1 /* Code conversion functions.
2 Copyright (C) 1991, 1995 Free Software Foundation, Inc.
3 Copyright (C) 1995 Sun Microsystems, Inc.
5 This file is part of XEmacs.
7 XEmacs is free software; you can redistribute it and/or modify it
8 under the terms of the GNU General Public License as published by the
9 Free Software Foundation; either version 2, or (at your option) any
12 XEmacs is distributed in the hope that it will be useful, but WITHOUT
13 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with XEmacs; see the file COPYING. If not, write to
19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 Boston, MA 02111-1307, USA. */
22 /* Synched up with: Mule 2.3. Not in FSF. */
24 /* Rewritten by Ben Wing <ben@xemacs.org>. */
37 #include "file-coding.h"
39 Lisp_Object Qbuffer_file_coding_system, Qcoding_system_error;
41 Lisp_Object Vkeyboard_coding_system;
42 Lisp_Object Vterminal_coding_system;
43 Lisp_Object Vcoding_system_for_read;
44 Lisp_Object Vcoding_system_for_write;
45 Lisp_Object Vfile_name_coding_system;
47 /* Table of symbols identifying each coding category. */
48 Lisp_Object coding_category_symbol[CODING_CATEGORY_LAST + 1];
50 /* Coding system currently associated with each coding category. */
51 Lisp_Object coding_category_system[CODING_CATEGORY_LAST + 1];
53 /* Table of all coding categories in decreasing order of priority.
54 This describes a permutation of the possible coding categories. */
55 int coding_category_by_priority[CODING_CATEGORY_LAST + 1];
57 Lisp_Object Qcoding_system_p;
59 Lisp_Object Qraw_text, Qno_conversion, Qccl, Qiso2022;
60 /* Qinternal in general.c */
62 Lisp_Object Qmnemonic, Qeol_type;
63 Lisp_Object Qcr, Qcrlf, Qlf;
64 Lisp_Object Qeol_cr, Qeol_crlf, Qeol_lf;
65 Lisp_Object Qpost_read_conversion;
66 Lisp_Object Qpre_write_conversion;
69 Lisp_Object Qucs4, Qutf8;
70 Lisp_Object Qbig5, Qshift_jis;
71 Lisp_Object Qcharset_g0, Qcharset_g1, Qcharset_g2, Qcharset_g3;
72 Lisp_Object Qforce_g0_on_output, Qforce_g1_on_output;
73 Lisp_Object Qforce_g2_on_output, Qforce_g3_on_output;
74 Lisp_Object Qno_iso6429;
75 Lisp_Object Qinput_charset_conversion, Qoutput_charset_conversion;
76 Lisp_Object Qctext, Qescape_quoted;
77 Lisp_Object Qshort, Qno_ascii_eol, Qno_ascii_cntl, Qseven, Qlock_shift;
79 Lisp_Object Qencode, Qdecode;
81 Lisp_Object Vcoding_system_hash_table;
83 int enable_multibyte_characters;
86 /* Additional information used by the ISO2022 decoder and detector. */
87 struct iso2022_decoder
89 /* CHARSET holds the character sets currently assigned to the G0
90 through G3 variables. It is initialized from the array
91 INITIAL_CHARSET in CODESYS. */
92 Lisp_Object charset[4];
94 /* Which registers are currently invoked into the left (GL) and
95 right (GR) halves of the 8-bit encoding space? */
96 int register_left, register_right;
98 /* ISO_ESC holds a value indicating part of an escape sequence
99 that has already been seen. */
100 enum iso_esc_flag esc;
102 /* This records the bytes we've seen so far in an escape sequence,
103 in case the sequence is invalid (we spit out the bytes unchanged). */
104 unsigned char esc_bytes[8];
106 /* Index for next byte to store in ISO escape sequence. */
109 #ifdef ENABLE_COMPOSITE_CHARS
110 /* Stuff seen so far when composing a string. */
111 unsigned_char_dynarr *composite_chars;
114 /* If we saw an invalid designation sequence for a particular
115 register, we flag it here and switch to ASCII. The next time we
116 see a valid designation for this register, we turn off the flag
117 and do the designation normally, but pretend the sequence was
118 invalid. The effect of all this is that (most of the time) the
119 escape sequences for both the switch to the unknown charset, and
120 the switch back to the known charset, get inserted literally into
121 the buffer and saved out as such. The hope is that we can
122 preserve the escape sequences so that the resulting written out
123 file makes sense. If we don't do any of this, the designation
124 to the invalid charset will be preserved but that switch back
125 to the known charset will probably get eaten because it was
126 the same charset that was already present in the register. */
127 unsigned char invalid_designated[4];
129 /* We try to do similar things as above for direction-switching
130 sequences. If we encountered a direction switch while an
131 invalid designation was present, or an invalid designation
132 just after a direction switch (i.e. no valid designation
133 encountered yet), we insert the direction-switch escape
134 sequence literally into the output stream, and later on
135 insert the corresponding direction-restoring escape sequence
137 unsigned int switched_dir_and_no_valid_charset_yet :1;
138 unsigned int invalid_switch_dir :1;
140 /* Tells the decoder to output the escape sequence literally
141 even though it was valid. Used in the games we play to
142 avoid lossage when we encounter invalid designations. */
143 unsigned int output_literally :1;
144 /* We encountered a direction switch followed by an invalid
145 designation. We didn't output the direction switch
146 literally because we didn't know about the invalid designation;
147 but we have to do so now. */
148 unsigned int output_direction_sequence :1;
151 EXFUN (Fcopy_coding_system, 2);
153 struct detection_state;
156 text_encode_generic (Lstream *encoding, CONST unsigned char *src,
157 unsigned_char_dynarr *dst, unsigned int n);
159 static int detect_coding_sjis (struct detection_state *st,
160 CONST unsigned char *src,
162 static void decode_coding_sjis (Lstream *decoding,
163 CONST unsigned char *src,
164 unsigned_char_dynarr *dst,
166 void char_encode_shift_jis (struct encoding_stream *str, Emchar c,
167 unsigned_char_dynarr *dst, unsigned int *flags);
168 void char_finish_shift_jis (struct encoding_stream *str,
169 unsigned_char_dynarr *dst, unsigned int *flags);
171 static int detect_coding_big5 (struct detection_state *st,
172 CONST unsigned char *src,
174 static void decode_coding_big5 (Lstream *decoding,
175 CONST unsigned char *src,
176 unsigned_char_dynarr *dst, unsigned int n);
177 static void encode_coding_big5 (Lstream *encoding,
178 CONST unsigned char *src,
179 unsigned_char_dynarr *dst, unsigned int n);
180 static int detect_coding_ucs4 (struct detection_state *st,
181 CONST unsigned char *src,
183 static void decode_coding_ucs4 (Lstream *decoding,
184 CONST unsigned char *src,
185 unsigned_char_dynarr *dst, unsigned int n);
186 void char_encode_ucs4 (struct encoding_stream *str, Emchar c,
187 unsigned_char_dynarr *dst, unsigned int *flags);
188 void char_finish_ucs4 (struct encoding_stream *str,
189 unsigned_char_dynarr *dst, unsigned int *flags);
191 static int detect_coding_utf8 (struct detection_state *st,
192 CONST unsigned char *src,
194 static void decode_coding_utf8 (Lstream *decoding,
195 CONST unsigned char *src,
196 unsigned_char_dynarr *dst, unsigned int n);
197 void char_encode_utf8 (struct encoding_stream *str, Emchar c,
198 unsigned_char_dynarr *dst, unsigned int *flags);
199 void char_finish_utf8 (struct encoding_stream *str,
200 unsigned_char_dynarr *dst, unsigned int *flags);
202 static int postprocess_iso2022_mask (int mask);
203 static void reset_iso2022 (Lisp_Object coding_system,
204 struct iso2022_decoder *iso);
205 static int detect_coding_iso2022 (struct detection_state *st,
206 CONST unsigned char *src,
208 static void decode_coding_iso2022 (Lstream *decoding,
209 CONST unsigned char *src,
210 unsigned_char_dynarr *dst, unsigned int n);
211 void char_encode_iso2022 (struct encoding_stream *str, Emchar c,
212 unsigned_char_dynarr *dst, unsigned int *flags);
213 void char_finish_iso2022 (struct encoding_stream *str,
214 unsigned_char_dynarr *dst, unsigned int *flags);
216 static void decode_coding_no_conversion (Lstream *decoding,
217 CONST unsigned char *src,
218 unsigned_char_dynarr *dst,
220 static void encode_coding_no_conversion (Lstream *encoding,
221 CONST unsigned char *src,
222 unsigned_char_dynarr *dst,
224 static void mule_decode (Lstream *decoding, CONST unsigned char *src,
225 unsigned_char_dynarr *dst, unsigned int n);
226 static void mule_encode (Lstream *encoding, CONST unsigned char *src,
227 unsigned_char_dynarr *dst, unsigned int n);
229 typedef struct codesys_prop codesys_prop;
238 Dynarr_declare (codesys_prop);
239 } codesys_prop_dynarr;
241 codesys_prop_dynarr *the_codesys_prop_dynarr;
243 enum codesys_prop_enum
246 CODESYS_PROP_ISO2022,
251 /************************************************************************/
252 /* Coding system functions */
253 /************************************************************************/
255 static Lisp_Object mark_coding_system (Lisp_Object, void (*) (Lisp_Object));
256 static void print_coding_system (Lisp_Object, Lisp_Object, int);
257 static void finalize_coding_system (void *header, int for_disksave);
260 static const struct lrecord_description ccs_description_1[] = {
261 { XD_LISP_OBJECT, offsetof(charset_conversion_spec, from_charset), 2 },
265 static const struct struct_description ccs_description = {
266 sizeof(charset_conversion_spec),
270 static const struct lrecord_description ccsd_description_1[] = {
271 XD_DYNARR_DESC(charset_conversion_spec_dynarr, &ccs_description),
275 static const struct struct_description ccsd_description = {
276 sizeof(charset_conversion_spec_dynarr),
281 static const struct lrecord_description coding_system_description[] = {
282 { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, name), 2 },
283 { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, mnemonic), 3 },
284 { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, eol_lf), 3 },
286 { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, iso2022.initial_charset), 4 },
287 { XD_STRUCT_PTR, offsetof(struct Lisp_Coding_System, iso2022.input_conv), 1, &ccsd_description },
288 { XD_STRUCT_PTR, offsetof(struct Lisp_Coding_System, iso2022.output_conv), 1, &ccsd_description },
289 { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, ccl.decode), 2 },
294 DEFINE_LRECORD_IMPLEMENTATION ("coding-system", coding_system,
295 mark_coding_system, print_coding_system,
296 finalize_coding_system,
297 0, 0, coding_system_description,
298 struct Lisp_Coding_System);
301 mark_coding_system (Lisp_Object obj, void (*markobj) (Lisp_Object))
303 Lisp_Coding_System *codesys = XCODING_SYSTEM (obj);
305 markobj (CODING_SYSTEM_NAME (codesys));
306 markobj (CODING_SYSTEM_DOC_STRING (codesys));
307 markobj (CODING_SYSTEM_MNEMONIC (codesys));
308 markobj (CODING_SYSTEM_EOL_LF (codesys));
309 markobj (CODING_SYSTEM_EOL_CRLF (codesys));
310 markobj (CODING_SYSTEM_EOL_CR (codesys));
312 switch (CODING_SYSTEM_TYPE (codesys))
316 case CODESYS_ISO2022:
317 for (i = 0; i < 4; i++)
318 markobj (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i));
319 if (codesys->iso2022.input_conv)
321 for (i = 0; i < Dynarr_length (codesys->iso2022.input_conv); i++)
323 struct charset_conversion_spec *ccs =
324 Dynarr_atp (codesys->iso2022.input_conv, i);
325 markobj (ccs->from_charset);
326 markobj (ccs->to_charset);
329 if (codesys->iso2022.output_conv)
331 for (i = 0; i < Dynarr_length (codesys->iso2022.output_conv); i++)
333 struct charset_conversion_spec *ccs =
334 Dynarr_atp (codesys->iso2022.output_conv, i);
335 markobj (ccs->from_charset);
336 markobj (ccs->to_charset);
342 markobj (CODING_SYSTEM_CCL_DECODE (codesys));
343 markobj (CODING_SYSTEM_CCL_ENCODE (codesys));
350 markobj (CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys));
351 return CODING_SYSTEM_POST_READ_CONVERSION (codesys);
355 print_coding_system (Lisp_Object obj, Lisp_Object printcharfun,
358 Lisp_Coding_System *c = XCODING_SYSTEM (obj);
360 error ("printing unreadable object #<coding_system 0x%x>",
363 write_c_string ("#<coding_system ", printcharfun);
364 print_internal (c->name, printcharfun, 1);
365 write_c_string (">", printcharfun);
369 finalize_coding_system (void *header, int for_disksave)
371 Lisp_Coding_System *c = (Lisp_Coding_System *) header;
372 /* Since coding systems never go away, this function is not
373 necessary. But it would be necessary if we changed things
374 so that coding systems could go away. */
375 if (!for_disksave) /* see comment in lstream.c */
377 switch (CODING_SYSTEM_TYPE (c))
380 case CODESYS_ISO2022:
381 if (c->iso2022.input_conv)
383 Dynarr_free (c->iso2022.input_conv);
384 c->iso2022.input_conv = 0;
386 if (c->iso2022.output_conv)
388 Dynarr_free (c->iso2022.output_conv);
389 c->iso2022.output_conv = 0;
400 symbol_to_eol_type (Lisp_Object symbol)
402 CHECK_SYMBOL (symbol);
403 if (NILP (symbol)) return EOL_AUTODETECT;
404 if (EQ (symbol, Qlf)) return EOL_LF;
405 if (EQ (symbol, Qcrlf)) return EOL_CRLF;
406 if (EQ (symbol, Qcr)) return EOL_CR;
408 signal_simple_error ("Unrecognized eol type", symbol);
409 return EOL_AUTODETECT; /* not reached */
413 eol_type_to_symbol (enum eol_type type)
418 case EOL_LF: return Qlf;
419 case EOL_CRLF: return Qcrlf;
420 case EOL_CR: return Qcr;
421 case EOL_AUTODETECT: return Qnil;
426 setup_eol_coding_systems (Lisp_Coding_System *codesys)
428 Lisp_Object codesys_obj;
429 int len = string_length (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name);
430 char *codesys_name = (char *) alloca (len + 7);
432 char *codesys_mnemonic=0;
434 Lisp_Object codesys_name_sym, sub_codesys_obj;
438 XSETCODING_SYSTEM (codesys_obj, codesys);
440 memcpy (codesys_name,
441 string_data (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name), len);
443 if (STRINGP (CODING_SYSTEM_MNEMONIC (codesys)))
445 mlen = XSTRING_LENGTH (CODING_SYSTEM_MNEMONIC (codesys));
446 codesys_mnemonic = (char *) alloca (mlen + 7);
447 memcpy (codesys_mnemonic,
448 XSTRING_DATA (CODING_SYSTEM_MNEMONIC (codesys)), mlen);
451 #define DEFINE_SUB_CODESYS(op_sys, op_sys_abbr, Type) do { \
452 strcpy (codesys_name + len, "-" op_sys); \
454 strcpy (codesys_mnemonic + mlen, op_sys_abbr); \
455 codesys_name_sym = intern (codesys_name); \
456 sub_codesys_obj = Fcopy_coding_system (codesys_obj, codesys_name_sym); \
457 XCODING_SYSTEM_EOL_TYPE (sub_codesys_obj) = Type; \
459 XCODING_SYSTEM_MNEMONIC(sub_codesys_obj) = \
460 build_string (codesys_mnemonic); \
461 CODING_SYSTEM_##Type (codesys) = sub_codesys_obj; \
464 DEFINE_SUB_CODESYS("unix", "", EOL_LF);
465 DEFINE_SUB_CODESYS("dos", ":T", EOL_CRLF);
466 DEFINE_SUB_CODESYS("mac", ":t", EOL_CR);
469 DEFUN ("coding-system-p", Fcoding_system_p, 1, 1, 0, /*
470 Return t if OBJECT is a coding system.
471 A coding system is an object that defines how text containing multiple
472 character sets is encoded into a stream of (typically 8-bit) bytes.
473 The coding system is used to decode the stream into a series of
474 characters (which may be from multiple charsets) when the text is read
475 from a file or process, and is used to encode the text back into the
476 same format when it is written out to a file or process.
478 For example, many ISO2022-compliant coding systems (such as Compound
479 Text, which is used for inter-client data under the X Window System)
480 use escape sequences to switch between different charsets -- Japanese
481 Kanji, for example, is invoked with "ESC $ ( B"; ASCII is invoked
482 with "ESC ( B"; and Cyrillic is invoked with "ESC - L". See
483 `make-coding-system' for more information.
485 Coding systems are normally identified using a symbol, and the
486 symbol is accepted in place of the actual coding system object whenever
487 a coding system is called for. (This is similar to how faces work.)
491 return CODING_SYSTEMP (object) ? Qt : Qnil;
494 DEFUN ("find-coding-system", Ffind_coding_system, 1, 1, 0, /*
495 Retrieve the coding system of the given name.
497 If CODING-SYSTEM-OR-NAME is a coding-system object, it is simply
498 returned. Otherwise, CODING-SYSTEM-OR-NAME should be a symbol.
499 If there is no such coding system, nil is returned. Otherwise the
500 associated coding system object is returned.
502 (coding_system_or_name))
504 if (CODING_SYSTEMP (coding_system_or_name))
505 return coding_system_or_name;
507 if (NILP (coding_system_or_name))
508 coding_system_or_name = Qbinary;
510 CHECK_SYMBOL (coding_system_or_name);
512 return Fgethash (coding_system_or_name, Vcoding_system_hash_table, Qnil);
515 DEFUN ("get-coding-system", Fget_coding_system, 1, 1, 0, /*
516 Retrieve the coding system of the given name.
517 Same as `find-coding-system' except that if there is no such
518 coding system, an error is signaled instead of returning nil.
522 Lisp_Object coding_system = Ffind_coding_system (name);
524 if (NILP (coding_system))
525 signal_simple_error ("No such coding system", name);
526 return coding_system;
529 /* We store the coding systems in hash tables with the names as the key and the
530 actual coding system object as the value. Occasionally we need to use them
531 in a list format. These routines provide us with that. */
532 struct coding_system_list_closure
534 Lisp_Object *coding_system_list;
538 add_coding_system_to_list_mapper (Lisp_Object key, Lisp_Object value,
539 void *coding_system_list_closure)
541 /* This function can GC */
542 struct coding_system_list_closure *cscl =
543 (struct coding_system_list_closure *) coding_system_list_closure;
544 Lisp_Object *coding_system_list = cscl->coding_system_list;
546 *coding_system_list = Fcons (XCODING_SYSTEM (value)->name,
547 *coding_system_list);
551 DEFUN ("coding-system-list", Fcoding_system_list, 0, 0, 0, /*
552 Return a list of the names of all defined coding systems.
556 Lisp_Object coding_system_list = Qnil;
558 struct coding_system_list_closure coding_system_list_closure;
560 GCPRO1 (coding_system_list);
561 coding_system_list_closure.coding_system_list = &coding_system_list;
562 elisp_maphash (add_coding_system_to_list_mapper, Vcoding_system_hash_table,
563 &coding_system_list_closure);
566 return coding_system_list;
569 DEFUN ("coding-system-name", Fcoding_system_name, 1, 1, 0, /*
570 Return the name of the given coding system.
574 coding_system = Fget_coding_system (coding_system);
575 return XCODING_SYSTEM_NAME (coding_system);
578 static Lisp_Coding_System *
579 allocate_coding_system (enum coding_system_type type, Lisp_Object name)
581 Lisp_Coding_System *codesys =
582 alloc_lcrecord_type (Lisp_Coding_System, &lrecord_coding_system);
584 zero_lcrecord (codesys);
585 CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = Qnil;
586 CODING_SYSTEM_POST_READ_CONVERSION (codesys) = Qnil;
587 CODING_SYSTEM_EOL_TYPE (codesys) = EOL_AUTODETECT;
588 CODING_SYSTEM_EOL_CRLF (codesys) = Qnil;
589 CODING_SYSTEM_EOL_CR (codesys) = Qnil;
590 CODING_SYSTEM_EOL_LF (codesys) = Qnil;
591 CODING_SYSTEM_TYPE (codesys) = type;
592 CODING_SYSTEM_MNEMONIC (codesys) = Qnil;
594 if (type == CODESYS_ISO2022)
597 for (i = 0; i < 4; i++)
598 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i) = Qnil;
600 else if (type == CODESYS_CCL)
602 CODING_SYSTEM_CCL_DECODE (codesys) = Qnil;
603 CODING_SYSTEM_CCL_ENCODE (codesys) = Qnil;
606 CODING_SYSTEM_NAME (codesys) = name;
612 /* Given a list of charset conversion specs as specified in a Lisp
613 program, parse it into STORE_HERE. */
616 parse_charset_conversion_specs (charset_conversion_spec_dynarr *store_here,
617 Lisp_Object spec_list)
621 EXTERNAL_LIST_LOOP (rest, spec_list)
623 Lisp_Object car = XCAR (rest);
624 Lisp_Object from, to;
625 struct charset_conversion_spec spec;
627 if (!CONSP (car) || !CONSP (XCDR (car)) || !NILP (XCDR (XCDR (car))))
628 signal_simple_error ("Invalid charset conversion spec", car);
629 from = Fget_charset (XCAR (car));
630 to = Fget_charset (XCAR (XCDR (car)));
631 if (XCHARSET_TYPE (from) != XCHARSET_TYPE (to))
632 signal_simple_error_2
633 ("Attempted conversion between different charset types",
635 spec.from_charset = from;
636 spec.to_charset = to;
638 Dynarr_add (store_here, spec);
642 /* Given a dynarr LOAD_HERE of internally-stored charset conversion
643 specs, return the equivalent as the Lisp programmer would see it.
645 If LOAD_HERE is 0, return Qnil. */
648 unparse_charset_conversion_specs (charset_conversion_spec_dynarr *load_here)
655 for (i = 0, result = Qnil; i < Dynarr_length (load_here); i++)
657 struct charset_conversion_spec *ccs = Dynarr_atp (load_here, i);
658 result = Fcons (list2 (ccs->from_charset, ccs->to_charset), result);
661 return Fnreverse (result);
666 DEFUN ("make-coding-system", Fmake_coding_system, 2, 4, 0, /*
667 Register symbol NAME as a coding system.
669 TYPE describes the conversion method used and should be one of
672 Automatic conversion. XEmacs attempts to detect the coding system
675 No conversion. Use this for binary files and such. On output,
676 graphic characters that are not in ASCII or Latin-1 will be
677 replaced by a ?. (For a no-conversion-encoded buffer, these
678 characters will only be present if you explicitly insert them.)
680 Shift-JIS (a Japanese encoding commonly used in PC operating systems).
682 ISO 10646 UCS-4 encoding.
684 ISO 10646 UTF-8 encoding.
686 Any ISO2022-compliant encoding. Among other things, this includes
687 JIS (the Japanese encoding commonly used for e-mail), EUC (the
688 standard Unix encoding for Japanese and other languages), and
689 Compound Text (the encoding used in X11). You can specify more
690 specific information about the conversion with the FLAGS argument.
692 Big5 (the encoding commonly used for Taiwanese).
694 The conversion is performed using a user-written pseudo-code
695 program. CCL (Code Conversion Language) is the name of this
698 Write out or read in the raw contents of the memory representing
699 the buffer's text. This is primarily useful for debugging
700 purposes, and is only enabled when XEmacs has been compiled with
701 DEBUG_XEMACS defined (via the --debug configure option).
702 WARNING: Reading in a file using 'internal conversion can result
703 in an internal inconsistency in the memory representing a
704 buffer's text, which will produce unpredictable results and may
705 cause XEmacs to crash. Under normal circumstances you should
706 never use 'internal conversion.
708 DOC-STRING is a string describing the coding system.
710 PROPS is a property list, describing the specific nature of the
711 character set. Recognized properties are:
714 String to be displayed in the modeline when this coding system is
718 End-of-line conversion to be used. It should be one of
721 Automatically detect the end-of-line type (LF, CRLF,
722 or CR). Also generate subsidiary coding systems named
723 `NAME-unix', `NAME-dos', and `NAME-mac', that are
724 identical to this coding system but have an EOL-TYPE
725 value of 'lf, 'crlf, and 'cr, respectively.
727 The end of a line is marked externally using ASCII LF.
728 Since this is also the way that XEmacs represents an
729 end-of-line internally, specifying this option results
730 in no end-of-line conversion. This is the standard
731 format for Unix text files.
733 The end of a line is marked externally using ASCII
734 CRLF. This is the standard format for MS-DOS text
737 The end of a line is marked externally using ASCII CR.
738 This is the standard format for Macintosh text files.
740 Automatically detect the end-of-line type but do not
741 generate subsidiary coding systems. (This value is
742 converted to nil when stored internally, and
743 `coding-system-property' will return nil.)
745 'post-read-conversion
746 Function called after a file has been read in, to perform the
747 decoding. Called with two arguments, BEG and END, denoting
748 a region of the current buffer to be decoded.
750 'pre-write-conversion
751 Function called before a file is written out, to perform the
752 encoding. Called with two arguments, BEG and END, denoting
753 a region of the current buffer to be encoded.
756 The following additional properties are recognized if TYPE is 'iso2022:
762 The character set initially designated to the G0 - G3 registers.
763 The value should be one of
765 -- A charset object (designate that character set)
766 -- nil (do not ever use this register)
767 -- t (no character set is initially designated to
768 the register, but may be later on; this automatically
769 sets the corresponding `force-g*-on-output' property)
775 If non-nil, send an explicit designation sequence on output before
776 using the specified register.
779 If non-nil, use the short forms "ESC $ @", "ESC $ A", and
780 "ESC $ B" on output in place of the full designation sequences
781 "ESC $ ( @", "ESC $ ( A", and "ESC $ ( B".
784 If non-nil, don't designate ASCII to G0 at each end of line on output.
785 Setting this to non-nil also suppresses other state-resetting that
786 normally happens at the end of a line.
789 If non-nil, don't designate ASCII to G0 before control chars on output.
792 If non-nil, use 7-bit environment on output. Otherwise, use 8-bit
796 If non-nil, use locking-shift (SO/SI) instead of single-shift
797 or designation by escape sequence.
800 If non-nil, don't use ISO6429's direction specification.
803 If non-nil, literal control characters that are the same as
804 the beginning of a recognized ISO2022 or ISO6429 escape sequence
805 (in particular, ESC (0x1B), SO (0x0E), SI (0x0F), SS2 (0x8E),
806 SS3 (0x8F), and CSI (0x9B)) are "quoted" with an escape character
807 so that they can be properly distinguished from an escape sequence.
808 (Note that doing this results in a non-portable encoding.) This
809 encoding flag is used for byte-compiled files. Note that ESC
810 is a good choice for a quoting character because there are no
811 escape sequences whose second byte is a character from the Control-0
812 or Control-1 character sets; this is explicitly disallowed by the
815 'input-charset-conversion
816 A list of conversion specifications, specifying conversion of
817 characters in one charset to another when decoding is performed.
818 Each specification is a list of two elements: the source charset,
819 and the destination charset.
821 'output-charset-conversion
822 A list of conversion specifications, specifying conversion of
823 characters in one charset to another when encoding is performed.
824 The form of each specification is the same as for
825 'input-charset-conversion.
828 The following additional properties are recognized (and required)
832 CCL program used for decoding (converting to internal format).
835 CCL program used for encoding (converting to external format).
837 (name, type, doc_string, props))
839 Lisp_Coding_System *codesys;
840 Lisp_Object rest, key, value;
841 enum coding_system_type ty;
842 int need_to_setup_eol_systems = 1;
844 /* Convert type to constant */
845 if (NILP (type) || EQ (type, Qundecided))
846 { ty = CODESYS_AUTODETECT; }
848 else if (EQ (type, Qshift_jis)) { ty = CODESYS_SHIFT_JIS; }
849 else if (EQ (type, Qiso2022)) { ty = CODESYS_ISO2022; }
850 else if (EQ (type, Qbig5)) { ty = CODESYS_BIG5; }
851 else if (EQ (type, Qucs4)) { ty = CODESYS_UCS4; }
852 else if (EQ (type, Qutf8)) { ty = CODESYS_UTF8; }
853 else if (EQ (type, Qccl)) { ty = CODESYS_CCL; }
855 else if (EQ (type, Qno_conversion)) { ty = CODESYS_NO_CONVERSION; }
857 else if (EQ (type, Qinternal)) { ty = CODESYS_INTERNAL; }
860 signal_simple_error ("Invalid coding system type", type);
864 codesys = allocate_coding_system (ty, name);
866 if (NILP (doc_string))
867 doc_string = build_string ("");
869 CHECK_STRING (doc_string);
870 CODING_SYSTEM_DOC_STRING (codesys) = doc_string;
872 EXTERNAL_PROPERTY_LIST_LOOP (rest, key, value, props)
874 if (EQ (key, Qmnemonic))
877 CHECK_STRING (value);
878 CODING_SYSTEM_MNEMONIC (codesys) = value;
881 else if (EQ (key, Qeol_type))
883 need_to_setup_eol_systems = NILP (value);
886 CODING_SYSTEM_EOL_TYPE (codesys) = symbol_to_eol_type (value);
889 else if (EQ (key, Qpost_read_conversion)) CODING_SYSTEM_POST_READ_CONVERSION (codesys) = value;
890 else if (EQ (key, Qpre_write_conversion)) CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = value;
892 else if (ty == CODESYS_ISO2022)
894 #define FROB_INITIAL_CHARSET(charset_num) \
895 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, charset_num) = \
896 ((EQ (value, Qt) || EQ (value, Qnil)) ? value : Fget_charset (value))
898 if (EQ (key, Qcharset_g0)) FROB_INITIAL_CHARSET (0);
899 else if (EQ (key, Qcharset_g1)) FROB_INITIAL_CHARSET (1);
900 else if (EQ (key, Qcharset_g2)) FROB_INITIAL_CHARSET (2);
901 else if (EQ (key, Qcharset_g3)) FROB_INITIAL_CHARSET (3);
903 #define FROB_FORCE_CHARSET(charset_num) \
904 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (codesys, charset_num) = !NILP (value)
906 else if (EQ (key, Qforce_g0_on_output)) FROB_FORCE_CHARSET (0);
907 else if (EQ (key, Qforce_g1_on_output)) FROB_FORCE_CHARSET (1);
908 else if (EQ (key, Qforce_g2_on_output)) FROB_FORCE_CHARSET (2);
909 else if (EQ (key, Qforce_g3_on_output)) FROB_FORCE_CHARSET (3);
911 #define FROB_BOOLEAN_PROPERTY(prop) \
912 CODING_SYSTEM_ISO2022_##prop (codesys) = !NILP (value)
914 else if (EQ (key, Qshort)) FROB_BOOLEAN_PROPERTY (SHORT);
915 else if (EQ (key, Qno_ascii_eol)) FROB_BOOLEAN_PROPERTY (NO_ASCII_EOL);
916 else if (EQ (key, Qno_ascii_cntl)) FROB_BOOLEAN_PROPERTY (NO_ASCII_CNTL);
917 else if (EQ (key, Qseven)) FROB_BOOLEAN_PROPERTY (SEVEN);
918 else if (EQ (key, Qlock_shift)) FROB_BOOLEAN_PROPERTY (LOCK_SHIFT);
919 else if (EQ (key, Qno_iso6429)) FROB_BOOLEAN_PROPERTY (NO_ISO6429);
920 else if (EQ (key, Qescape_quoted)) FROB_BOOLEAN_PROPERTY (ESCAPE_QUOTED);
922 else if (EQ (key, Qinput_charset_conversion))
924 codesys->iso2022.input_conv =
925 Dynarr_new (charset_conversion_spec);
926 parse_charset_conversion_specs (codesys->iso2022.input_conv,
929 else if (EQ (key, Qoutput_charset_conversion))
931 codesys->iso2022.output_conv =
932 Dynarr_new (charset_conversion_spec);
933 parse_charset_conversion_specs (codesys->iso2022.output_conv,
937 signal_simple_error ("Unrecognized property", key);
939 else if (EQ (type, Qccl))
941 if (EQ (key, Qdecode))
943 CHECK_VECTOR (value);
944 CODING_SYSTEM_CCL_DECODE (codesys) = value;
946 else if (EQ (key, Qencode))
948 CHECK_VECTOR (value);
949 CODING_SYSTEM_CCL_ENCODE (codesys) = value;
952 signal_simple_error ("Unrecognized property", key);
956 signal_simple_error ("Unrecognized property", key);
959 if (need_to_setup_eol_systems)
960 setup_eol_coding_systems (codesys);
963 Lisp_Object codesys_obj;
964 XSETCODING_SYSTEM (codesys_obj, codesys);
965 Fputhash (name, codesys_obj, Vcoding_system_hash_table);
970 DEFUN ("copy-coding-system", Fcopy_coding_system, 2, 2, 0, /*
971 Copy OLD-CODING-SYSTEM to NEW-NAME.
972 If NEW-NAME does not name an existing coding system, a new one will
975 (old_coding_system, new_name))
977 Lisp_Object new_coding_system;
978 old_coding_system = Fget_coding_system (old_coding_system);
979 new_coding_system = Ffind_coding_system (new_name);
980 if (NILP (new_coding_system))
982 XSETCODING_SYSTEM (new_coding_system,
983 allocate_coding_system
984 (XCODING_SYSTEM_TYPE (old_coding_system),
986 Fputhash (new_name, new_coding_system, Vcoding_system_hash_table);
990 Lisp_Coding_System *to = XCODING_SYSTEM (new_coding_system);
991 Lisp_Coding_System *from = XCODING_SYSTEM (old_coding_system);
992 memcpy (((char *) to ) + sizeof (to->header),
993 ((char *) from) + sizeof (from->header),
994 sizeof (*from) - sizeof (from->header));
997 return new_coding_system;
1000 DEFUN ("define-coding-system-alias", Fdefine_coding_system_alias, 2, 2, 0, /*
1001 Define symbol ALIAS as an alias for coding system CODING-SYSTEM.
1003 (alias, coding_system))
1005 CHECK_SYMBOL (alias);
1006 if (!NILP (Ffind_coding_system (alias)))
1007 signal_simple_error ("Symbol already names a coding system", alias);
1008 coding_system = Fget_coding_system (coding_system);
1009 Fputhash (alias, coding_system, Vcoding_system_hash_table);
1011 /* Set up aliases for subsidiaries. */
1012 if (XCODING_SYSTEM_EOL_TYPE (coding_system) == EOL_AUTODETECT)
1015 XSETSTRING (str, symbol_name (XSYMBOL (alias)));
1016 #define FROB(type, name) \
1018 Lisp_Object subsidiary = XCODING_SYSTEM_EOL_##type (coding_system); \
1019 if (!NILP (subsidiary)) \
1020 Fdefine_coding_system_alias \
1021 (Fintern (concat2 (str, build_string (name)), Qnil), subsidiary); \
1024 FROB (CRLF, "-dos");
1028 /* FSF return value is a vector of [ALIAS-unix ALIAS-doc ALIAS-mac],
1029 but it doesn't look intentional, so I'd rather return something
1030 meaningful or nothing at all. */
1035 subsidiary_coding_system (Lisp_Object coding_system, enum eol_type type)
1037 Lisp_Coding_System *cs = XCODING_SYSTEM (coding_system);
1038 Lisp_Object new_coding_system;
1040 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
1041 return coding_system;
1045 case EOL_AUTODETECT: return coding_system;
1046 case EOL_LF: new_coding_system = CODING_SYSTEM_EOL_LF (cs); break;
1047 case EOL_CR: new_coding_system = CODING_SYSTEM_EOL_CR (cs); break;
1048 case EOL_CRLF: new_coding_system = CODING_SYSTEM_EOL_CRLF (cs); break;
1052 return NILP (new_coding_system) ? coding_system : new_coding_system;
1055 DEFUN ("subsidiary-coding-system", Fsubsidiary_coding_system, 2, 2, 0, /*
1056 Return the subsidiary coding system of CODING-SYSTEM with eol type EOL-TYPE.
1058 (coding_system, eol_type))
1060 coding_system = Fget_coding_system (coding_system);
1062 return subsidiary_coding_system (coding_system,
1063 symbol_to_eol_type (eol_type));
1067 /************************************************************************/
1068 /* Coding system accessors */
1069 /************************************************************************/
1071 DEFUN ("coding-system-doc-string", Fcoding_system_doc_string, 1, 1, 0, /*
1072 Return the doc string for CODING-SYSTEM.
1076 coding_system = Fget_coding_system (coding_system);
1077 return XCODING_SYSTEM_DOC_STRING (coding_system);
1080 DEFUN ("coding-system-type", Fcoding_system_type, 1, 1, 0, /*
1081 Return the type of CODING-SYSTEM.
1085 switch (XCODING_SYSTEM_TYPE (Fget_coding_system (coding_system)))
1088 case CODESYS_AUTODETECT: return Qundecided;
1090 case CODESYS_SHIFT_JIS: return Qshift_jis;
1091 case CODESYS_ISO2022: return Qiso2022;
1092 case CODESYS_BIG5: return Qbig5;
1093 case CODESYS_UCS4: return Qucs4;
1094 case CODESYS_UTF8: return Qutf8;
1095 case CODESYS_CCL: return Qccl;
1097 case CODESYS_NO_CONVERSION: return Qno_conversion;
1099 case CODESYS_INTERNAL: return Qinternal;
1106 Lisp_Object coding_system_charset (Lisp_Object coding_system, int gnum)
1109 = XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, gnum);
1111 return CHARSETP (cs) ? XCHARSET_NAME (cs) : Qnil;
1114 DEFUN ("coding-system-charset", Fcoding_system_charset, 2, 2, 0, /*
1115 Return initial charset of CODING-SYSTEM designated to GNUM.
1118 (coding_system, gnum))
1120 coding_system = Fget_coding_system (coding_system);
1123 return coding_system_charset (coding_system, XINT (gnum));
1127 DEFUN ("coding-system-property", Fcoding_system_property, 2, 2, 0, /*
1128 Return the PROP property of CODING-SYSTEM.
1130 (coding_system, prop))
1133 enum coding_system_type type;
1135 coding_system = Fget_coding_system (coding_system);
1136 CHECK_SYMBOL (prop);
1137 type = XCODING_SYSTEM_TYPE (coding_system);
1139 for (i = 0; !ok && i < Dynarr_length (the_codesys_prop_dynarr); i++)
1140 if (EQ (Dynarr_at (the_codesys_prop_dynarr, i).sym, prop))
1143 switch (Dynarr_at (the_codesys_prop_dynarr, i).prop_type)
1145 case CODESYS_PROP_ALL_OK:
1148 case CODESYS_PROP_ISO2022:
1149 if (type != CODESYS_ISO2022)
1151 ("Property only valid in ISO2022 coding systems",
1155 case CODESYS_PROP_CCL:
1156 if (type != CODESYS_CCL)
1158 ("Property only valid in CCL coding systems",
1168 signal_simple_error ("Unrecognized property", prop);
1170 if (EQ (prop, Qname))
1171 return XCODING_SYSTEM_NAME (coding_system);
1172 else if (EQ (prop, Qtype))
1173 return Fcoding_system_type (coding_system);
1174 else if (EQ (prop, Qdoc_string))
1175 return XCODING_SYSTEM_DOC_STRING (coding_system);
1176 else if (EQ (prop, Qmnemonic))
1177 return XCODING_SYSTEM_MNEMONIC (coding_system);
1178 else if (EQ (prop, Qeol_type))
1179 return eol_type_to_symbol (XCODING_SYSTEM_EOL_TYPE (coding_system));
1180 else if (EQ (prop, Qeol_lf))
1181 return XCODING_SYSTEM_EOL_LF (coding_system);
1182 else if (EQ (prop, Qeol_crlf))
1183 return XCODING_SYSTEM_EOL_CRLF (coding_system);
1184 else if (EQ (prop, Qeol_cr))
1185 return XCODING_SYSTEM_EOL_CR (coding_system);
1186 else if (EQ (prop, Qpost_read_conversion))
1187 return XCODING_SYSTEM_POST_READ_CONVERSION (coding_system);
1188 else if (EQ (prop, Qpre_write_conversion))
1189 return XCODING_SYSTEM_PRE_WRITE_CONVERSION (coding_system);
1191 else if (type == CODESYS_ISO2022)
1193 if (EQ (prop, Qcharset_g0))
1194 return coding_system_charset (coding_system, 0);
1195 else if (EQ (prop, Qcharset_g1))
1196 return coding_system_charset (coding_system, 1);
1197 else if (EQ (prop, Qcharset_g2))
1198 return coding_system_charset (coding_system, 2);
1199 else if (EQ (prop, Qcharset_g3))
1200 return coding_system_charset (coding_system, 3);
1202 #define FORCE_CHARSET(charset_num) \
1203 (XCODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT \
1204 (coding_system, charset_num) ? Qt : Qnil)
1206 else if (EQ (prop, Qforce_g0_on_output)) return FORCE_CHARSET (0);
1207 else if (EQ (prop, Qforce_g1_on_output)) return FORCE_CHARSET (1);
1208 else if (EQ (prop, Qforce_g2_on_output)) return FORCE_CHARSET (2);
1209 else if (EQ (prop, Qforce_g3_on_output)) return FORCE_CHARSET (3);
1211 #define LISP_BOOLEAN(prop) \
1212 (XCODING_SYSTEM_ISO2022_##prop (coding_system) ? Qt : Qnil)
1214 else if (EQ (prop, Qshort)) return LISP_BOOLEAN (SHORT);
1215 else if (EQ (prop, Qno_ascii_eol)) return LISP_BOOLEAN (NO_ASCII_EOL);
1216 else if (EQ (prop, Qno_ascii_cntl)) return LISP_BOOLEAN (NO_ASCII_CNTL);
1217 else if (EQ (prop, Qseven)) return LISP_BOOLEAN (SEVEN);
1218 else if (EQ (prop, Qlock_shift)) return LISP_BOOLEAN (LOCK_SHIFT);
1219 else if (EQ (prop, Qno_iso6429)) return LISP_BOOLEAN (NO_ISO6429);
1220 else if (EQ (prop, Qescape_quoted)) return LISP_BOOLEAN (ESCAPE_QUOTED);
1222 else if (EQ (prop, Qinput_charset_conversion))
1224 unparse_charset_conversion_specs
1225 (XCODING_SYSTEM (coding_system)->iso2022.input_conv);
1226 else if (EQ (prop, Qoutput_charset_conversion))
1228 unparse_charset_conversion_specs
1229 (XCODING_SYSTEM (coding_system)->iso2022.output_conv);
1233 else if (type == CODESYS_CCL)
1235 if (EQ (prop, Qdecode))
1236 return XCODING_SYSTEM_CCL_DECODE (coding_system);
1237 else if (EQ (prop, Qencode))
1238 return XCODING_SYSTEM_CCL_ENCODE (coding_system);
1246 return Qnil; /* not reached */
1250 /************************************************************************/
1251 /* Coding category functions */
1252 /************************************************************************/
1255 decode_coding_category (Lisp_Object symbol)
1259 CHECK_SYMBOL (symbol);
1260 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1261 if (EQ (coding_category_symbol[i], symbol))
1264 signal_simple_error ("Unrecognized coding category", symbol);
1265 return 0; /* not reached */
1268 DEFUN ("coding-category-list", Fcoding_category_list, 0, 0, 0, /*
1269 Return a list of all recognized coding categories.
1274 Lisp_Object list = Qnil;
1276 for (i = CODING_CATEGORY_LAST; i >= 0; i--)
1277 list = Fcons (coding_category_symbol[i], list);
1281 DEFUN ("set-coding-priority-list", Fset_coding_priority_list, 1, 1, 0, /*
1282 Change the priority order of the coding categories.
1283 LIST should be list of coding categories, in descending order of
1284 priority. Unspecified coding categories will be lower in priority
1285 than all specified ones, in the same relative order they were in
1290 int category_to_priority[CODING_CATEGORY_LAST + 1];
1294 /* First generate a list that maps coding categories to priorities. */
1296 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1297 category_to_priority[i] = -1;
1299 /* Highest priority comes from the specified list. */
1301 EXTERNAL_LIST_LOOP (rest, list)
1303 int cat = decode_coding_category (XCAR (rest));
1305 if (category_to_priority[cat] >= 0)
1306 signal_simple_error ("Duplicate coding category in list", XCAR (rest));
1307 category_to_priority[cat] = i++;
1310 /* Now go through the existing categories by priority to retrieve
1311 the categories not yet specified and preserve their priority
1313 for (j = 0; j <= CODING_CATEGORY_LAST; j++)
1315 int cat = coding_category_by_priority[j];
1316 if (category_to_priority[cat] < 0)
1317 category_to_priority[cat] = i++;
1320 /* Now we need to construct the inverse of the mapping we just
1323 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1324 coding_category_by_priority[category_to_priority[i]] = i;
1326 /* Phew! That was confusing. */
1330 DEFUN ("coding-priority-list", Fcoding_priority_list, 0, 0, 0, /*
1331 Return a list of coding categories in descending order of priority.
1336 Lisp_Object list = Qnil;
1338 for (i = CODING_CATEGORY_LAST; i >= 0; i--)
1339 list = Fcons (coding_category_symbol[coding_category_by_priority[i]],
1344 DEFUN ("set-coding-category-system", Fset_coding_category_system, 2, 2, 0, /*
1345 Change the coding system associated with a coding category.
1347 (coding_category, coding_system))
1349 int cat = decode_coding_category (coding_category);
1351 coding_system = Fget_coding_system (coding_system);
1352 coding_category_system[cat] = coding_system;
1356 DEFUN ("coding-category-system", Fcoding_category_system, 1, 1, 0, /*
1357 Return the coding system associated with a coding category.
1361 int cat = decode_coding_category (coding_category);
1362 Lisp_Object sys = coding_category_system[cat];
1365 return XCODING_SYSTEM_NAME (sys);
1370 /************************************************************************/
1371 /* Detecting the encoding of data */
1372 /************************************************************************/
1374 struct detection_state
1376 enum eol_type eol_type;
1412 struct iso2022_decoder iso;
1414 int high_byte_count;
1415 unsigned int saw_single_shift:1;
1428 acceptable_control_char_p (int c)
1432 /* Allow and ignore control characters that you might
1433 reasonably see in a text file */
1438 case 8: /* backspace */
1439 case 11: /* vertical tab */
1440 case 12: /* form feed */
1441 case 26: /* MS-DOS C-z junk */
1442 case 31: /* '^_' -- for info */
1450 mask_has_at_most_one_bit_p (int mask)
1452 /* Perhaps the only thing useful you learn from intensive Microsoft
1453 technical interviews */
1454 return (mask & (mask - 1)) == 0;
1457 static enum eol_type
1458 detect_eol_type (struct detection_state *st, CONST unsigned char *src,
1467 st->eol.just_saw_cr = 1;
1472 if (st->eol.just_saw_cr)
1474 else if (st->eol.seen_anything)
1477 else if (st->eol.just_saw_cr)
1479 st->eol.just_saw_cr = 0;
1481 st->eol.seen_anything = 1;
1484 return EOL_AUTODETECT;
1487 /* Attempt to determine the encoding and EOL type of the given text.
1488 Before calling this function for the first type, you must initialize
1489 st->eol_type as appropriate and initialize st->mask to ~0.
1491 st->eol_type holds the determined EOL type, or EOL_AUTODETECT if
1494 st->mask holds the determined coding category mask, or ~0 if only
1495 ASCII has been seen so far.
1499 0 == st->eol_type is EOL_AUTODETECT and/or more than coding category
1500 is present in st->mask
1501 1 == definitive answers are here for both st->eol_type and st->mask
1505 detect_coding_type (struct detection_state *st, CONST unsigned char *src,
1506 unsigned int n, int just_do_eol)
1510 if (st->eol_type == EOL_AUTODETECT)
1511 st->eol_type = detect_eol_type (st, src, n);
1514 return st->eol_type != EOL_AUTODETECT;
1516 if (!st->seen_non_ascii)
1518 for (; n; n--, src++)
1521 if ((c < 0x20 && !acceptable_control_char_p (c)) || c >= 0x80)
1523 st->seen_non_ascii = 1;
1525 st->shift_jis.mask = ~0;
1529 st->iso2022.mask = ~0;
1539 if (!mask_has_at_most_one_bit_p (st->iso2022.mask))
1540 st->iso2022.mask = detect_coding_iso2022 (st, src, n);
1541 if (!mask_has_at_most_one_bit_p (st->shift_jis.mask))
1542 st->shift_jis.mask = detect_coding_sjis (st, src, n);
1543 if (!mask_has_at_most_one_bit_p (st->big5.mask))
1544 st->big5.mask = detect_coding_big5 (st, src, n);
1545 if (!mask_has_at_most_one_bit_p (st->utf8.mask))
1546 st->utf8.mask = detect_coding_utf8 (st, src, n);
1547 if (!mask_has_at_most_one_bit_p (st->ucs4.mask))
1548 st->ucs4.mask = detect_coding_ucs4 (st, src, n);
1551 = st->iso2022.mask | st->shift_jis.mask | st->big5.mask
1552 | st->utf8.mask | st->ucs4.mask;
1555 int retval = mask_has_at_most_one_bit_p (st->mask);
1556 st->mask |= CODING_CATEGORY_NO_CONVERSION_MASK;
1557 return retval && st->eol_type != EOL_AUTODETECT;
1562 coding_system_from_mask (int mask)
1566 /* If the file was entirely or basically ASCII, use the
1567 default value of `buffer-file-coding-system'. */
1568 Lisp_Object retval =
1569 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system;
1572 retval = Ffind_coding_system (retval);
1576 (Qbad_variable, Qwarning,
1577 "Invalid `default-buffer-file-coding-system', set to nil");
1578 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system = Qnil;
1582 retval = Fget_coding_system (Qraw_text);
1590 mask = postprocess_iso2022_mask (mask);
1592 /* Look through the coding categories by priority and find
1593 the first one that is allowed. */
1594 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1596 cat = coding_category_by_priority[i];
1597 if ((mask & (1 << cat)) &&
1598 !NILP (coding_category_system[cat]))
1602 return coding_category_system[cat];
1604 return Fget_coding_system (Qraw_text);
1608 /* Given a seekable read stream and potential coding system and EOL type
1609 as specified, do any autodetection that is called for. If the
1610 coding system and/or EOL type are not autodetect, they will be left
1611 alone; but this function will never return an autodetect coding system
1614 This function does not automatically fetch subsidiary coding systems;
1615 that should be unnecessary with the explicit eol-type argument. */
1618 determine_real_coding_system (Lstream *stream, Lisp_Object *codesys_in_out,
1619 enum eol_type *eol_type_in_out)
1621 struct detection_state decst;
1623 if (*eol_type_in_out == EOL_AUTODETECT)
1624 *eol_type_in_out = XCODING_SYSTEM_EOL_TYPE (*codesys_in_out);
1627 decst.eol_type = *eol_type_in_out;
1630 /* If autodetection is called for, do it now. */
1631 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT ||
1632 *eol_type_in_out == EOL_AUTODETECT)
1634 unsigned char random_buffer[4096];
1636 Lisp_Object coding_system = Qnil;
1638 nread = Lstream_read (stream, random_buffer, sizeof (random_buffer));
1641 unsigned char *cp = random_buffer;
1643 while (cp < random_buffer + nread)
1645 if ((*cp++ == 'c') && (cp < random_buffer + nread) &&
1646 (*cp++ == 'o') && (cp < random_buffer + nread) &&
1647 (*cp++ == 'd') && (cp < random_buffer + nread) &&
1648 (*cp++ == 'i') && (cp < random_buffer + nread) &&
1649 (*cp++ == 'n') && (cp < random_buffer + nread) &&
1650 (*cp++ == 'g') && (cp < random_buffer + nread) &&
1651 (*cp++ == ':') && (cp < random_buffer + nread))
1653 unsigned char coding_system_name[4096 - 6];
1654 unsigned char *np = coding_system_name;
1656 while ( (cp < random_buffer + nread)
1657 && ((*cp == ' ') || (*cp == '\t')) )
1661 while ( (cp < random_buffer + nread) &&
1662 (*cp != ' ') && (*cp != '\t') && (*cp != ';') )
1668 = Ffind_coding_system (intern (coding_system_name));
1672 if (EQ(coding_system, Qnil))
1674 if (detect_coding_type (&decst, random_buffer, nread,
1675 XCODING_SYSTEM_TYPE (*codesys_in_out)
1676 != CODESYS_AUTODETECT))
1678 nread = Lstream_read (stream,
1679 random_buffer, sizeof (random_buffer));
1684 *eol_type_in_out = decst.eol_type;
1685 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT)
1687 if (EQ(coding_system, Qnil))
1688 *codesys_in_out = coding_system_from_mask (decst.mask);
1690 *codesys_in_out = coding_system;
1693 /* If we absolutely can't determine the EOL type, just assume LF. */
1694 if (*eol_type_in_out == EOL_AUTODETECT)
1695 *eol_type_in_out = EOL_LF;
1697 Lstream_rewind (stream);
1700 DEFUN ("detect-coding-region", Fdetect_coding_region, 2, 3, 0, /*
1701 Detect coding system of the text in the region between START and END.
1702 Returned a list of possible coding systems ordered by priority.
1703 If only ASCII characters are found, it returns 'undecided or one of
1704 its subsidiary coding systems according to a detected end-of-line
1705 type. Optional arg BUFFER defaults to the current buffer.
1707 (start, end, buffer))
1709 Lisp_Object val = Qnil;
1710 struct buffer *buf = decode_buffer (buffer, 0);
1712 Lisp_Object instream, lb_instream;
1713 Lstream *istr, *lb_istr;
1714 struct detection_state decst;
1715 struct gcpro gcpro1, gcpro2;
1717 get_buffer_range_char (buf, start, end, &b, &e, 0);
1718 lb_instream = make_lisp_buffer_input_stream (buf, b, e, 0);
1719 lb_istr = XLSTREAM (lb_instream);
1720 instream = make_encoding_input_stream (lb_istr, Fget_coding_system (Qbinary));
1721 istr = XLSTREAM (instream);
1722 GCPRO2 (instream, lb_instream);
1724 decst.eol_type = EOL_AUTODETECT;
1728 unsigned char random_buffer[4096];
1729 int nread = Lstream_read (istr, random_buffer, sizeof (random_buffer));
1733 if (detect_coding_type (&decst, random_buffer, nread, 0))
1737 if (decst.mask == ~0)
1738 val = subsidiary_coding_system (Fget_coding_system (Qundecided),
1746 decst.mask = postprocess_iso2022_mask (decst.mask);
1748 for (i = CODING_CATEGORY_LAST; i >= 0; i--)
1750 int sys = coding_category_by_priority[i];
1751 if (decst.mask & (1 << sys))
1753 Lisp_Object codesys = coding_category_system[sys];
1754 if (!NILP (codesys))
1755 codesys = subsidiary_coding_system (codesys, decst.eol_type);
1756 val = Fcons (codesys, val);
1760 Lstream_close (istr);
1762 Lstream_delete (istr);
1763 Lstream_delete (lb_istr);
1768 /************************************************************************/
1769 /* Converting to internal Mule format ("decoding") */
1770 /************************************************************************/
1772 /* A decoding stream is a stream used for decoding text (i.e.
1773 converting from some external format to internal format).
1774 The decoding-stream object keeps track of the actual coding
1775 stream, the stream that is at the other end, and data that
1776 needs to be persistent across the lifetime of the stream. */
1778 /* Handle the EOL stuff related to just-read-in character C.
1779 EOL_TYPE is the EOL type of the coding stream.
1780 FLAGS is the current value of FLAGS in the coding stream, and may
1781 be modified by this macro. (The macro only looks at the
1782 CODING_STATE_CR flag.) DST is the Dynarr to which the decoded
1783 bytes are to be written. You need to also define a local goto
1784 label "label_continue_loop" that is at the end of the main
1785 character-reading loop.
1787 If C is a CR character, then this macro handles it entirely and
1788 jumps to label_continue_loop. Otherwise, this macro does not add
1789 anything to DST, and continues normally. You should continue
1790 processing C normally after this macro. */
1792 #define DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst) \
1796 if (eol_type == EOL_CR) \
1797 Dynarr_add (dst, '\n'); \
1798 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR) \
1799 Dynarr_add (dst, c); \
1801 flags |= CODING_STATE_CR; \
1802 goto label_continue_loop; \
1804 else if (flags & CODING_STATE_CR) \
1805 { /* eol_type == CODING_SYSTEM_EOL_CRLF */ \
1807 Dynarr_add (dst, '\r'); \
1808 flags &= ~CODING_STATE_CR; \
1812 /* C should be a binary character in the range 0 - 255; convert
1813 to internal format and add to Dynarr DST. */
1816 #define DECODE_ADD_BINARY_CHAR(c, dst) \
1818 if (BYTE_ASCII_P (c)) \
1819 Dynarr_add (dst, c); \
1822 Dynarr_add (dst, (c >> 6) | 0xc0); \
1823 Dynarr_add (dst, (c & 0x3f) | 0x80); \
1828 DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst)
1832 Dynarr_add (dst, c);
1834 else if ( c <= 0x7ff )
1836 Dynarr_add (dst, (c >> 6) | 0xc0);
1837 Dynarr_add (dst, (c & 0x3f) | 0x80);
1839 else if ( c <= 0xffff )
1841 Dynarr_add (dst, (c >> 12) | 0xe0);
1842 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
1843 Dynarr_add (dst, (c & 0x3f) | 0x80);
1845 else if ( c <= 0x1fffff )
1847 Dynarr_add (dst, (c >> 18) | 0xf0);
1848 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
1849 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
1850 Dynarr_add (dst, (c & 0x3f) | 0x80);
1852 else if ( c <= 0x3ffffff )
1854 Dynarr_add (dst, (c >> 24) | 0xf8);
1855 Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
1856 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
1857 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
1858 Dynarr_add (dst, (c & 0x3f) | 0x80);
1862 Dynarr_add (dst, (c >> 30) | 0xfc);
1863 Dynarr_add (dst, ((c >> 24) & 0x3f) | 0x80);
1864 Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
1865 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
1866 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
1867 Dynarr_add (dst, (c & 0x3f) | 0x80);
1871 #define DECODE_ADD_BINARY_CHAR(c, dst) \
1873 if (BYTE_ASCII_P (c)) \
1874 Dynarr_add (dst, c); \
1875 else if (BYTE_C1_P (c)) \
1877 Dynarr_add (dst, LEADING_BYTE_CONTROL_1); \
1878 Dynarr_add (dst, c + 0x20); \
1882 Dynarr_add (dst, LEADING_BYTE_LATIN_ISO8859_1); \
1883 Dynarr_add (dst, c); \
1888 #define DECODE_OUTPUT_PARTIAL_CHAR(ch) \
1892 DECODE_ADD_BINARY_CHAR (ch, dst); \
1897 #define DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst) \
1899 if (flags & CODING_STATE_END) \
1901 DECODE_OUTPUT_PARTIAL_CHAR (ch); \
1902 if (flags & CODING_STATE_CR) \
1903 Dynarr_add (dst, '\r'); \
1907 #define DECODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, decoding)
1909 struct decoding_stream
1911 /* Coding system that governs the conversion. */
1912 Lisp_Coding_System *codesys;
1914 /* Stream that we read the encoded data from or
1915 write the decoded data to. */
1918 /* If we are reading, then we can return only a fixed amount of
1919 data, so if the conversion resulted in too much data, we store it
1920 here for retrieval the next time around. */
1921 unsigned_char_dynarr *runoff;
1923 /* FLAGS holds flags indicating the current state of the decoding.
1924 Some of these flags are dependent on the coding system. */
1927 /* CH holds a partially built-up character. Since we only deal
1928 with one- and two-byte characters at the moment, we only use
1929 this to store the first byte of a two-byte character. */
1932 /* EOL_TYPE specifies the type of end-of-line conversion that
1933 currently applies. We need to keep this separate from the
1934 EOL type stored in CODESYS because the latter might indicate
1935 automatic EOL-type detection while the former will always
1936 indicate a particular EOL type. */
1937 enum eol_type eol_type;
1939 /* Additional ISO2022 information. We define the structure above
1940 because it's also needed by the detection routines. */
1941 struct iso2022_decoder iso2022;
1943 /* Additional information (the state of the running CCL program)
1944 used by the CCL decoder. */
1945 struct ccl_program ccl;
1947 /* counter for UTF-8 or UCS-4 */
1948 unsigned char counter;
1950 struct detection_state decst;
1953 static int decoding_reader (Lstream *stream, unsigned char *data, size_t size);
1954 static int decoding_writer (Lstream *stream, CONST unsigned char *data, size_t size);
1955 static int decoding_rewinder (Lstream *stream);
1956 static int decoding_seekable_p (Lstream *stream);
1957 static int decoding_flusher (Lstream *stream);
1958 static int decoding_closer (Lstream *stream);
1960 static Lisp_Object decoding_marker (Lisp_Object stream,
1961 void (*markobj) (Lisp_Object));
1963 DEFINE_LSTREAM_IMPLEMENTATION ("decoding", lstream_decoding,
1964 sizeof (struct decoding_stream));
1967 decoding_marker (Lisp_Object stream, void (*markobj) (Lisp_Object))
1969 Lstream *str = DECODING_STREAM_DATA (XLSTREAM (stream))->other_end;
1970 Lisp_Object str_obj;
1972 /* We do not need to mark the coding systems or charsets stored
1973 within the stream because they are stored in a global list
1974 and automatically marked. */
1976 XSETLSTREAM (str_obj, str);
1978 if (str->imp->marker)
1979 return (str->imp->marker) (str_obj, markobj);
1984 /* Read SIZE bytes of data and store it into DATA. We are a decoding stream
1985 so we read data from the other end, decode it, and store it into DATA. */
1988 decoding_reader (Lstream *stream, unsigned char *data, size_t size)
1990 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
1991 unsigned char *orig_data = data;
1993 int error_occurred = 0;
1995 /* We need to interface to mule_decode(), which expects to take some
1996 amount of data and store the result into a Dynarr. We have
1997 mule_decode() store into str->runoff, and take data from there
2000 /* We loop until we have enough data, reading chunks from the other
2001 end and decoding it. */
2004 /* Take data from the runoff if we can. Make sure to take at
2005 most SIZE bytes, and delete the data from the runoff. */
2006 if (Dynarr_length (str->runoff) > 0)
2008 size_t chunk = min (size, (size_t) Dynarr_length (str->runoff));
2009 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2010 Dynarr_delete_many (str->runoff, 0, chunk);
2016 break; /* No more room for data */
2018 if (str->flags & CODING_STATE_END)
2019 /* This means that on the previous iteration, we hit the EOF on
2020 the other end. We loop once more so that mule_decode() can
2021 output any final stuff it may be holding, or any "go back
2022 to a sane state" escape sequences. (This latter makes sense
2023 during encoding.) */
2026 /* Exhausted the runoff, so get some more. DATA has at least
2027 SIZE bytes left of storage in it, so it's OK to read directly
2028 into it. (We'll be overwriting above, after we've decoded it
2029 into the runoff.) */
2030 read_size = Lstream_read (str->other_end, data, size);
2037 /* There might be some more end data produced in the translation.
2038 See the comment above. */
2039 str->flags |= CODING_STATE_END;
2040 mule_decode (stream, data, str->runoff, read_size);
2043 if (data - orig_data == 0)
2044 return error_occurred ? -1 : 0;
2046 return data - orig_data;
2050 decoding_writer (Lstream *stream, CONST unsigned char *data, size_t size)
2052 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2055 /* Decode all our data into the runoff, and then attempt to write
2056 it all out to the other end. Remove whatever chunk we succeeded
2058 mule_decode (stream, data, str->runoff, size);
2059 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
2060 Dynarr_length (str->runoff));
2062 Dynarr_delete_many (str->runoff, 0, retval);
2063 /* Do NOT return retval. The return value indicates how much
2064 of the incoming data was written, not how many bytes were
2070 reset_decoding_stream (struct decoding_stream *str)
2073 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_ISO2022)
2075 Lisp_Object coding_system;
2076 XSETCODING_SYSTEM (coding_system, str->codesys);
2077 reset_iso2022 (coding_system, &str->iso2022);
2079 else if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_CCL)
2081 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_DECODE (str->codesys));
2085 str->flags = str->ch = 0;
2089 decoding_rewinder (Lstream *stream)
2091 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2092 reset_decoding_stream (str);
2093 Dynarr_reset (str->runoff);
2094 return Lstream_rewind (str->other_end);
2098 decoding_seekable_p (Lstream *stream)
2100 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2101 return Lstream_seekable_p (str->other_end);
2105 decoding_flusher (Lstream *stream)
2107 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2108 return Lstream_flush (str->other_end);
2112 decoding_closer (Lstream *stream)
2114 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2115 if (stream->flags & LSTREAM_FL_WRITE)
2117 str->flags |= CODING_STATE_END;
2118 decoding_writer (stream, 0, 0);
2120 Dynarr_free (str->runoff);
2122 #ifdef ENABLE_COMPOSITE_CHARS
2123 if (str->iso2022.composite_chars)
2124 Dynarr_free (str->iso2022.composite_chars);
2127 return Lstream_close (str->other_end);
2131 decoding_stream_coding_system (Lstream *stream)
2133 Lisp_Object coding_system;
2134 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2136 XSETCODING_SYSTEM (coding_system, str->codesys);
2137 return subsidiary_coding_system (coding_system, str->eol_type);
2141 set_decoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
2143 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
2144 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2146 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
2147 str->eol_type = CODING_SYSTEM_EOL_TYPE (cs);
2148 reset_decoding_stream (str);
2151 /* WARNING WARNING WARNING WARNING!!!!! If you open up a decoding
2152 stream for writing, no automatic code detection will be performed.
2153 The reason for this is that automatic code detection requires a
2154 seekable input. Things will also fail if you open a decoding
2155 stream for reading using a non-fully-specified coding system and
2156 a non-seekable input stream. */
2159 make_decoding_stream_1 (Lstream *stream, Lisp_Object codesys,
2162 Lstream *lstr = Lstream_new (lstream_decoding, mode);
2163 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2167 str->other_end = stream;
2168 str->runoff = (unsigned_char_dynarr *) Dynarr_new (unsigned_char);
2169 str->eol_type = EOL_AUTODETECT;
2170 if (!strcmp (mode, "r")
2171 && Lstream_seekable_p (stream))
2172 /* We can determine the coding system now. */
2173 determine_real_coding_system (stream, &codesys, &str->eol_type);
2174 set_decoding_stream_coding_system (lstr, codesys);
2175 str->decst.eol_type = str->eol_type;
2176 str->decst.mask = ~0;
2177 XSETLSTREAM (obj, lstr);
2182 make_decoding_input_stream (Lstream *stream, Lisp_Object codesys)
2184 return make_decoding_stream_1 (stream, codesys, "r");
2188 make_decoding_output_stream (Lstream *stream, Lisp_Object codesys)
2190 return make_decoding_stream_1 (stream, codesys, "w");
2193 /* Note: the decode_coding_* functions all take the same
2194 arguments as mule_decode(), which is to say some SRC data of
2195 size N, which is to be stored into dynamic array DST.
2196 DECODING is the stream within which the decoding is
2197 taking place, but no data is actually read from or
2198 written to that stream; that is handled in decoding_reader()
2199 or decoding_writer(). This allows the same functions to
2200 be used for both reading and writing. */
2203 mule_decode (Lstream *decoding, CONST unsigned char *src,
2204 unsigned_char_dynarr *dst, unsigned int n)
2206 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
2208 /* If necessary, do encoding-detection now. We do this when
2209 we're a writing stream or a non-seekable reading stream,
2210 meaning that we can't just process the whole input,
2211 rewind, and start over. */
2213 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT ||
2214 str->eol_type == EOL_AUTODETECT)
2216 Lisp_Object codesys;
2218 XSETCODING_SYSTEM (codesys, str->codesys);
2219 detect_coding_type (&str->decst, src, n,
2220 CODING_SYSTEM_TYPE (str->codesys) !=
2221 CODESYS_AUTODETECT);
2222 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT &&
2223 str->decst.mask != ~0)
2224 /* #### This is cheesy. What we really ought to do is
2225 buffer up a certain amount of data so as to get a
2226 less random result. */
2227 codesys = coding_system_from_mask (str->decst.mask);
2228 str->eol_type = str->decst.eol_type;
2229 if (XCODING_SYSTEM (codesys) != str->codesys)
2231 /* Preserve the CODING_STATE_END flag in case it was set.
2232 If we erase it, bad things might happen. */
2233 int was_end = str->flags & CODING_STATE_END;
2234 set_decoding_stream_coding_system (decoding, codesys);
2236 str->flags |= CODING_STATE_END;
2240 switch (CODING_SYSTEM_TYPE (str->codesys))
2243 case CODESYS_INTERNAL:
2244 Dynarr_add_many (dst, src, n);
2247 case CODESYS_AUTODETECT:
2248 /* If we got this far and still haven't decided on the coding
2249 system, then do no conversion. */
2250 case CODESYS_NO_CONVERSION:
2251 decode_coding_no_conversion (decoding, src, dst, n);
2254 case CODESYS_SHIFT_JIS:
2255 decode_coding_sjis (decoding, src, dst, n);
2258 decode_coding_big5 (decoding, src, dst, n);
2261 decode_coding_ucs4 (decoding, src, dst, n);
2264 decode_coding_utf8 (decoding, src, dst, n);
2267 str->ccl.last_block = str->flags & CODING_STATE_END;
2268 ccl_driver (&str->ccl, src, dst, n, 0, CCL_MODE_DECODING);
2270 case CODESYS_ISO2022:
2271 decode_coding_iso2022 (decoding, src, dst, n);
2279 DEFUN ("decode-coding-region", Fdecode_coding_region, 3, 4, 0, /*
2280 Decode the text between START and END which is encoded in CODING-SYSTEM.
2281 This is useful if you've read in encoded text from a file without decoding
2282 it (e.g. you read in a JIS-formatted file but used the `binary' or
2283 `no-conversion' coding system, so that it shows up as "^[$B!<!+^[(B").
2284 Return length of decoded text.
2285 BUFFER defaults to the current buffer if unspecified.
2287 (start, end, coding_system, buffer))
2290 struct buffer *buf = decode_buffer (buffer, 0);
2291 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2292 Lstream *istr, *ostr;
2293 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2295 get_buffer_range_char (buf, start, end, &b, &e, 0);
2297 barf_if_buffer_read_only (buf, b, e);
2299 coding_system = Fget_coding_system (coding_system);
2300 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2301 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
2302 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
2304 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
2305 Fget_coding_system (Qbinary));
2306 istr = XLSTREAM (instream);
2307 ostr = XLSTREAM (outstream);
2308 GCPRO4 (instream, lb_outstream, de_outstream, outstream);
2310 /* The chain of streams looks like this:
2312 [BUFFER] <----- send through
2313 ------> [ENCODE AS BINARY]
2314 ------> [DECODE AS SPECIFIED]
2320 char tempbuf[1024]; /* some random amount */
2321 Bufpos newpos, even_newer_pos;
2322 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
2323 int size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
2327 newpos = lisp_buffer_stream_startpos (istr);
2328 Lstream_write (ostr, tempbuf, size_in_bytes);
2329 even_newer_pos = lisp_buffer_stream_startpos (istr);
2330 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
2333 Lstream_close (istr);
2334 Lstream_close (ostr);
2336 Lstream_delete (istr);
2337 Lstream_delete (ostr);
2338 Lstream_delete (XLSTREAM (de_outstream));
2339 Lstream_delete (XLSTREAM (lb_outstream));
2344 /************************************************************************/
2345 /* Converting to an external encoding ("encoding") */
2346 /************************************************************************/
2348 /* An encoding stream is an output stream. When you create the
2349 stream, you specify the coding system that governs the encoding
2350 and another stream that the resulting encoded data is to be
2351 sent to, and then start sending data to it. */
2353 #define ENCODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, encoding)
2355 struct encoding_stream
2357 /* Coding system that governs the conversion. */
2358 Lisp_Coding_System *codesys;
2360 /* Stream that we read the encoded data from or
2361 write the decoded data to. */
2364 /* If we are reading, then we can return only a fixed amount of
2365 data, so if the conversion resulted in too much data, we store it
2366 here for retrieval the next time around. */
2367 unsigned_char_dynarr *runoff;
2369 /* FLAGS holds flags indicating the current state of the encoding.
2370 Some of these flags are dependent on the coding system. */
2373 /* CH holds a partially built-up character. Since we only deal
2374 with one- and two-byte characters at the moment, we only use
2375 this to store the first byte of a two-byte character. */
2378 /* Additional information used by the ISO2022 encoder. */
2381 /* CHARSET holds the character sets currently assigned to the G0
2382 through G3 registers. It is initialized from the array
2383 INITIAL_CHARSET in CODESYS. */
2384 Lisp_Object charset[4];
2386 /* Which registers are currently invoked into the left (GL) and
2387 right (GR) halves of the 8-bit encoding space? */
2388 int register_left, register_right;
2390 /* Whether we need to explicitly designate the charset in the
2391 G? register before using it. It is initialized from the
2392 array FORCE_CHARSET_ON_OUTPUT in CODESYS. */
2393 unsigned char force_charset_on_output[4];
2395 /* Other state variables that need to be preserved across
2397 Lisp_Object current_charset;
2399 int current_char_boundary;
2402 void (*encode_char) (struct encoding_stream *str, Emchar c,
2403 unsigned_char_dynarr *dst, unsigned int *flags);
2404 void (*finish) (struct encoding_stream *str,
2405 unsigned_char_dynarr *dst, unsigned int *flags);
2407 /* Additional information (the state of the running CCL program)
2408 used by the CCL encoder. */
2409 struct ccl_program ccl;
2413 static int encoding_reader (Lstream *stream, unsigned char *data, size_t size);
2414 static int encoding_writer (Lstream *stream, CONST unsigned char *data,
2416 static int encoding_rewinder (Lstream *stream);
2417 static int encoding_seekable_p (Lstream *stream);
2418 static int encoding_flusher (Lstream *stream);
2419 static int encoding_closer (Lstream *stream);
2421 static Lisp_Object encoding_marker (Lisp_Object stream,
2422 void (*markobj) (Lisp_Object));
2424 DEFINE_LSTREAM_IMPLEMENTATION ("encoding", lstream_encoding,
2425 sizeof (struct encoding_stream));
2428 encoding_marker (Lisp_Object stream, void (*markobj) (Lisp_Object))
2430 Lstream *str = ENCODING_STREAM_DATA (XLSTREAM (stream))->other_end;
2431 Lisp_Object str_obj;
2433 /* We do not need to mark the coding systems or charsets stored
2434 within the stream because they are stored in a global list
2435 and automatically marked. */
2437 XSETLSTREAM (str_obj, str);
2439 if (str->imp->marker)
2440 return (str->imp->marker) (str_obj, markobj);
2445 /* Read SIZE bytes of data and store it into DATA. We are a encoding stream
2446 so we read data from the other end, encode it, and store it into DATA. */
2449 encoding_reader (Lstream *stream, unsigned char *data, size_t size)
2451 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2452 unsigned char *orig_data = data;
2454 int error_occurred = 0;
2456 /* We need to interface to mule_encode(), which expects to take some
2457 amount of data and store the result into a Dynarr. We have
2458 mule_encode() store into str->runoff, and take data from there
2461 /* We loop until we have enough data, reading chunks from the other
2462 end and encoding it. */
2465 /* Take data from the runoff if we can. Make sure to take at
2466 most SIZE bytes, and delete the data from the runoff. */
2467 if (Dynarr_length (str->runoff) > 0)
2469 int chunk = min ((int) size, Dynarr_length (str->runoff));
2470 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2471 Dynarr_delete_many (str->runoff, 0, chunk);
2477 break; /* No more room for data */
2479 if (str->flags & CODING_STATE_END)
2480 /* This means that on the previous iteration, we hit the EOF on
2481 the other end. We loop once more so that mule_encode() can
2482 output any final stuff it may be holding, or any "go back
2483 to a sane state" escape sequences. (This latter makes sense
2484 during encoding.) */
2487 /* Exhausted the runoff, so get some more. DATA at least SIZE bytes
2488 left of storage in it, so it's OK to read directly into it.
2489 (We'll be overwriting above, after we've encoded it into the
2491 read_size = Lstream_read (str->other_end, data, size);
2498 /* There might be some more end data produced in the translation.
2499 See the comment above. */
2500 str->flags |= CODING_STATE_END;
2501 mule_encode (stream, data, str->runoff, read_size);
2504 if (data == orig_data)
2505 return error_occurred ? -1 : 0;
2507 return data - orig_data;
2511 encoding_writer (Lstream *stream, CONST unsigned char *data, size_t size)
2513 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2516 /* Encode all our data into the runoff, and then attempt to write
2517 it all out to the other end. Remove whatever chunk we succeeded
2519 mule_encode (stream, data, str->runoff, size);
2520 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
2521 Dynarr_length (str->runoff));
2523 Dynarr_delete_many (str->runoff, 0, retval);
2524 /* Do NOT return retval. The return value indicates how much
2525 of the incoming data was written, not how many bytes were
2531 reset_encoding_stream (struct encoding_stream *str)
2534 switch (CODING_SYSTEM_TYPE (str->codesys))
2536 case CODESYS_ISO2022:
2540 str->encode_char = &char_encode_iso2022;
2541 str->finish = &char_finish_iso2022;
2542 for (i = 0; i < 4; i++)
2544 str->iso2022.charset[i] =
2545 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, i);
2546 str->iso2022.force_charset_on_output[i] =
2547 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (str->codesys, i);
2549 str->iso2022.register_left = 0;
2550 str->iso2022.register_right = 1;
2551 str->iso2022.current_charset = Qnil;
2552 str->iso2022.current_half = 0;
2556 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_ENCODE (str->codesys));
2559 str->encode_char = &char_encode_utf8;
2560 str->finish = &char_finish_utf8;
2562 str->encode_char = &char_encode_ucs4;
2563 str->finish = &char_finish_ucs4;
2564 case CODESYS_SHIFT_JIS:
2565 str->encode_char = &char_encode_shift_jis;
2566 str->finish = &char_finish_shift_jis;
2571 str->iso2022.current_char_boundary = 0;
2572 str->flags = str->ch = 0;
2576 encoding_rewinder (Lstream *stream)
2578 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2579 reset_encoding_stream (str);
2580 Dynarr_reset (str->runoff);
2581 return Lstream_rewind (str->other_end);
2585 encoding_seekable_p (Lstream *stream)
2587 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2588 return Lstream_seekable_p (str->other_end);
2592 encoding_flusher (Lstream *stream)
2594 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2595 return Lstream_flush (str->other_end);
2599 encoding_closer (Lstream *stream)
2601 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2602 if (stream->flags & LSTREAM_FL_WRITE)
2604 str->flags |= CODING_STATE_END;
2605 encoding_writer (stream, 0, 0);
2607 Dynarr_free (str->runoff);
2608 return Lstream_close (str->other_end);
2612 encoding_stream_coding_system (Lstream *stream)
2614 Lisp_Object coding_system;
2615 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2617 XSETCODING_SYSTEM (coding_system, str->codesys);
2618 return coding_system;
2622 set_encoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
2624 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
2625 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
2627 reset_encoding_stream (str);
2631 make_encoding_stream_1 (Lstream *stream, Lisp_Object codesys,
2634 Lstream *lstr = Lstream_new (lstream_encoding, mode);
2635 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
2639 str->runoff = Dynarr_new (unsigned_char);
2640 str->other_end = stream;
2641 set_encoding_stream_coding_system (lstr, codesys);
2642 XSETLSTREAM (obj, lstr);
2647 make_encoding_input_stream (Lstream *stream, Lisp_Object codesys)
2649 return make_encoding_stream_1 (stream, codesys, "r");
2653 make_encoding_output_stream (Lstream *stream, Lisp_Object codesys)
2655 return make_encoding_stream_1 (stream, codesys, "w");
2658 /* Convert N bytes of internally-formatted data stored in SRC to an
2659 external format, according to the encoding stream ENCODING.
2660 Store the encoded data into DST. */
2663 mule_encode (Lstream *encoding, CONST unsigned char *src,
2664 unsigned_char_dynarr *dst, unsigned int n)
2666 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
2668 switch (CODING_SYSTEM_TYPE (str->codesys))
2671 case CODESYS_INTERNAL:
2672 Dynarr_add_many (dst, src, n);
2675 case CODESYS_AUTODETECT:
2676 /* If we got this far and still haven't decided on the coding
2677 system, then do no conversion. */
2678 case CODESYS_NO_CONVERSION:
2679 encode_coding_no_conversion (encoding, src, dst, n);
2683 encode_coding_big5 (encoding, src, dst, n);
2686 str->ccl.last_block = str->flags & CODING_STATE_END;
2687 ccl_driver (&str->ccl, src, dst, n, 0, CCL_MODE_ENCODING);
2691 text_encode_generic (encoding, src, dst, n);
2695 DEFUN ("encode-coding-region", Fencode_coding_region, 3, 4, 0, /*
2696 Encode the text between START and END using CODING-SYSTEM.
2697 This will, for example, convert Japanese characters into stuff such as
2698 "^[$B!<!+^[(B" if you use the JIS encoding. Return length of encoded
2699 text. BUFFER defaults to the current buffer if unspecified.
2701 (start, end, coding_system, buffer))
2704 struct buffer *buf = decode_buffer (buffer, 0);
2705 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2706 Lstream *istr, *ostr;
2707 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2709 get_buffer_range_char (buf, start, end, &b, &e, 0);
2711 barf_if_buffer_read_only (buf, b, e);
2713 coding_system = Fget_coding_system (coding_system);
2714 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2715 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
2716 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
2717 Fget_coding_system (Qbinary));
2718 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
2720 istr = XLSTREAM (instream);
2721 ostr = XLSTREAM (outstream);
2722 GCPRO4 (instream, outstream, de_outstream, lb_outstream);
2723 /* The chain of streams looks like this:
2725 [BUFFER] <----- send through
2726 ------> [ENCODE AS SPECIFIED]
2727 ------> [DECODE AS BINARY]
2732 char tempbuf[1024]; /* some random amount */
2733 Bufpos newpos, even_newer_pos;
2734 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
2735 int size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
2739 newpos = lisp_buffer_stream_startpos (istr);
2740 Lstream_write (ostr, tempbuf, size_in_bytes);
2741 even_newer_pos = lisp_buffer_stream_startpos (istr);
2742 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
2748 lisp_buffer_stream_startpos (XLSTREAM (instream)) - b;
2749 Lstream_close (istr);
2750 Lstream_close (ostr);
2752 Lstream_delete (istr);
2753 Lstream_delete (ostr);
2754 Lstream_delete (XLSTREAM (de_outstream));
2755 Lstream_delete (XLSTREAM (lb_outstream));
2756 return make_int (retlen);
2763 text_encode_generic (Lstream *encoding, CONST unsigned char *src,
2764 unsigned_char_dynarr *dst, unsigned int n)
2767 unsigned char char_boundary;
2768 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
2769 unsigned int flags = str->flags;
2770 Emchar ch = str->ch;
2772 char_boundary = str->iso2022.current_char_boundary;
2778 if (char_boundary == 0)
2806 (*str->encode_char) (str, c, dst, &flags);
2808 else if (char_boundary == 1)
2810 (*str->encode_char) (str, (ch << 6) | (c & 0x3f), dst, &flags);
2816 ch = (ch << 6) | (c & 0x3f);
2821 if ((char_boundary == 0) && (flags & CODING_STATE_END))
2823 (*str->finish) (str, dst, &flags);
2828 str->iso2022.current_char_boundary = char_boundary;
2832 /************************************************************************/
2833 /* Shift-JIS methods */
2834 /************************************************************************/
2836 /* Shift-JIS is a coding system encoding three character sets: ASCII, right
2837 half of JISX0201-Kana, and JISX0208. An ASCII character is encoded
2838 as is. A character of JISX0201-Kana (DIMENSION1_CHARS94 character set) is
2839 encoded by "position-code + 0x80". A character of JISX0208
2840 (DIMENSION2_CHARS94 character set) is encoded in 2-byte but two
2841 position-codes are divided and shifted so that it fit in the range
2844 --- CODE RANGE of Shift-JIS ---
2845 (character set) (range)
2847 JISX0201-Kana 0xA0 .. 0xDF
2848 JISX0208 (1st byte) 0x80 .. 0x9F and 0xE0 .. 0xEF
2849 (2nd byte) 0x40 .. 0x7E and 0x80 .. 0xFC
2850 -------------------------------
2854 /* Is this the first byte of a Shift-JIS two-byte char? */
2856 #define BYTE_SJIS_TWO_BYTE_1_P(c) \
2857 (((c) >= 0x81 && (c) <= 0x9F) || ((c) >= 0xE0 && (c) <= 0xEF))
2859 /* Is this the second byte of a Shift-JIS two-byte char? */
2861 #define BYTE_SJIS_TWO_BYTE_2_P(c) \
2862 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0x80 && (c) <= 0xFC))
2864 #define BYTE_SJIS_KATAKANA_P(c) \
2865 ((c) >= 0xA1 && (c) <= 0xDF)
2868 detect_coding_sjis (struct detection_state *st, CONST unsigned char *src,
2876 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
2878 if (st->shift_jis.in_second_byte)
2880 st->shift_jis.in_second_byte = 0;
2884 else if ((c >= 0x80 && c < 0xA0) || c >= 0xE0)
2885 st->shift_jis.in_second_byte = 1;
2887 return CODING_CATEGORY_SHIFT_JIS_MASK;
2890 /* Convert Shift-JIS data to internal format. */
2893 decode_coding_sjis (Lstream *decoding, CONST unsigned char *src,
2894 unsigned_char_dynarr *dst, unsigned int n)
2897 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
2898 unsigned int flags = str->flags;
2899 unsigned int ch = str->ch;
2900 eol_type_t eol_type = str->eol_type;
2908 /* Previous character was first byte of Shift-JIS Kanji char. */
2909 if (BYTE_SJIS_TWO_BYTE_2_P (c))
2911 unsigned char e1, e2;
2913 DECODE_SJIS (ch, c, e1, e2);
2915 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_japanese_jisx0208,
2919 Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208);
2920 Dynarr_add (dst, e1);
2921 Dynarr_add (dst, e2);
2926 DECODE_ADD_BINARY_CHAR (ch, dst);
2927 DECODE_ADD_BINARY_CHAR (c, dst);
2933 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
2934 if (BYTE_SJIS_TWO_BYTE_1_P (c))
2936 else if (BYTE_SJIS_KATAKANA_P (c))
2939 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_katakana_jisx0201,
2942 Dynarr_add (dst, LEADING_BYTE_KATAKANA_JISX0201);
2943 Dynarr_add (dst, c);
2948 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_latin_jisx0201,
2952 DECODE_ADD_BINARY_CHAR (c, dst);
2954 label_continue_loop:;
2957 DECODE_HANDLE_END_OF_CONVERSION (flags, ch, dst);
2963 /* Convert internal character representation to Shift_JIS. */
2966 char_encode_shift_jis (struct encoding_stream *str, Emchar ch,
2967 unsigned_char_dynarr *dst, unsigned int *flags)
2969 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
2973 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
2974 Dynarr_add (dst, '\r');
2975 if (eol_type != EOL_CR)
2976 Dynarr_add (dst, ch);
2980 Lisp_Object charset;
2981 unsigned int c1, c2, s1, s2;
2985 get_byte_from_character_table
2986 (ch, XCHARSET_TO_BYTE1_TABLE (Vcharset_latin_jisx0201))) )
2988 charset = Vcharset_latin_jisx0201;
2993 BREAKUP_CHAR (ch, charset, c1, c2);
2995 if (EQ(charset, Vcharset_katakana_jisx0201))
2997 Dynarr_add (dst, c1 | 0x80);
3001 Dynarr_add (dst, c1);
3003 else if (EQ(charset, Vcharset_japanese_jisx0208))
3005 ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
3006 Dynarr_add (dst, s1);
3007 Dynarr_add (dst, s2);
3010 Dynarr_add (dst, '?');
3015 char_finish_shift_jis (struct encoding_stream *str, unsigned_char_dynarr *dst,
3016 unsigned int *flags)
3020 DEFUN ("decode-shift-jis-char", Fdecode_shift_jis_char, 1, 1, 0, /*
3021 Decode a JISX0208 character of Shift-JIS coding-system.
3022 CODE is the character code in Shift-JIS as a cons of type bytes.
3023 Return the corresponding character.
3027 unsigned char c1, c2, s1, s2;
3030 CHECK_INT (XCAR (code));
3031 CHECK_INT (XCDR (code));
3032 s1 = XINT (XCAR (code));
3033 s2 = XINT (XCDR (code));
3034 if (BYTE_SJIS_TWO_BYTE_1_P (s1) &&
3035 BYTE_SJIS_TWO_BYTE_2_P (s2))
3037 DECODE_SJIS (s1, s2, c1, c2);
3038 return make_char (MAKE_CHAR (Vcharset_japanese_jisx0208,
3039 c1 & 0x7F, c2 & 0x7F));
3045 DEFUN ("encode-shift-jis-char", Fencode_shift_jis_char, 1, 1, 0, /*
3046 Encode a JISX0208 character CHAR to SHIFT-JIS coding-system.
3047 Return the corresponding character code in SHIFT-JIS as a cons of two bytes.
3051 Lisp_Object charset;
3054 CHECK_CHAR_COERCE_INT (ch);
3055 BREAKUP_CHAR (XCHAR (ch), charset, c1, c2);
3056 if (EQ (charset, Vcharset_japanese_jisx0208))
3058 ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
3059 return Fcons (make_int (s1), make_int (s2));
3066 /************************************************************************/
3068 /************************************************************************/
3070 /* BIG5 is a coding system encoding two character sets: ASCII and
3071 Big5. An ASCII character is encoded as is. Big5 is a two-byte
3072 character set and is encoded in two-byte.
3074 --- CODE RANGE of BIG5 ---
3075 (character set) (range)
3077 Big5 (1st byte) 0xA1 .. 0xFE
3078 (2nd byte) 0x40 .. 0x7E and 0xA1 .. 0xFE
3079 --------------------------
3081 Since the number of characters in Big5 is larger than maximum
3082 characters in Emacs' charset (96x96), it can't be handled as one
3083 charset. So, in Emacs, Big5 is divided into two: `charset-big5-1'
3084 and `charset-big5-2'. Both <type>s are DIMENSION2_CHARS94. The former
3085 contains frequently used characters and the latter contains less
3086 frequently used characters. */
3088 #define BYTE_BIG5_TWO_BYTE_1_P(c) \
3089 ((c) >= 0xA1 && (c) <= 0xFE)
3091 /* Is this the second byte of a Shift-JIS two-byte char? */
3093 #define BYTE_BIG5_TWO_BYTE_2_P(c) \
3094 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0xA1 && (c) <= 0xFE))
3096 /* Number of Big5 characters which have the same code in 1st byte. */
3098 #define BIG5_SAME_ROW (0xFF - 0xA1 + 0x7F - 0x40)
3100 /* Code conversion macros. These are macros because they are used in
3101 inner loops during code conversion.
3103 Note that temporary variables in macros introduce the classic
3104 dynamic-scoping problems with variable names. We use capital-
3105 lettered variables in the assumption that XEmacs does not use
3106 capital letters in variables except in a very formalized way
3109 /* Convert Big5 code (b1, b2) into its internal string representation
3112 /* There is a much simpler way to split the Big5 charset into two.
3113 For the moment I'm going to leave the algorithm as-is because it
3114 claims to separate out the most-used characters into a single
3115 charset, which perhaps will lead to optimizations in various
3118 The way the algorithm works is something like this:
3120 Big5 can be viewed as a 94x157 charset, where the row is
3121 encoded into the bytes 0xA1 .. 0xFE and the column is encoded
3122 into the bytes 0x40 .. 0x7E and 0xA1 .. 0xFE. As for frequency,
3123 the split between low and high column numbers is apparently
3124 meaningless; ascending rows produce less and less frequent chars.
3125 Therefore, we assign the lower half of rows (0xA1 .. 0xC8) to
3126 the first charset, and the upper half (0xC9 .. 0xFE) to the
3127 second. To do the conversion, we convert the character into
3128 a single number where 0 .. 156 is the first row, 157 .. 313
3129 is the second, etc. That way, the characters are ordered by
3130 decreasing frequency. Then we just chop the space in two
3131 and coerce the result into a 94x94 space.
3134 #define DECODE_BIG5(b1, b2, lb, c1, c2) do \
3136 int B1 = b1, B2 = b2; \
3138 = (B1 - 0xA1) * BIG5_SAME_ROW + B2 - (B2 < 0x7F ? 0x40 : 0x62); \
3142 lb = LEADING_BYTE_CHINESE_BIG5_1; \
3146 lb = LEADING_BYTE_CHINESE_BIG5_2; \
3147 I -= (BIG5_SAME_ROW) * (0xC9 - 0xA1); \
3149 c1 = I / (0xFF - 0xA1) + 0xA1; \
3150 c2 = I % (0xFF - 0xA1) + 0xA1; \
3153 /* Convert the internal string representation of a Big5 character
3154 (lb, c1, c2) into Big5 code (b1, b2). */
3156 #define ENCODE_BIG5(lb, c1, c2, b1, b2) do \
3158 unsigned int I = ((c1) - 0xA1) * (0xFF - 0xA1) + ((c2) - 0xA1); \
3160 if (lb == LEADING_BYTE_CHINESE_BIG5_2) \
3162 I += BIG5_SAME_ROW * (0xC9 - 0xA1); \
3164 b1 = I / BIG5_SAME_ROW + 0xA1; \
3165 b2 = I % BIG5_SAME_ROW; \
3166 b2 += b2 < 0x3F ? 0x40 : 0x62; \
3170 detect_coding_big5 (struct detection_state *st, CONST unsigned char *src,
3178 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO ||
3179 (c >= 0x80 && c <= 0xA0))
3181 if (st->big5.in_second_byte)
3183 st->big5.in_second_byte = 0;
3184 if (c < 0x40 || (c >= 0x80 && c <= 0xA0))
3188 st->big5.in_second_byte = 1;
3190 return CODING_CATEGORY_BIG5_MASK;
3193 /* Convert Big5 data to internal format. */
3196 decode_coding_big5 (Lstream *decoding, CONST unsigned char *src,
3197 unsigned_char_dynarr *dst, unsigned int n)
3200 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3201 unsigned int flags = str->flags;
3202 unsigned int ch = str->ch;
3203 eol_type_t eol_type = str->eol_type;
3210 /* Previous character was first byte of Big5 char. */
3211 if (BYTE_BIG5_TWO_BYTE_2_P (c))
3213 unsigned char b1, b2, b3;
3214 DECODE_BIG5 (ch, c, b1, b2, b3);
3215 Dynarr_add (dst, b1);
3216 Dynarr_add (dst, b2);
3217 Dynarr_add (dst, b3);
3221 DECODE_ADD_BINARY_CHAR (ch, dst);
3222 DECODE_ADD_BINARY_CHAR (c, dst);
3228 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3229 if (BYTE_BIG5_TWO_BYTE_1_P (c))
3232 DECODE_ADD_BINARY_CHAR (c, dst);
3234 label_continue_loop:;
3237 DECODE_HANDLE_END_OF_CONVERSION (flags, ch, dst);
3243 /* Convert internally-formatted data to Big5. */
3246 encode_coding_big5 (Lstream *encoding, CONST unsigned char *src,
3247 unsigned_char_dynarr *dst, unsigned int n)
3251 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3252 unsigned int flags = str->flags;
3253 unsigned int ch = str->ch;
3254 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3261 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3262 Dynarr_add (dst, '\r');
3263 if (eol_type != EOL_CR)
3264 Dynarr_add (dst, '\n');
3266 else if (BYTE_ASCII_P (c))
3269 Dynarr_add (dst, c);
3271 else if (BUFBYTE_LEADING_BYTE_P (c))
3273 if (c == LEADING_BYTE_CHINESE_BIG5_1 ||
3274 c == LEADING_BYTE_CHINESE_BIG5_2)
3276 /* A recognized leading byte. */
3278 continue; /* not done with this character. */
3280 /* otherwise just ignore this character. */
3282 else if (ch == LEADING_BYTE_CHINESE_BIG5_1 ||
3283 ch == LEADING_BYTE_CHINESE_BIG5_2)
3285 /* Previous char was a recognized leading byte. */
3287 continue; /* not done with this character. */
3291 /* Encountering second byte of a Big5 character. */
3292 unsigned char b1, b2;
3294 ENCODE_BIG5 (ch >> 8, ch & 0xFF, c, b1, b2);
3295 Dynarr_add (dst, b1);
3296 Dynarr_add (dst, b2);
3308 DEFUN ("decode-big5-char", Fdecode_big5_char, 1, 1, 0, /*
3309 Decode a Big5 character CODE of BIG5 coding-system.
3310 CODE is the character code in BIG5, a cons of two integers.
3311 Return the corresponding character.
3315 unsigned char c1, c2, b1, b2;
3318 CHECK_INT (XCAR (code));
3319 CHECK_INT (XCDR (code));
3320 b1 = XINT (XCAR (code));
3321 b2 = XINT (XCDR (code));
3322 if (BYTE_BIG5_TWO_BYTE_1_P (b1) &&
3323 BYTE_BIG5_TWO_BYTE_2_P (b2))
3325 Charset_ID leading_byte;
3326 Lisp_Object charset;
3327 DECODE_BIG5 (b1, b2, leading_byte, c1, c2);
3328 charset = CHARSET_BY_LEADING_BYTE (leading_byte);
3329 return make_char (MAKE_CHAR (charset, c1 & 0x7F, c2 & 0x7F));
3335 DEFUN ("encode-big5-char", Fencode_big5_char, 1, 1, 0, /*
3336 Encode the Big5 character CH to BIG5 coding-system.
3337 Return the corresponding character code in Big5.
3341 Lisp_Object charset;
3344 CHECK_CHAR_COERCE_INT (ch);
3345 BREAKUP_CHAR (XCHAR (ch), charset, c1, c2);
3346 if (EQ (charset, Vcharset_chinese_big5_1) ||
3347 EQ (charset, Vcharset_chinese_big5_2))
3349 ENCODE_BIG5 (XCHARSET_LEADING_BYTE (charset), c1 | 0x80, c2 | 0x80,
3351 return Fcons (make_int (b1), make_int (b2));
3358 /************************************************************************/
3360 /************************************************************************/
3363 detect_coding_ucs4 (struct detection_state *st, CONST unsigned char *src,
3369 switch (st->ucs4.in_byte)
3378 st->ucs4.in_byte = 0;
3384 return CODING_CATEGORY_UCS4_MASK;
3388 decode_coding_ucs4 (Lstream *decoding, CONST unsigned char *src,
3389 unsigned_char_dynarr *dst, unsigned int n)
3391 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3392 unsigned int flags = str->flags;
3393 unsigned int ch = str->ch;
3394 unsigned char counter = str->counter;
3398 unsigned char c = *src++;
3406 DECODE_ADD_UCS_CHAR ((ch << 8) | c, dst);
3411 ch = ( ch << 8 ) | c;
3415 if (counter & CODING_STATE_END)
3416 DECODE_OUTPUT_PARTIAL_CHAR (ch);
3420 str->counter = counter;
3424 char_encode_ucs4 (struct encoding_stream *str, Emchar ch,
3425 unsigned_char_dynarr *dst, unsigned int *flags)
3427 Dynarr_add (dst, ch >> 24);
3428 Dynarr_add (dst, (ch >> 16) & 255);
3429 Dynarr_add (dst, (ch >> 8) & 255);
3430 Dynarr_add (dst, ch & 255);
3434 char_finish_ucs4 (struct encoding_stream *str, unsigned_char_dynarr *dst,
3435 unsigned int *flags)
3440 /************************************************************************/
3442 /************************************************************************/
3445 detect_coding_utf8 (struct detection_state *st, CONST unsigned char *src,
3450 unsigned char c = *src++;
3451 switch (st->utf8.in_byte)
3454 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
3457 st->utf8.in_byte = 5;
3459 st->utf8.in_byte = 4;
3461 st->utf8.in_byte = 3;
3463 st->utf8.in_byte = 2;
3465 st->utf8.in_byte = 1;
3470 if ((c & 0xc0) != 0x80)
3476 return CODING_CATEGORY_UTF8_MASK;
3480 decode_coding_utf8 (Lstream *decoding, CONST unsigned char *src,
3481 unsigned_char_dynarr *dst, unsigned int n)
3483 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3484 unsigned int flags = str->flags;
3485 unsigned int ch = str->ch;
3486 eol_type_t eol_type = str->eol_type;
3487 unsigned char counter = str->counter;
3491 unsigned char c = *src++;
3500 else if ( c >= 0xf8 )
3505 else if ( c >= 0xf0 )
3510 else if ( c >= 0xe0 )
3515 else if ( c >= 0xc0 )
3522 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3523 DECODE_ADD_UCS_CHAR (c, dst);
3527 ch = ( ch << 6 ) | ( c & 0x3f );
3528 DECODE_ADD_UCS_CHAR (ch, dst);
3533 ch = ( ch << 6 ) | ( c & 0x3f );
3536 label_continue_loop:;
3539 if (flags & CODING_STATE_END)
3540 DECODE_OUTPUT_PARTIAL_CHAR (ch);
3544 str->counter = counter;
3548 char_encode_utf8 (struct encoding_stream *str, Emchar ch,
3549 unsigned_char_dynarr *dst, unsigned int *flags)
3551 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3555 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3556 Dynarr_add (dst, '\r');
3557 if (eol_type != EOL_CR)
3558 Dynarr_add (dst, ch);
3560 else if (ch <= 0x7f)
3562 Dynarr_add (dst, ch);
3564 else if (ch <= 0x7ff)
3566 Dynarr_add (dst, (ch >> 6) | 0xc0);
3567 Dynarr_add (dst, (ch & 0x3f) | 0x80);
3569 else if (ch <= 0xffff)
3571 Dynarr_add (dst, (ch >> 12) | 0xe0);
3572 Dynarr_add (dst, ((ch >> 6) & 0x3f) | 0x80);
3573 Dynarr_add (dst, (ch & 0x3f) | 0x80);
3575 else if (ch <= 0x1fffff)
3577 Dynarr_add (dst, (ch >> 18) | 0xf0);
3578 Dynarr_add (dst, ((ch >> 12) & 0x3f) | 0x80);
3579 Dynarr_add (dst, ((ch >> 6) & 0x3f) | 0x80);
3580 Dynarr_add (dst, (ch & 0x3f) | 0x80);
3582 else if (ch <= 0x3ffffff)
3584 Dynarr_add (dst, (ch >> 24) | 0xf8);
3585 Dynarr_add (dst, ((ch >> 18) & 0x3f) | 0x80);
3586 Dynarr_add (dst, ((ch >> 12) & 0x3f) | 0x80);
3587 Dynarr_add (dst, ((ch >> 6) & 0x3f) | 0x80);
3588 Dynarr_add (dst, (ch & 0x3f) | 0x80);
3592 Dynarr_add (dst, (ch >> 30) | 0xfc);
3593 Dynarr_add (dst, ((ch >> 24) & 0x3f) | 0x80);
3594 Dynarr_add (dst, ((ch >> 18) & 0x3f) | 0x80);
3595 Dynarr_add (dst, ((ch >> 12) & 0x3f) | 0x80);
3596 Dynarr_add (dst, ((ch >> 6) & 0x3f) | 0x80);
3597 Dynarr_add (dst, (ch & 0x3f) | 0x80);
3602 char_finish_utf8 (struct encoding_stream *str, unsigned_char_dynarr *dst,
3603 unsigned int *flags)
3608 /************************************************************************/
3609 /* ISO2022 methods */
3610 /************************************************************************/
3612 /* The following note describes the coding system ISO2022 briefly.
3613 Since the intention of this note is to help understand the
3614 functions in this file, some parts are NOT ACCURATE or OVERLY
3615 SIMPLIFIED. For thorough understanding, please refer to the
3616 original document of ISO2022.
3618 ISO2022 provides many mechanisms to encode several character sets
3619 in 7-bit and 8-bit environments. For 7-bit environments, all text
3620 is encoded using bytes less than 128. This may make the encoded
3621 text a little bit longer, but the text passes more easily through
3622 several gateways, some of which strip off MSB (Most Signigant Bit).
3624 There are two kinds of character sets: control character set and
3625 graphic character set. The former contains control characters such
3626 as `newline' and `escape' to provide control functions (control
3627 functions are also provided by escape sequences). The latter
3628 contains graphic characters such as 'A' and '-'. Emacs recognizes
3629 two control character sets and many graphic character sets.
3631 Graphic character sets are classified into one of the following
3632 four classes, according to the number of bytes (DIMENSION) and
3633 number of characters in one dimension (CHARS) of the set:
3634 - DIMENSION1_CHARS94
3635 - DIMENSION1_CHARS96
3636 - DIMENSION2_CHARS94
3637 - DIMENSION2_CHARS96
3639 In addition, each character set is assigned an identification tag,
3640 unique for each set, called "final character" (denoted as <F>
3641 hereafter). The <F> of each character set is decided by ECMA(*)
3642 when it is registered in ISO. The code range of <F> is 0x30..0x7F
3643 (0x30..0x3F are for private use only).
3645 Note (*): ECMA = European Computer Manufacturers Association
3647 Here are examples of graphic character set [NAME(<F>)]:
3648 o DIMENSION1_CHARS94 -- ASCII('B'), right-half-of-JISX0201('I'), ...
3649 o DIMENSION1_CHARS96 -- right-half-of-ISO8859-1('A'), ...
3650 o DIMENSION2_CHARS94 -- GB2312('A'), JISX0208('B'), ...
3651 o DIMENSION2_CHARS96 -- none for the moment
3653 A code area (1 byte = 8 bits) is divided into 4 areas, C0, GL, C1, and GR.
3654 C0 [0x00..0x1F] -- control character plane 0
3655 GL [0x20..0x7F] -- graphic character plane 0
3656 C1 [0x80..0x9F] -- control character plane 1
3657 GR [0xA0..0xFF] -- graphic character plane 1
3659 A control character set is directly designated and invoked to C0 or
3660 C1 by an escape sequence. The most common case is that:
3661 - ISO646's control character set is designated/invoked to C0, and
3662 - ISO6429's control character set is designated/invoked to C1,
3663 and usually these designations/invocations are omitted in encoded
3664 text. In a 7-bit environment, only C0 can be used, and a control
3665 character for C1 is encoded by an appropriate escape sequence to
3666 fit into the environment. All control characters for C1 are
3667 defined to have corresponding escape sequences.
3669 A graphic character set is at first designated to one of four
3670 graphic registers (G0 through G3), then these graphic registers are
3671 invoked to GL or GR. These designations and invocations can be
3672 done independently. The most common case is that G0 is invoked to
3673 GL, G1 is invoked to GR, and ASCII is designated to G0. Usually
3674 these invocations and designations are omitted in encoded text.
3675 In a 7-bit environment, only GL can be used.
3677 When a graphic character set of CHARS94 is invoked to GL, codes
3678 0x20 and 0x7F of the GL area work as control characters SPACE and
3679 DEL respectively, and codes 0xA0 and 0xFF of the GR area should not
3682 There are two ways of invocation: locking-shift and single-shift.
3683 With locking-shift, the invocation lasts until the next different
3684 invocation, whereas with single-shift, the invocation affects the
3685 following character only and doesn't affect the locking-shift
3686 state. Invocations are done by the following control characters or
3689 ----------------------------------------------------------------------
3690 abbrev function cntrl escape seq description
3691 ----------------------------------------------------------------------
3692 SI/LS0 (shift-in) 0x0F none invoke G0 into GL
3693 SO/LS1 (shift-out) 0x0E none invoke G1 into GL
3694 LS2 (locking-shift-2) none ESC 'n' invoke G2 into GL
3695 LS3 (locking-shift-3) none ESC 'o' invoke G3 into GL
3696 LS1R (locking-shift-1 right) none ESC '~' invoke G1 into GR (*)
3697 LS2R (locking-shift-2 right) none ESC '}' invoke G2 into GR (*)
3698 LS3R (locking-shift 3 right) none ESC '|' invoke G3 into GR (*)
3699 SS2 (single-shift-2) 0x8E ESC 'N' invoke G2 for one char
3700 SS3 (single-shift-3) 0x8F ESC 'O' invoke G3 for one char
3701 ----------------------------------------------------------------------
3702 (*) These are not used by any known coding system.
3704 Control characters for these functions are defined by macros
3705 ISO_CODE_XXX in `coding.h'.
3707 Designations are done by the following escape sequences:
3708 ----------------------------------------------------------------------
3709 escape sequence description
3710 ----------------------------------------------------------------------
3711 ESC '(' <F> designate DIMENSION1_CHARS94<F> to G0
3712 ESC ')' <F> designate DIMENSION1_CHARS94<F> to G1
3713 ESC '*' <F> designate DIMENSION1_CHARS94<F> to G2
3714 ESC '+' <F> designate DIMENSION1_CHARS94<F> to G3
3715 ESC ',' <F> designate DIMENSION1_CHARS96<F> to G0 (*)
3716 ESC '-' <F> designate DIMENSION1_CHARS96<F> to G1
3717 ESC '.' <F> designate DIMENSION1_CHARS96<F> to G2
3718 ESC '/' <F> designate DIMENSION1_CHARS96<F> to G3
3719 ESC '$' '(' <F> designate DIMENSION2_CHARS94<F> to G0 (**)
3720 ESC '$' ')' <F> designate DIMENSION2_CHARS94<F> to G1
3721 ESC '$' '*' <F> designate DIMENSION2_CHARS94<F> to G2
3722 ESC '$' '+' <F> designate DIMENSION2_CHARS94<F> to G3
3723 ESC '$' ',' <F> designate DIMENSION2_CHARS96<F> to G0 (*)
3724 ESC '$' '-' <F> designate DIMENSION2_CHARS96<F> to G1
3725 ESC '$' '.' <F> designate DIMENSION2_CHARS96<F> to G2
3726 ESC '$' '/' <F> designate DIMENSION2_CHARS96<F> to G3
3727 ----------------------------------------------------------------------
3729 In this list, "DIMENSION1_CHARS94<F>" means a graphic character set
3730 of dimension 1, chars 94, and final character <F>, etc...
3732 Note (*): Although these designations are not allowed in ISO2022,
3733 Emacs accepts them on decoding, and produces them on encoding
3734 CHARS96 character sets in a coding system which is characterized as
3735 7-bit environment, non-locking-shift, and non-single-shift.
3737 Note (**): If <F> is '@', 'A', or 'B', the intermediate character
3738 '(' can be omitted. We refer to this as "short-form" hereafter.
3740 Now you may notice that there are a lot of ways for encoding the
3741 same multilingual text in ISO2022. Actually, there exist many
3742 coding systems such as Compound Text (used in X11's inter client
3743 communication, ISO-2022-JP (used in Japanese internet), ISO-2022-KR
3744 (used in Korean internet), EUC (Extended UNIX Code, used in Asian
3745 localized platforms), and all of these are variants of ISO2022.
3747 In addition to the above, Emacs handles two more kinds of escape
3748 sequences: ISO6429's direction specification and Emacs' private
3749 sequence for specifying character composition.
3751 ISO6429's direction specification takes the following form:
3752 o CSI ']' -- end of the current direction
3753 o CSI '0' ']' -- end of the current direction
3754 o CSI '1' ']' -- start of left-to-right text
3755 o CSI '2' ']' -- start of right-to-left text
3756 The control character CSI (0x9B: control sequence introducer) is
3757 abbreviated to the escape sequence ESC '[' in a 7-bit environment.
3759 Character composition specification takes the following form:
3760 o ESC '0' -- start character composition
3761 o ESC '1' -- end character composition
3762 Since these are not standard escape sequences of any ISO standard,
3763 their use with these meanings is restricted to Emacs only. */
3766 reset_iso2022 (Lisp_Object coding_system, struct iso2022_decoder *iso)
3770 for (i = 0; i < 4; i++)
3772 if (!NILP (coding_system))
3774 XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, i);
3776 iso->charset[i] = Qt;
3777 iso->invalid_designated[i] = 0;
3779 iso->esc = ISO_ESC_NOTHING;
3780 iso->esc_bytes_index = 0;
3781 iso->register_left = 0;
3782 iso->register_right = 1;
3783 iso->switched_dir_and_no_valid_charset_yet = 0;
3784 iso->invalid_switch_dir = 0;
3785 iso->output_direction_sequence = 0;
3786 iso->output_literally = 0;
3787 #ifdef ENABLE_COMPOSITE_CHARS
3788 if (iso->composite_chars)
3789 Dynarr_reset (iso->composite_chars);
3794 fit_to_be_escape_quoted (unsigned char c)
3811 /* Parse one byte of an ISO2022 escape sequence.
3812 If the result is an invalid escape sequence, return 0 and
3813 do not change anything in STR. Otherwise, if the result is
3814 an incomplete escape sequence, update ISO2022.ESC and
3815 ISO2022.ESC_BYTES and return -1. Otherwise, update
3816 all the state variables (but not ISO2022.ESC_BYTES) and
3819 If CHECK_INVALID_CHARSETS is non-zero, check for designation
3820 or invocation of an invalid character set and treat that as
3821 an unrecognized escape sequence. */
3824 parse_iso2022_esc (Lisp_Object codesys, struct iso2022_decoder *iso,
3825 unsigned char c, unsigned int *flags,
3826 int check_invalid_charsets)
3828 /* (1) If we're at the end of a designation sequence, CS is the
3829 charset being designated and REG is the register to designate
3832 (2) If we're at the end of a locking-shift sequence, REG is
3833 the register to invoke and HALF (0 == left, 1 == right) is
3834 the half to invoke it into.
3836 (3) If we're at the end of a single-shift sequence, REG is
3837 the register to invoke. */
3838 Lisp_Object cs = Qnil;
3841 /* NOTE: This code does goto's all over the fucking place.
3842 The reason for this is that we're basically implementing
3843 a state machine here, and hierarchical languages like C
3844 don't really provide a clean way of doing this. */
3846 if (! (*flags & CODING_STATE_ESCAPE))
3847 /* At beginning of escape sequence; we need to reset our
3848 escape-state variables. */
3849 iso->esc = ISO_ESC_NOTHING;
3851 iso->output_literally = 0;
3852 iso->output_direction_sequence = 0;
3856 case ISO_ESC_NOTHING:
3857 iso->esc_bytes_index = 0;
3860 case ISO_CODE_ESC: /* Start escape sequence */
3861 *flags |= CODING_STATE_ESCAPE;
3865 case ISO_CODE_CSI: /* ISO6429 (specifying directionality) */
3866 *flags |= CODING_STATE_ESCAPE;
3867 iso->esc = ISO_ESC_5_11;
3870 case ISO_CODE_SO: /* locking shift 1 */
3873 case ISO_CODE_SI: /* locking shift 0 */
3877 case ISO_CODE_SS2: /* single shift */
3880 case ISO_CODE_SS3: /* single shift */
3884 default: /* Other control characters */
3891 /**** single shift ****/
3893 case 'N': /* single shift 2 */
3896 case 'O': /* single shift 3 */
3900 /**** locking shift ****/
3902 case '~': /* locking shift 1 right */
3905 case 'n': /* locking shift 2 */
3908 case '}': /* locking shift 2 right */
3911 case 'o': /* locking shift 3 */
3914 case '|': /* locking shift 3 right */
3918 #ifdef ENABLE_COMPOSITE_CHARS
3919 /**** composite ****/
3922 iso->esc = ISO_ESC_START_COMPOSITE;
3923 *flags = (*flags & CODING_STATE_ISO2022_LOCK) |
3924 CODING_STATE_COMPOSITE;
3928 iso->esc = ISO_ESC_END_COMPOSITE;
3929 *flags = (*flags & CODING_STATE_ISO2022_LOCK) &
3930 ~CODING_STATE_COMPOSITE;
3932 #endif /* ENABLE_COMPOSITE_CHARS */
3934 /**** directionality ****/
3937 iso->esc = ISO_ESC_5_11;
3940 /**** designation ****/
3942 case '$': /* multibyte charset prefix */
3943 iso->esc = ISO_ESC_2_4;
3947 if (0x28 <= c && c <= 0x2F)
3949 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_8);
3953 /* This function is called with CODESYS equal to nil when
3954 doing coding-system detection. */
3956 && XCODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
3957 && fit_to_be_escape_quoted (c))
3959 iso->esc = ISO_ESC_LITERAL;
3960 *flags &= CODING_STATE_ISO2022_LOCK;
3970 /**** directionality ****/
3972 case ISO_ESC_5_11: /* ISO6429 direction control */
3975 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
3976 goto directionality;
3978 if (c == '0') iso->esc = ISO_ESC_5_11_0;
3979 else if (c == '1') iso->esc = ISO_ESC_5_11_1;
3980 else if (c == '2') iso->esc = ISO_ESC_5_11_2;
3984 case ISO_ESC_5_11_0:
3987 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
3988 goto directionality;
3992 case ISO_ESC_5_11_1:
3995 *flags = (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
3996 goto directionality;
4000 case ISO_ESC_5_11_2:
4003 *flags = (*flags & CODING_STATE_ISO2022_LOCK) | CODING_STATE_R2L;
4004 goto directionality;
4009 iso->esc = ISO_ESC_DIRECTIONALITY;
4010 /* Various junk here to attempt to preserve the direction sequences
4011 literally in the text if they would otherwise be swallowed due
4012 to invalid designations that don't show up as actual charset
4013 changes in the text. */
4014 if (iso->invalid_switch_dir)
4016 /* We already inserted a direction switch literally into the
4017 text. We assume (#### this may not be right) that the
4018 next direction switch is the one going the other way,
4019 and we need to output that literally as well. */
4020 iso->output_literally = 1;
4021 iso->invalid_switch_dir = 0;
4027 /* If we are in the thrall of an invalid designation,
4028 then stick the directionality sequence literally into the
4029 output stream so it ends up in the original text again. */
4030 for (jj = 0; jj < 4; jj++)
4031 if (iso->invalid_designated[jj])
4035 iso->output_literally = 1;
4036 iso->invalid_switch_dir = 1;
4039 /* Indicate that we haven't yet seen a valid designation,
4040 so that if a switch-dir is directly followed by an
4041 invalid designation, both get inserted literally. */
4042 iso->switched_dir_and_no_valid_charset_yet = 1;
4047 /**** designation ****/
4050 if (0x28 <= c && c <= 0x2F)
4052 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_4_8);
4055 if (0x40 <= c && c <= 0x42)
4057 cs = CHARSET_BY_ATTRIBUTES (CHARSET_TYPE_94X94, c,
4058 *flags & CODING_STATE_R2L ?
4059 CHARSET_RIGHT_TO_LEFT :
4060 CHARSET_LEFT_TO_RIGHT);
4070 if (c < '0' || c > '~')
4071 return 0; /* bad final byte */
4073 if (iso->esc >= ISO_ESC_2_8 &&
4074 iso->esc <= ISO_ESC_2_15)
4076 type = ((iso->esc >= ISO_ESC_2_12) ?
4077 CHARSET_TYPE_96 : CHARSET_TYPE_94);
4078 reg = (iso->esc - ISO_ESC_2_8) & 3;
4080 else if (iso->esc >= ISO_ESC_2_4_8 &&
4081 iso->esc <= ISO_ESC_2_4_15)
4083 type = ((iso->esc >= ISO_ESC_2_4_12) ?
4084 CHARSET_TYPE_96X96 : CHARSET_TYPE_94X94);
4085 reg = (iso->esc - ISO_ESC_2_4_8) & 3;
4089 /* Can this ever be reached? -slb */
4093 cs = CHARSET_BY_ATTRIBUTES (type, c,
4094 *flags & CODING_STATE_R2L ?
4095 CHARSET_RIGHT_TO_LEFT :
4096 CHARSET_LEFT_TO_RIGHT);
4102 iso->esc_bytes[iso->esc_bytes_index++] = (unsigned char) c;
4106 if (check_invalid_charsets && !CHARSETP (iso->charset[reg]))
4107 /* can't invoke something that ain't there. */
4109 iso->esc = ISO_ESC_SINGLE_SHIFT;
4110 *flags &= CODING_STATE_ISO2022_LOCK;
4112 *flags |= CODING_STATE_SS2;
4114 *flags |= CODING_STATE_SS3;
4118 if (check_invalid_charsets &&
4119 !CHARSETP (iso->charset[reg]))
4120 /* can't invoke something that ain't there. */
4123 iso->register_right = reg;
4125 iso->register_left = reg;
4126 *flags &= CODING_STATE_ISO2022_LOCK;
4127 iso->esc = ISO_ESC_LOCKING_SHIFT;
4131 if (NILP (cs) && check_invalid_charsets)
4133 iso->invalid_designated[reg] = 1;
4134 iso->charset[reg] = Vcharset_ascii;
4135 iso->esc = ISO_ESC_DESIGNATE;
4136 *flags &= CODING_STATE_ISO2022_LOCK;
4137 iso->output_literally = 1;
4138 if (iso->switched_dir_and_no_valid_charset_yet)
4140 /* We encountered a switch-direction followed by an
4141 invalid designation. Ensure that the switch-direction
4142 gets outputted; otherwise it will probably get eaten
4143 when the text is written out again. */
4144 iso->switched_dir_and_no_valid_charset_yet = 0;
4145 iso->output_direction_sequence = 1;
4146 /* And make sure that the switch-dir going the other
4147 way gets outputted, as well. */
4148 iso->invalid_switch_dir = 1;
4152 /* This function is called with CODESYS equal to nil when
4153 doing coding-system detection. */
4154 if (!NILP (codesys))
4156 charset_conversion_spec_dynarr *dyn =
4157 XCODING_SYSTEM (codesys)->iso2022.input_conv;
4163 for (i = 0; i < Dynarr_length (dyn); i++)
4165 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
4166 if (EQ (cs, spec->from_charset))
4167 cs = spec->to_charset;
4172 iso->charset[reg] = cs;
4173 iso->esc = ISO_ESC_DESIGNATE;
4174 *flags &= CODING_STATE_ISO2022_LOCK;
4175 if (iso->invalid_designated[reg])
4177 iso->invalid_designated[reg] = 0;
4178 iso->output_literally = 1;
4180 if (iso->switched_dir_and_no_valid_charset_yet)
4181 iso->switched_dir_and_no_valid_charset_yet = 0;
4186 detect_coding_iso2022 (struct detection_state *st, CONST unsigned char *src,
4191 /* #### There are serious deficiencies in the recognition mechanism
4192 here. This needs to be much smarter if it's going to cut it.
4193 The sequence "\xff\x0f" is currently detected as LOCK_SHIFT while
4194 it should be detected as Latin-1.
4195 All the ISO2022 stuff in this file should be synced up with the
4196 code from FSF Emacs-20.4, in which Mule should be more or less stable.
4197 Perhaps we should wait till R2L works in FSF Emacs? */
4199 if (!st->iso2022.initted)
4201 reset_iso2022 (Qnil, &st->iso2022.iso);
4202 st->iso2022.mask = (CODING_CATEGORY_ISO_7_MASK |
4203 CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
4204 CODING_CATEGORY_ISO_8_1_MASK |
4205 CODING_CATEGORY_ISO_8_2_MASK |
4206 CODING_CATEGORY_ISO_LOCK_SHIFT_MASK);
4207 st->iso2022.flags = 0;
4208 st->iso2022.high_byte_count = 0;
4209 st->iso2022.saw_single_shift = 0;
4210 st->iso2022.initted = 1;
4213 mask = st->iso2022.mask;
4220 mask &= ~CODING_CATEGORY_ISO_7_MASK;
4221 st->iso2022.high_byte_count++;
4225 if (st->iso2022.high_byte_count && !st->iso2022.saw_single_shift)
4227 if (st->iso2022.high_byte_count & 1)
4228 /* odd number of high bytes; assume not iso-8-2 */
4229 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
4231 st->iso2022.high_byte_count = 0;
4232 st->iso2022.saw_single_shift = 0;
4234 mask &= ~CODING_CATEGORY_ISO_7_MASK;
4236 if (!(st->iso2022.flags & CODING_STATE_ESCAPE)
4237 && (BYTE_C0_P (c) || BYTE_C1_P (c)))
4238 { /* control chars */
4241 /* Allow and ignore control characters that you might
4242 reasonably see in a text file */
4247 case 8: /* backspace */
4248 case 11: /* vertical tab */
4249 case 12: /* form feed */
4250 case 26: /* MS-DOS C-z junk */
4251 case 31: /* '^_' -- for info */
4252 goto label_continue_loop;
4259 if ((st->iso2022.flags & CODING_STATE_ESCAPE) || BYTE_C0_P (c)
4262 if (parse_iso2022_esc (Qnil, &st->iso2022.iso, c,
4263 &st->iso2022.flags, 0))
4265 switch (st->iso2022.iso.esc)
4267 case ISO_ESC_DESIGNATE:
4268 mask &= ~CODING_CATEGORY_ISO_8_1_MASK;
4269 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
4271 case ISO_ESC_LOCKING_SHIFT:
4272 mask = CODING_CATEGORY_ISO_LOCK_SHIFT_MASK;
4273 goto ran_out_of_chars;
4274 case ISO_ESC_SINGLE_SHIFT:
4275 mask &= ~CODING_CATEGORY_ISO_8_DESIGNATE_MASK;
4276 st->iso2022.saw_single_shift = 1;
4285 goto ran_out_of_chars;
4288 label_continue_loop:;
4297 postprocess_iso2022_mask (int mask)
4299 /* #### kind of cheesy */
4300 /* If seven-bit ISO is allowed, then assume that the encoding is
4301 entirely seven-bit and turn off the eight-bit ones. */
4302 if (mask & CODING_CATEGORY_ISO_7_MASK)
4303 mask &= ~ (CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
4304 CODING_CATEGORY_ISO_8_1_MASK |
4305 CODING_CATEGORY_ISO_8_2_MASK);
4309 /* If FLAGS is a null pointer or specifies right-to-left motion,
4310 output a switch-dir-to-left-to-right sequence to DST.
4311 Also update FLAGS if it is not a null pointer.
4312 If INTERNAL_P is set, we are outputting in internal format and
4313 need to handle the CSI differently. */
4316 restore_left_to_right_direction (Lisp_Coding_System *codesys,
4317 unsigned_char_dynarr *dst,
4318 unsigned int *flags,
4321 if (!flags || (*flags & CODING_STATE_R2L))
4323 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
4325 Dynarr_add (dst, ISO_CODE_ESC);
4326 Dynarr_add (dst, '[');
4328 else if (internal_p)
4329 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
4331 Dynarr_add (dst, ISO_CODE_CSI);
4332 Dynarr_add (dst, '0');
4333 Dynarr_add (dst, ']');
4335 *flags &= ~CODING_STATE_R2L;
4339 /* If FLAGS is a null pointer or specifies a direction different from
4340 DIRECTION (which should be either CHARSET_RIGHT_TO_LEFT or
4341 CHARSET_LEFT_TO_RIGHT), output the appropriate switch-dir escape
4342 sequence to DST. Also update FLAGS if it is not a null pointer.
4343 If INTERNAL_P is set, we are outputting in internal format and
4344 need to handle the CSI differently. */
4347 ensure_correct_direction (int direction, Lisp_Coding_System *codesys,
4348 unsigned_char_dynarr *dst, unsigned int *flags,
4351 if ((!flags || (*flags & CODING_STATE_R2L)) &&
4352 direction == CHARSET_LEFT_TO_RIGHT)
4353 restore_left_to_right_direction (codesys, dst, flags, internal_p);
4354 else if (!CODING_SYSTEM_ISO2022_NO_ISO6429 (codesys)
4355 && (!flags || !(*flags & CODING_STATE_R2L)) &&
4356 direction == CHARSET_RIGHT_TO_LEFT)
4358 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
4360 Dynarr_add (dst, ISO_CODE_ESC);
4361 Dynarr_add (dst, '[');
4363 else if (internal_p)
4364 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
4366 Dynarr_add (dst, ISO_CODE_CSI);
4367 Dynarr_add (dst, '2');
4368 Dynarr_add (dst, ']');
4370 *flags |= CODING_STATE_R2L;
4374 /* Convert ISO2022-format data to internal format. */
4377 decode_coding_iso2022 (Lstream *decoding, CONST unsigned char *src,
4378 unsigned_char_dynarr *dst, unsigned int n)
4380 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
4381 unsigned int flags = str->flags;
4382 unsigned int ch = str->ch;
4383 eol_type_t eol_type = str->eol_type;
4384 #ifdef ENABLE_COMPOSITE_CHARS
4385 unsigned_char_dynarr *real_dst = dst;
4387 Lisp_Object coding_system;
4389 XSETCODING_SYSTEM (coding_system, str->codesys);
4391 #ifdef ENABLE_COMPOSITE_CHARS
4392 if (flags & CODING_STATE_COMPOSITE)
4393 dst = str->iso2022.composite_chars;
4394 #endif /* ENABLE_COMPOSITE_CHARS */
4398 unsigned char c = *src++;
4399 if (flags & CODING_STATE_ESCAPE)
4400 { /* Within ESC sequence */
4401 int retval = parse_iso2022_esc (coding_system, &str->iso2022,
4406 switch (str->iso2022.esc)
4408 #ifdef ENABLE_COMPOSITE_CHARS
4409 case ISO_ESC_START_COMPOSITE:
4410 if (str->iso2022.composite_chars)
4411 Dynarr_reset (str->iso2022.composite_chars);
4413 str->iso2022.composite_chars = Dynarr_new (unsigned_char);
4414 dst = str->iso2022.composite_chars;
4416 case ISO_ESC_END_COMPOSITE:
4418 Bufbyte comstr[MAX_EMCHAR_LEN];
4420 Emchar emch = lookup_composite_char (Dynarr_atp (dst, 0),
4421 Dynarr_length (dst));
4423 len = set_charptr_emchar (comstr, emch);
4424 Dynarr_add_many (dst, comstr, len);
4427 #endif /* ENABLE_COMPOSITE_CHARS */
4429 case ISO_ESC_LITERAL:
4430 DECODE_ADD_BINARY_CHAR (c, dst);
4434 /* Everything else handled already */
4439 /* Attempted error recovery. */
4440 if (str->iso2022.output_direction_sequence)
4441 ensure_correct_direction (flags & CODING_STATE_R2L ?
4442 CHARSET_RIGHT_TO_LEFT :
4443 CHARSET_LEFT_TO_RIGHT,
4444 str->codesys, dst, 0, 1);
4445 /* More error recovery. */
4446 if (!retval || str->iso2022.output_literally)
4448 /* Output the (possibly invalid) sequence */
4450 for (i = 0; i < str->iso2022.esc_bytes_index; i++)
4451 DECODE_ADD_BINARY_CHAR (str->iso2022.esc_bytes[i], dst);
4452 flags &= CODING_STATE_ISO2022_LOCK;
4454 n++, src--;/* Repeat the loop with the same character. */
4457 /* No sense in reprocessing the final byte of the
4458 escape sequence; it could mess things up anyway.
4460 DECODE_ADD_BINARY_CHAR (c, dst);
4465 else if (BYTE_C0_P (c) || BYTE_C1_P (c))
4466 { /* Control characters */
4468 /***** Error-handling *****/
4470 /* If we were in the middle of a character, dump out the
4471 partial character. */
4472 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4474 /* If we just saw a single-shift character, dump it out.
4475 This may dump out the wrong sort of single-shift character,
4476 but least it will give an indication that something went
4478 if (flags & CODING_STATE_SS2)
4480 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS2, dst);
4481 flags &= ~CODING_STATE_SS2;
4483 if (flags & CODING_STATE_SS3)
4485 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS3, dst);
4486 flags &= ~CODING_STATE_SS3;
4489 /***** Now handle the control characters. *****/
4492 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
4494 flags &= CODING_STATE_ISO2022_LOCK;
4496 if (!parse_iso2022_esc (coding_system, &str->iso2022, c, &flags, 1))
4497 DECODE_ADD_BINARY_CHAR (c, dst);
4500 { /* Graphic characters */
4501 Lisp_Object charset;
4507 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
4509 /* Now determine the charset. */
4510 reg = ((flags & CODING_STATE_SS2) ? 2
4511 : (flags & CODING_STATE_SS3) ? 3
4512 : !BYTE_ASCII_P (c) ? str->iso2022.register_right
4513 : str->iso2022.register_left);
4514 charset = str->iso2022.charset[reg];
4516 /* Error checking: */
4517 if (! CHARSETP (charset)
4518 || str->iso2022.invalid_designated[reg]
4519 || (((c & 0x7F) == ' ' || (c & 0x7F) == ISO_CODE_DEL)
4520 && XCHARSET_CHARS (charset) == 94))
4521 /* Mrmph. We are trying to invoke a register that has no
4522 or an invalid charset in it, or trying to add a character
4523 outside the range of the charset. Insert that char literally
4524 to preserve it for the output. */
4526 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4527 DECODE_ADD_BINARY_CHAR (c, dst);
4532 /* Things are probably hunky-dorey. */
4534 /* Fetch reverse charset, maybe. */
4535 if (((flags & CODING_STATE_R2L) &&
4536 XCHARSET_DIRECTION (charset) == CHARSET_LEFT_TO_RIGHT)
4538 (!(flags & CODING_STATE_R2L) &&
4539 XCHARSET_DIRECTION (charset) == CHARSET_RIGHT_TO_LEFT))
4541 Lisp_Object new_charset =
4542 XCHARSET_REVERSE_DIRECTION_CHARSET (charset);
4543 if (!NILP (new_charset))
4544 charset = new_charset;
4548 if (XCHARSET_DIMENSION (charset) == 1)
4550 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4552 (MAKE_CHAR (charset, c & 0x7F, 0), dst);
4557 (MAKE_CHAR (charset, ch & 0x7F, c & 0x7F), dst);
4563 lb = XCHARSET_LEADING_BYTE (charset);
4564 switch (XCHARSET_REP_BYTES (charset))
4567 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4568 Dynarr_add (dst, c & 0x7F);
4571 case 2: /* one-byte official */
4572 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4573 Dynarr_add (dst, lb);
4574 Dynarr_add (dst, c | 0x80);
4577 case 3: /* one-byte private or two-byte official */
4578 if (XCHARSET_PRIVATE_P (charset))
4580 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4581 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_1);
4582 Dynarr_add (dst, lb);
4583 Dynarr_add (dst, c | 0x80);
4589 Dynarr_add (dst, lb);
4590 Dynarr_add (dst, ch | 0x80);
4591 Dynarr_add (dst, c | 0x80);
4599 default: /* two-byte private */
4602 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_2);
4603 Dynarr_add (dst, lb);
4604 Dynarr_add (dst, ch | 0x80);
4605 Dynarr_add (dst, c | 0x80);
4615 flags &= CODING_STATE_ISO2022_LOCK;
4618 label_continue_loop:;
4621 if (flags & CODING_STATE_END)
4622 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4629 /***** ISO2022 encoder *****/
4631 /* Designate CHARSET into register REG. */
4634 iso2022_designate (Lisp_Object charset, unsigned char reg,
4635 struct encoding_stream *str, unsigned_char_dynarr *dst)
4637 static CONST char inter94[] = "()*+";
4638 static CONST char inter96[] = ",-./";
4640 unsigned char final;
4641 Lisp_Object old_charset = str->iso2022.charset[reg];
4643 str->iso2022.charset[reg] = charset;
4644 if (!CHARSETP (charset))
4645 /* charset might be an initial nil or t. */
4647 type = XCHARSET_TYPE (charset);
4648 final = XCHARSET_FINAL (charset);
4649 if (!str->iso2022.force_charset_on_output[reg] &&
4650 CHARSETP (old_charset) &&
4651 XCHARSET_TYPE (old_charset) == type &&
4652 XCHARSET_FINAL (old_charset) == final)
4655 str->iso2022.force_charset_on_output[reg] = 0;
4658 charset_conversion_spec_dynarr *dyn =
4659 str->codesys->iso2022.output_conv;
4665 for (i = 0; i < Dynarr_length (dyn); i++)
4667 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
4668 if (EQ (charset, spec->from_charset))
4669 charset = spec->to_charset;
4674 Dynarr_add (dst, ISO_CODE_ESC);
4677 case CHARSET_TYPE_94:
4678 Dynarr_add (dst, inter94[reg]);
4680 case CHARSET_TYPE_96:
4681 Dynarr_add (dst, inter96[reg]);
4683 case CHARSET_TYPE_94X94:
4684 Dynarr_add (dst, '$');
4686 || !(CODING_SYSTEM_ISO2022_SHORT (str->codesys))
4689 Dynarr_add (dst, inter94[reg]);
4691 case CHARSET_TYPE_96X96:
4692 Dynarr_add (dst, '$');
4693 Dynarr_add (dst, inter96[reg]);
4696 Dynarr_add (dst, final);
4700 ensure_normal_shift (struct encoding_stream *str, unsigned_char_dynarr *dst)
4702 if (str->iso2022.register_left != 0)
4704 Dynarr_add (dst, ISO_CODE_SI);
4705 str->iso2022.register_left = 0;
4710 ensure_shift_out (struct encoding_stream *str, unsigned_char_dynarr *dst)
4712 if (str->iso2022.register_left != 1)
4714 Dynarr_add (dst, ISO_CODE_SO);
4715 str->iso2022.register_left = 1;
4720 char_encode_iso2022 (struct encoding_stream *str, Emchar ch,
4721 unsigned_char_dynarr *dst, unsigned int *flags)
4723 unsigned char charmask;
4724 Lisp_Coding_System* codesys = str->codesys;
4725 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
4727 Lisp_Object charset = str->iso2022.current_charset;
4728 int half = str->iso2022.current_half;
4729 unsigned int byte1, byte2;
4733 restore_left_to_right_direction (codesys, dst, flags, 0);
4735 /* Make sure G0 contains ASCII */
4736 if ((ch > ' ' && ch < ISO_CODE_DEL)
4737 || !CODING_SYSTEM_ISO2022_NO_ASCII_CNTL (codesys))
4739 ensure_normal_shift (str, dst);
4740 iso2022_designate (Vcharset_ascii, 0, str, dst);
4743 /* If necessary, restore everything to the default state
4745 if (ch == '\n' && !(CODING_SYSTEM_ISO2022_NO_ASCII_EOL (codesys)))
4747 restore_left_to_right_direction (codesys, dst, flags, 0);
4749 ensure_normal_shift (str, dst);
4751 for (i = 0; i < 4; i++)
4753 Lisp_Object initial_charset =
4754 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
4755 iso2022_designate (initial_charset, i, str, dst);
4760 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
4761 Dynarr_add (dst, '\r');
4762 if (eol_type != EOL_CR)
4763 Dynarr_add (dst, ch);
4767 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
4768 && fit_to_be_escape_quoted (ch))
4769 Dynarr_add (dst, ISO_CODE_ESC);
4770 Dynarr_add (dst, ch);
4773 else if ( (0x80 <= ch) && (ch <= 0x9f) )
4775 charmask = (half == 0 ? 0x00 : 0x80);
4777 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
4778 && fit_to_be_escape_quoted (ch))
4779 Dynarr_add (dst, ISO_CODE_ESC);
4780 /* you asked for it ... */
4781 Dynarr_add (dst, ch);
4787 BREAKUP_CHAR (ch, charset, byte1, byte2);
4788 ensure_correct_direction (XCHARSET_DIRECTION (charset),
4789 codesys, dst, flags, 0);
4791 /* Now determine which register to use. */
4793 for (i = 0; i < 4; i++)
4795 if (EQ (charset, str->iso2022.charset[i]) ||
4797 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i)))
4806 if (XCHARSET_GRAPHIC (charset) != 0)
4808 if (!NILP (str->iso2022.charset[1]) &&
4809 (!CODING_SYSTEM_ISO2022_SEVEN (codesys)
4810 || CODING_SYSTEM_ISO2022_LOCK_SHIFT (codesys)))
4812 else if (!NILP (str->iso2022.charset[2]))
4814 else if (!NILP (str->iso2022.charset[3]))
4823 iso2022_designate (charset, reg, str, dst);
4825 /* Now invoke that register. */
4829 ensure_normal_shift (str, dst);
4833 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
4835 ensure_shift_out (str, dst);
4842 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
4844 Dynarr_add (dst, ISO_CODE_ESC);
4845 Dynarr_add (dst, 'N');
4850 Dynarr_add (dst, ISO_CODE_SS2);
4855 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
4857 Dynarr_add (dst, ISO_CODE_ESC);
4858 Dynarr_add (dst, 'O');
4863 Dynarr_add (dst, ISO_CODE_SS3);
4871 charmask = (half == 0 ? 0x00 : 0x80);
4873 switch (XCHARSET_DIMENSION (charset))
4876 Dynarr_add (dst, byte1 | charmask);
4879 Dynarr_add (dst, byte1 | charmask);
4880 Dynarr_add (dst, byte2 | charmask);
4886 str->iso2022.current_charset = charset;
4887 str->iso2022.current_half = half;
4891 char_finish_iso2022 (struct encoding_stream *str, unsigned_char_dynarr *dst,
4892 unsigned int *flags)
4894 Lisp_Coding_System* codesys = str->codesys;
4897 restore_left_to_right_direction (codesys, dst, flags, 0);
4898 ensure_normal_shift (str, dst);
4899 for (i = 0; i < 4; i++)
4901 Lisp_Object initial_charset
4902 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
4903 iso2022_designate (initial_charset, i, str, dst);
4908 /************************************************************************/
4909 /* No-conversion methods */
4910 /************************************************************************/
4912 /* This is used when reading in "binary" files -- i.e. files that may
4913 contain all 256 possible byte values and that are not to be
4914 interpreted as being in any particular decoding. */
4916 decode_coding_no_conversion (Lstream *decoding, CONST unsigned char *src,
4917 unsigned_char_dynarr *dst, unsigned int n)
4920 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
4921 unsigned int flags = str->flags;
4922 unsigned int ch = str->ch;
4923 eol_type_t eol_type = str->eol_type;
4929 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
4930 DECODE_ADD_BINARY_CHAR (c, dst);
4931 label_continue_loop:;
4934 DECODE_HANDLE_END_OF_CONVERSION (flags, ch, dst);
4941 encode_coding_no_conversion (Lstream *encoding, CONST unsigned char *src,
4942 unsigned_char_dynarr *dst, unsigned int n)
4945 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
4946 unsigned int flags = str->flags;
4947 unsigned int ch = str->ch;
4948 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
4950 unsigned char char_boundary = str->iso2022.current_char_boundary;
4957 if (char_boundary == 0)
4963 else if ( c >= 0xf8 )
4968 else if ( c >= 0xf0 )
4973 else if ( c >= 0xe0 )
4978 else if ( c >= 0xc0 )
4988 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
4989 Dynarr_add (dst, '\r');
4990 if (eol_type != EOL_CR)
4991 Dynarr_add (dst, c);
4994 Dynarr_add (dst, c);
4997 else if (char_boundary == 1)
4999 ch = ( ch << 6 ) | ( c & 0x3f );
5000 Dynarr_add (dst, ch & 0xff);
5005 ch = ( ch << 6 ) | ( c & 0x3f );
5008 #else /* not UTF2000 */
5011 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5012 Dynarr_add (dst, '\r');
5013 if (eol_type != EOL_CR)
5014 Dynarr_add (dst, '\n');
5017 else if (BYTE_ASCII_P (c))
5020 Dynarr_add (dst, c);
5022 else if (BUFBYTE_LEADING_BYTE_P (c))
5025 if (c == LEADING_BYTE_LATIN_ISO8859_1 ||
5026 c == LEADING_BYTE_CONTROL_1)
5029 Dynarr_add (dst, '~'); /* untranslatable character */
5033 if (ch == LEADING_BYTE_LATIN_ISO8859_1)
5034 Dynarr_add (dst, c);
5035 else if (ch == LEADING_BYTE_CONTROL_1)
5038 Dynarr_add (dst, c - 0x20);
5040 /* else it should be the second or third byte of an
5041 untranslatable character, so ignore it */
5044 #endif /* not UTF2000 */
5050 str->iso2022.current_char_boundary = char_boundary;
5055 /************************************************************************/
5056 /* Simple internal/external functions */
5057 /************************************************************************/
5059 static Extbyte_dynarr *conversion_out_dynarr;
5060 static Bufbyte_dynarr *conversion_in_dynarr;
5062 /* Determine coding system from coding format */
5064 /* #### not correct for all values of `fmt'! */
5066 external_data_format_to_coding_system (enum external_data_format fmt)
5070 case FORMAT_FILENAME:
5071 case FORMAT_TERMINAL:
5072 if (EQ (Vfile_name_coding_system, Qnil) ||
5073 EQ (Vfile_name_coding_system, Qbinary))
5076 return Fget_coding_system (Vfile_name_coding_system);
5079 return Fget_coding_system (Qctext);
5087 convert_to_external_format (CONST Bufbyte *ptr,
5090 enum external_data_format fmt)
5092 Lisp_Object coding_system = external_data_format_to_coding_system (fmt);
5094 if (!conversion_out_dynarr)
5095 conversion_out_dynarr = Dynarr_new (Extbyte);
5097 Dynarr_reset (conversion_out_dynarr);
5099 if (NILP (coding_system))
5101 CONST Bufbyte *end = ptr + len;
5107 (*ptr < 0xc0) ? *ptr :
5108 ((*ptr & 0x1f) << 6) | (*(ptr+1) & 0x3f);
5111 (BYTE_ASCII_P (*ptr)) ? *ptr :
5112 (*ptr == LEADING_BYTE_CONTROL_1) ? (*(ptr+1) - 0x20) :
5113 (*ptr == LEADING_BYTE_LATIN_ISO8859_1) ? (*(ptr+1)) :
5116 Dynarr_add (conversion_out_dynarr, (Extbyte) c);
5120 #ifdef ERROR_CHECK_BUFPOS
5121 assert (ptr == end);
5126 Lisp_Object instream, outstream, da_outstream;
5127 Lstream *istr, *ostr;
5128 struct gcpro gcpro1, gcpro2, gcpro3;
5129 char tempbuf[1024]; /* some random amount */
5131 instream = make_fixed_buffer_input_stream ((unsigned char *) ptr, len);
5132 da_outstream = make_dynarr_output_stream
5133 ((unsigned_char_dynarr *) conversion_out_dynarr);
5135 make_encoding_output_stream (XLSTREAM (da_outstream), coding_system);
5136 istr = XLSTREAM (instream);
5137 ostr = XLSTREAM (outstream);
5138 GCPRO3 (instream, outstream, da_outstream);
5141 int size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
5144 Lstream_write (ostr, tempbuf, size_in_bytes);
5146 Lstream_close (istr);
5147 Lstream_close (ostr);
5149 Lstream_delete (istr);
5150 Lstream_delete (ostr);
5151 Lstream_delete (XLSTREAM (da_outstream));
5154 *len_out = Dynarr_length (conversion_out_dynarr);
5155 Dynarr_add (conversion_out_dynarr, 0); /* remember to zero-terminate! */
5156 return Dynarr_atp (conversion_out_dynarr, 0);
5160 convert_from_external_format (CONST Extbyte *ptr,
5163 enum external_data_format fmt)
5165 Lisp_Object coding_system = external_data_format_to_coding_system (fmt);
5167 if (!conversion_in_dynarr)
5168 conversion_in_dynarr = Dynarr_new (Bufbyte);
5170 Dynarr_reset (conversion_in_dynarr);
5172 if (NILP (coding_system))
5174 CONST Extbyte *end = ptr + len;
5175 for (; ptr < end; ptr++)
5178 DECODE_ADD_BINARY_CHAR (c, conversion_in_dynarr);
5183 Lisp_Object instream, outstream, da_outstream;
5184 Lstream *istr, *ostr;
5185 struct gcpro gcpro1, gcpro2, gcpro3;
5186 char tempbuf[1024]; /* some random amount */
5188 instream = make_fixed_buffer_input_stream ((unsigned char *) ptr, len);
5189 da_outstream = make_dynarr_output_stream
5190 ((unsigned_char_dynarr *) conversion_in_dynarr);
5192 make_decoding_output_stream (XLSTREAM (da_outstream), coding_system);
5193 istr = XLSTREAM (instream);
5194 ostr = XLSTREAM (outstream);
5195 GCPRO3 (instream, outstream, da_outstream);
5198 int size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
5201 Lstream_write (ostr, tempbuf, size_in_bytes);
5203 Lstream_close (istr);
5204 Lstream_close (ostr);
5206 Lstream_delete (istr);
5207 Lstream_delete (ostr);
5208 Lstream_delete (XLSTREAM (da_outstream));
5211 *len_out = Dynarr_length (conversion_in_dynarr);
5212 Dynarr_add (conversion_in_dynarr, 0); /* remember to zero-terminate! */
5213 return Dynarr_atp (conversion_in_dynarr, 0);
5217 /************************************************************************/
5218 /* Initialization */
5219 /************************************************************************/
5222 syms_of_file_coding (void)
5224 defsymbol (&Qbuffer_file_coding_system, "buffer-file-coding-system");
5225 deferror (&Qcoding_system_error, "coding-system-error",
5226 "Coding-system error", Qio_error);
5228 DEFSUBR (Fcoding_system_p);
5229 DEFSUBR (Ffind_coding_system);
5230 DEFSUBR (Fget_coding_system);
5231 DEFSUBR (Fcoding_system_list);
5232 DEFSUBR (Fcoding_system_name);
5233 DEFSUBR (Fmake_coding_system);
5234 DEFSUBR (Fcopy_coding_system);
5235 DEFSUBR (Fdefine_coding_system_alias);
5236 DEFSUBR (Fsubsidiary_coding_system);
5238 DEFSUBR (Fcoding_system_type);
5239 DEFSUBR (Fcoding_system_doc_string);
5241 DEFSUBR (Fcoding_system_charset);
5243 DEFSUBR (Fcoding_system_property);
5245 DEFSUBR (Fcoding_category_list);
5246 DEFSUBR (Fset_coding_priority_list);
5247 DEFSUBR (Fcoding_priority_list);
5248 DEFSUBR (Fset_coding_category_system);
5249 DEFSUBR (Fcoding_category_system);
5251 DEFSUBR (Fdetect_coding_region);
5252 DEFSUBR (Fdecode_coding_region);
5253 DEFSUBR (Fencode_coding_region);
5255 DEFSUBR (Fdecode_shift_jis_char);
5256 DEFSUBR (Fencode_shift_jis_char);
5257 DEFSUBR (Fdecode_big5_char);
5258 DEFSUBR (Fencode_big5_char);
5260 defsymbol (&Qcoding_system_p, "coding-system-p");
5261 defsymbol (&Qno_conversion, "no-conversion");
5262 defsymbol (&Qraw_text, "raw-text");
5264 defsymbol (&Qbig5, "big5");
5265 defsymbol (&Qshift_jis, "shift-jis");
5266 defsymbol (&Qucs4, "ucs-4");
5267 defsymbol (&Qutf8, "utf-8");
5268 defsymbol (&Qccl, "ccl");
5269 defsymbol (&Qiso2022, "iso2022");
5271 defsymbol (&Qmnemonic, "mnemonic");
5272 defsymbol (&Qeol_type, "eol-type");
5273 defsymbol (&Qpost_read_conversion, "post-read-conversion");
5274 defsymbol (&Qpre_write_conversion, "pre-write-conversion");
5276 defsymbol (&Qcr, "cr");
5277 defsymbol (&Qlf, "lf");
5278 defsymbol (&Qcrlf, "crlf");
5279 defsymbol (&Qeol_cr, "eol-cr");
5280 defsymbol (&Qeol_lf, "eol-lf");
5281 defsymbol (&Qeol_crlf, "eol-crlf");
5283 defsymbol (&Qcharset_g0, "charset-g0");
5284 defsymbol (&Qcharset_g1, "charset-g1");
5285 defsymbol (&Qcharset_g2, "charset-g2");
5286 defsymbol (&Qcharset_g3, "charset-g3");
5287 defsymbol (&Qforce_g0_on_output, "force-g0-on-output");
5288 defsymbol (&Qforce_g1_on_output, "force-g1-on-output");
5289 defsymbol (&Qforce_g2_on_output, "force-g2-on-output");
5290 defsymbol (&Qforce_g3_on_output, "force-g3-on-output");
5291 defsymbol (&Qno_iso6429, "no-iso6429");
5292 defsymbol (&Qinput_charset_conversion, "input-charset-conversion");
5293 defsymbol (&Qoutput_charset_conversion, "output-charset-conversion");
5295 defsymbol (&Qshort, "short");
5296 defsymbol (&Qno_ascii_eol, "no-ascii-eol");
5297 defsymbol (&Qno_ascii_cntl, "no-ascii-cntl");
5298 defsymbol (&Qseven, "seven");
5299 defsymbol (&Qlock_shift, "lock-shift");
5300 defsymbol (&Qescape_quoted, "escape-quoted");
5302 defsymbol (&Qencode, "encode");
5303 defsymbol (&Qdecode, "decode");
5306 defsymbol (&Qctext, "ctext");
5307 defsymbol (&coding_category_symbol[CODING_CATEGORY_SHIFT_JIS],
5309 defsymbol (&coding_category_symbol[CODING_CATEGORY_BIG5],
5311 defsymbol (&coding_category_symbol[CODING_CATEGORY_UCS4],
5313 defsymbol (&coding_category_symbol[CODING_CATEGORY_UTF8],
5315 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_7],
5317 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_DESIGNATE],
5319 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_1],
5321 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_2],
5323 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_LOCK_SHIFT],
5326 defsymbol (&coding_category_symbol[CODING_CATEGORY_NO_CONVERSION],
5331 lstream_type_create_file_coding (void)
5333 LSTREAM_HAS_METHOD (decoding, reader);
5334 LSTREAM_HAS_METHOD (decoding, writer);
5335 LSTREAM_HAS_METHOD (decoding, rewinder);
5336 LSTREAM_HAS_METHOD (decoding, seekable_p);
5337 LSTREAM_HAS_METHOD (decoding, flusher);
5338 LSTREAM_HAS_METHOD (decoding, closer);
5339 LSTREAM_HAS_METHOD (decoding, marker);
5341 LSTREAM_HAS_METHOD (encoding, reader);
5342 LSTREAM_HAS_METHOD (encoding, writer);
5343 LSTREAM_HAS_METHOD (encoding, rewinder);
5344 LSTREAM_HAS_METHOD (encoding, seekable_p);
5345 LSTREAM_HAS_METHOD (encoding, flusher);
5346 LSTREAM_HAS_METHOD (encoding, closer);
5347 LSTREAM_HAS_METHOD (encoding, marker);
5351 vars_of_file_coding (void)
5355 /* Initialize to something reasonable ... */
5356 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
5358 coding_category_system[i] = Qnil;
5359 coding_category_by_priority[i] = i;
5362 Fprovide (intern ("file-coding"));
5364 DEFVAR_LISP ("keyboard-coding-system", &Vkeyboard_coding_system /*
5365 Coding system used for TTY keyboard input.
5366 Not used under a windowing system.
5368 Vkeyboard_coding_system = Qnil;
5370 DEFVAR_LISP ("terminal-coding-system", &Vterminal_coding_system /*
5371 Coding system used for TTY display output.
5372 Not used under a windowing system.
5374 Vterminal_coding_system = Qnil;
5376 DEFVAR_LISP ("coding-system-for-read", &Vcoding_system_for_read /*
5377 Overriding coding system used when writing a file or process.
5378 You should *bind* this, not set it. If this is non-nil, it specifies
5379 the coding system that will be used when a file or process is read
5380 in, and overrides `buffer-file-coding-system-for-read',
5381 `insert-file-contents-pre-hook', etc. Use those variables instead of
5382 this one for permanent changes to the environment.
5384 Vcoding_system_for_read = Qnil;
5386 DEFVAR_LISP ("coding-system-for-write",
5387 &Vcoding_system_for_write /*
5388 Overriding coding system used when writing a file or process.
5389 You should *bind* this, not set it. If this is non-nil, it specifies
5390 the coding system that will be used when a file or process is wrote
5391 in, and overrides `buffer-file-coding-system',
5392 `write-region-pre-hook', etc. Use those variables instead of this one
5393 for permanent changes to the environment.
5395 Vcoding_system_for_write = Qnil;
5397 DEFVAR_LISP ("file-name-coding-system", &Vfile_name_coding_system /*
5398 Coding system used to convert pathnames when accessing files.
5400 Vfile_name_coding_system = Qnil;
5402 DEFVAR_BOOL ("enable-multibyte-characters", &enable_multibyte_characters /*
5403 Non-nil means the buffer contents are regarded as multi-byte form
5404 of characters, not a binary code. This affects the display, file I/O,
5405 and behaviors of various editing commands.
5407 Setting this to nil does not do anything.
5409 enable_multibyte_characters = 1;
5413 complex_vars_of_file_coding (void)
5415 staticpro (&Vcoding_system_hash_table);
5416 Vcoding_system_hash_table =
5417 make_lisp_hash_table (50, HASH_TABLE_NON_WEAK, HASH_TABLE_EQ);
5419 the_codesys_prop_dynarr = Dynarr_new (codesys_prop);
5421 #define DEFINE_CODESYS_PROP(Prop_Type, Sym) do \
5423 struct codesys_prop csp; \
5425 csp.prop_type = (Prop_Type); \
5426 Dynarr_add (the_codesys_prop_dynarr, csp); \
5429 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qmnemonic);
5430 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_type);
5431 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_cr);
5432 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_crlf);
5433 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_lf);
5434 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpost_read_conversion);
5435 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpre_write_conversion);
5437 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g0);
5438 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g1);
5439 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g2);
5440 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g3);
5441 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g0_on_output);
5442 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g1_on_output);
5443 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g2_on_output);
5444 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g3_on_output);
5445 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qshort);
5446 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_eol);
5447 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_cntl);
5448 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qseven);
5449 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qlock_shift);
5450 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_iso6429);
5451 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qescape_quoted);
5452 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qinput_charset_conversion);
5453 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qoutput_charset_conversion);
5455 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qencode);
5456 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qdecode);
5458 /* Need to create this here or we're really screwed. */
5460 (Qraw_text, Qno_conversion,
5461 build_string ("Raw text, which means it converts only line-break-codes."),
5462 list2 (Qmnemonic, build_string ("Raw")));
5465 (Qbinary, Qno_conversion,
5466 build_string ("Binary, which means it does not convert anything."),
5467 list4 (Qeol_type, Qlf,
5468 Qmnemonic, build_string ("Binary")));
5473 build_string ("Coding-system of ISO/IEC 10646 UTF-8."),
5474 list2 (Qmnemonic, build_string ("UTF8")));
5477 Fdefine_coding_system_alias (Qno_conversion, Qraw_text);
5479 /* Need this for bootstrapping */
5480 coding_category_system[CODING_CATEGORY_NO_CONVERSION] =
5481 Fget_coding_system (Qraw_text);
5484 coding_category_system[CODING_CATEGORY_UTF8]
5485 = Fget_coding_system (Qutf8);