saved table is restored, even in case of an abnormal exit.
2227 Value is what BODY returns."
2228 (let ((old-table (make-symbol "old-table"))
2229 (old-buffer (make-symbol "old-buffer")))
2230 `(let ((,old-table
(category-table))
2231 (,old-buffer
(current-buffer)))
2234 (set-category-table ,table
)
2236 (with-current-buffer ,old-buffer
2237 (set-category-table ,old-table
))))))
2239 (defun define-translation-hash-table (symbol table
)
2240 "Define SYMBOL as the name of the hash translation TABLE for use in CCL.
2242 Analogous to `define-translation-table', but updates
2243 `translation-hash-table-vector' and the table is for use in the CCL
2244 `lookup-integer' and `lookup-character' functions."
2245 (unless (and (symbolp symbol
)
2246 (hash-table-p table
))
2247 (error "Bad args to define-translation-hash-table"))
2248 (let ((len (length translation-hash-table-vector
))
2251 (put symbol
'translation-hash-table table
)
2254 (setq translation-hash-table-vector
2255 (vconcat translation-hash-table-vector
[nil])))
2256 (let ((slot (aref translation-hash-table-vector id)))
2258 (eq (car slot) symbol))
2260 (aset translation-hash-table-vector id (cons symbol table))
2262 (setq id (1+ id)))))
2263 (put symbol 'translation-hash-table-id id)
2266 ;;; Initialize some variables.
2268 (put 'use-default-ascent 'char-table-extra-slots 0)
2269 (setq use-default-ascent (make-char-table 'use-default-ascent))
2270 (put 'ignore-relative-composition 'char-table-extra-slots 0)
2271 (setq ignore-relative-composition
2272 (make-char-table 'ignore-relative-composition))
2274 (make-obsolete 'set-char-table-default
2275 "generic characters no longer exist." "23.1")
2277 ;;; Built-in auto-coding-functions:
2279 (defun sgml-xml-auto-coding-function (size)
2280 "Determine whether the buffer is XML, and if so, its encoding.
2281 This function is intended to be added to `auto-coding-functions'."
2282 (setq size (+ (point) size))
2283 (when (re-search-forward "\\`[[:space:]\n]*<\\?xml" size t)
2284 (let ((end (save-excursion
2286 (re-search-forward "[\"']\\s-*\\?>" size t))))
2288 (if (re-search-forward "encoding=[\"']\\(.+?\\)[\"']" end t)
2289 (let* ((match (match-string 1))
2290 (sym (intern (downcase match))))
2291 (if (coding-system-p sym)
2293 (message "Warning: unknown coding system \"%s\"" match)
2295 ;; Files without an encoding tag should be UTF-8. But users
2296 ;; may be naive about encodings, and have saved the file from
2297 ;; another editor that does not help them get the encoding right.
2298 ;; Detect the encoding and warn the user if it is detected as
2299 ;; something other than UTF-8.
2301 (with-coding-priority '(utf-8)
2303 (detect-coding-region (point-min) size t)))))
2304 ;; Pure ASCII always comes back as undecided.
2305 (if (memq detected '(utf-8 undecided))
2307 (warn "File contents detected as %s.
2308 Consider adding an encoding attribute to the xml declaration,
2309 or saving as utf-8, as mandated by the xml specification." detected)
2312 (defun sgml-html-meta-auto-coding-function (size)
2313 "If the buffer has an HTML meta tag, use it to determine encoding.
2314 This function is intended to be added to `auto-coding-functions'."
2315 (let ((case-fold-search t))
2316 (setq size (min (+ (point) size)
2318 ;; Limit the search by the end of the HTML header.
2319 (or (search-forward "</head>" (+ (point) size) t)
2320 ;; In case of no header, search only 10 lines.
2323 ;; Make sure that the buffer really contains an HTML document, by
2324 ;; checking that it starts with a doctype or a <HTML> start tag
2325 ;; (allowing for whitespace at bob). Note: 'DOCTYPE NETSCAPE' is
2326 ;; useful for Mozilla bookmark files.
2327 (when (and (re-search-forward "\\`[[:space:]\n]*\\(<!doctype[[:space:]\n]+\\(html\\|netscape\\)\\|<html\\)" size t)
2328 (re-search-forward "<meta\\s-+http-equiv=[\"']?content-type[\"']?\\s-+content=[\"']text/\\sw+;\\s-*charset=\\(.+?\\)[\"']" size t))
2329 (let* ((match (match-string 1))
2330 (sym (intern (downcase match))))
2331 (if (coding-system-p sym)
2333 (message "Warning: unknown coding system \"%s\"" match)
2336 (defun xml-find-file-coding-system (args)
2337 "Determine the coding system of an XML file without a declaration.
2338 Strictly speaking, the file should be utf-8, but mistakes are
2339 made, and there are genuine cases where XML fragments are saved,
2340 with the encoding properly specified in a master document, or
2341 added by processing software."
2342 (if (eq (car args) 'insert-file-contents)
2344 (with-coding-priority '(utf-8)
2346 (detect-coding-region (point-min) (point-max) t)))))
2347 ;; Pure ASCII always comes back as undecided.
2349 ((memq detected '(utf-8 undecided))
2351 ((eq detected 'utf-16le-with-signature) 'utf-16le-with-signature)
2352 ((eq detected 'utf-16be-with-signature) 'utf-16be-with-signature)
2354 (warn "File contents detected as %s.
2355 Consider adding an xml declaration with the encoding specified,
2356 or saving as utf-8, as mandated by the xml specification." detected)
2358 ;; Don't interfere with the user's wishes for saving the buffer.
2359 ;; We did what we could when the buffer was created to ensure the
2360 ;; correct encoding was used, or the user was warned, so any
2361 ;; non-conformity here is deliberate on the part of the user.
2367 ;; arch-tag: 9aebaa6e-0e8a-40a9-b857-cb5d04a39e7c
2368 ;;; mule.el ends here
HCoop / bpt / emacs.git / blob