1 /* Code conversion functions.
2 Copyright (C) 1991, 1995 Free Software Foundation, Inc.
3 Copyright (C) 1995 Sun Microsystems, Inc.
4 Copyright (C) 1999,2000,2001,2002 MORIOKA Tomohiko
6 This file is part of XEmacs.
8 XEmacs is free software; you can redistribute it and/or modify it
9 under the terms of the GNU General Public License as published by the
10 Free Software Foundation; either version 2, or (at your option) any
13 XEmacs is distributed in the hope that it will be useful, but WITHOUT
14 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
18 You should have received a copy of the GNU General Public License
19 along with XEmacs; see the file COPYING. If not, write to
20 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
21 Boston, MA 02111-1307, USA. */
23 /* Synched up with: Mule 2.3. Not in FSF. */
25 /* Rewritten by Ben Wing <ben@xemacs.org>. */
26 /* Rewritten by MORIOKA Tomohiko <tomo@m17n.org> for XEmacs UTF-2000. */
40 #include "file-coding.h"
42 Lisp_Object Qcoding_system_error;
44 Lisp_Object Vkeyboard_coding_system;
45 Lisp_Object Vterminal_coding_system;
46 Lisp_Object Vcoding_system_for_read;
47 Lisp_Object Vcoding_system_for_write;
48 Lisp_Object Vfile_name_coding_system;
50 Lisp_Object Vcoded_charset_entity_reference_alist;
52 /* Table of symbols identifying each coding category. */
53 Lisp_Object coding_category_symbol[CODING_CATEGORY_LAST];
57 struct file_coding_dump {
58 /* Coding system currently associated with each coding category. */
59 Lisp_Object coding_category_system[CODING_CATEGORY_LAST];
61 /* Table of all coding categories in decreasing order of priority.
62 This describes a permutation of the possible coding categories. */
63 int coding_category_by_priority[CODING_CATEGORY_LAST];
65 #if defined(MULE) && !defined(UTF2000)
66 Lisp_Object ucs_to_mule_table[65536];
70 static const struct lrecord_description fcd_description_1[] = {
71 { XD_LISP_OBJECT_ARRAY, offsetof (struct file_coding_dump, coding_category_system), CODING_CATEGORY_LAST },
72 #if defined(MULE) && !defined(UTF2000)
73 { XD_LISP_OBJECT_ARRAY, offsetof (struct file_coding_dump, ucs_to_mule_table), countof (fcd->ucs_to_mule_table) },
78 static const struct struct_description fcd_description = {
79 sizeof (struct file_coding_dump),
83 Lisp_Object mule_to_ucs_table;
85 Lisp_Object Qcoding_systemp;
87 Lisp_Object Qraw_text, Qno_conversion, Qccl, Qiso2022;
88 /* Qinternal in general.c */
90 Lisp_Object Qmnemonic, Qeol_type;
91 Lisp_Object Qcr, Qcrlf, Qlf;
92 Lisp_Object Qeol_cr, Qeol_crlf, Qeol_lf;
93 Lisp_Object Qpost_read_conversion;
94 Lisp_Object Qpre_write_conversion;
97 Lisp_Object Qucs4, Qutf8;
98 Lisp_Object Qbig5, Qshift_jis;
99 Lisp_Object Qcharset_g0, Qcharset_g1, Qcharset_g2, Qcharset_g3;
100 Lisp_Object Qforce_g0_on_output, Qforce_g1_on_output;
101 Lisp_Object Qforce_g2_on_output, Qforce_g3_on_output;
102 Lisp_Object Qno_iso6429;
103 Lisp_Object Qinput_charset_conversion, Qoutput_charset_conversion;
104 Lisp_Object Qescape_quoted;
105 Lisp_Object Qshort, Qno_ascii_eol, Qno_ascii_cntl, Qseven, Qlock_shift;
108 Lisp_Object Qdisable_composition;
109 Lisp_Object Quse_entity_reference;
110 Lisp_Object Qd, Qx, QX;
112 Lisp_Object Qencode, Qdecode;
114 Lisp_Object Vcoding_system_hash_table;
116 int enable_multibyte_characters;
119 /* Additional information used by the ISO2022 decoder and detector. */
120 struct iso2022_decoder
122 /* CHARSET holds the character sets currently assigned to the G0
123 through G3 variables. It is initialized from the array
124 INITIAL_CHARSET in CODESYS. */
125 Lisp_Object charset[4];
127 /* Which registers are currently invoked into the left (GL) and
128 right (GR) halves of the 8-bit encoding space? */
129 int register_left, register_right;
131 /* ISO_ESC holds a value indicating part of an escape sequence
132 that has already been seen. */
133 enum iso_esc_flag esc;
135 /* This records the bytes we've seen so far in an escape sequence,
136 in case the sequence is invalid (we spit out the bytes unchanged). */
137 unsigned char esc_bytes[8];
139 /* Index for next byte to store in ISO escape sequence. */
142 #ifdef ENABLE_COMPOSITE_CHARS
143 /* Stuff seen so far when composing a string. */
144 unsigned_char_dynarr *composite_chars;
147 /* If we saw an invalid designation sequence for a particular
148 register, we flag it here and switch to ASCII. The next time we
149 see a valid designation for this register, we turn off the flag
150 and do the designation normally, but pretend the sequence was
151 invalid. The effect of all this is that (most of the time) the
152 escape sequences for both the switch to the unknown charset, and
153 the switch back to the known charset, get inserted literally into
154 the buffer and saved out as such. The hope is that we can
155 preserve the escape sequences so that the resulting written out
156 file makes sense. If we don't do any of this, the designation
157 to the invalid charset will be preserved but that switch back
158 to the known charset will probably get eaten because it was
159 the same charset that was already present in the register. */
160 unsigned char invalid_designated[4];
162 /* We try to do similar things as above for direction-switching
163 sequences. If we encountered a direction switch while an
164 invalid designation was present, or an invalid designation
165 just after a direction switch (i.e. no valid designation
166 encountered yet), we insert the direction-switch escape
167 sequence literally into the output stream, and later on
168 insert the corresponding direction-restoring escape sequence
170 unsigned int switched_dir_and_no_valid_charset_yet :1;
171 unsigned int invalid_switch_dir :1;
173 /* Tells the decoder to output the escape sequence literally
174 even though it was valid. Used in the games we play to
175 avoid lossage when we encounter invalid designations. */
176 unsigned int output_literally :1;
177 /* We encountered a direction switch followed by an invalid
178 designation. We didn't output the direction switch
179 literally because we didn't know about the invalid designation;
180 but we have to do so now. */
181 unsigned int output_direction_sequence :1;
184 EXFUN (Fcopy_coding_system, 2);
186 struct detection_state;
189 text_encode_generic (Lstream *encoding, const Bufbyte *src,
190 unsigned_char_dynarr *dst, size_t n);
192 static int detect_coding_sjis (struct detection_state *st,
193 const Extbyte *src, size_t n);
194 static void decode_coding_sjis (Lstream *decoding, const Extbyte *src,
195 unsigned_char_dynarr *dst, size_t n);
196 void char_encode_shift_jis (struct encoding_stream *str, Emchar c,
197 unsigned_char_dynarr *dst, unsigned int *flags);
198 void char_finish_shift_jis (struct encoding_stream *str,
199 unsigned_char_dynarr *dst, unsigned int *flags);
201 static int detect_coding_big5 (struct detection_state *st,
202 const Extbyte *src, size_t n);
203 static void decode_coding_big5 (Lstream *decoding, const Extbyte *src,
204 unsigned_char_dynarr *dst, size_t n);
205 void char_encode_big5 (struct encoding_stream *str, Emchar c,
206 unsigned_char_dynarr *dst, unsigned int *flags);
207 void char_finish_big5 (struct encoding_stream *str,
208 unsigned_char_dynarr *dst, unsigned int *flags);
210 static int detect_coding_ucs4 (struct detection_state *st,
211 const Extbyte *src, size_t n);
212 static void decode_coding_ucs4 (Lstream *decoding, const Extbyte *src,
213 unsigned_char_dynarr *dst, size_t n);
214 void char_encode_ucs4 (struct encoding_stream *str, Emchar c,
215 unsigned_char_dynarr *dst, unsigned int *flags);
216 void char_finish_ucs4 (struct encoding_stream *str,
217 unsigned_char_dynarr *dst, unsigned int *flags);
219 static int detect_coding_utf8 (struct detection_state *st,
220 const Extbyte *src, size_t n);
221 static void decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
222 unsigned_char_dynarr *dst, size_t n);
223 void char_encode_utf8 (struct encoding_stream *str, Emchar c,
224 unsigned_char_dynarr *dst, unsigned int *flags);
225 void char_finish_utf8 (struct encoding_stream *str,
226 unsigned_char_dynarr *dst, unsigned int *flags);
228 static int postprocess_iso2022_mask (int mask);
229 static void reset_iso2022 (Lisp_Object coding_system,
230 struct iso2022_decoder *iso);
231 static int detect_coding_iso2022 (struct detection_state *st,
232 const Extbyte *src, size_t n);
233 static void decode_coding_iso2022 (Lstream *decoding, const Extbyte *src,
234 unsigned_char_dynarr *dst, size_t n);
235 void char_encode_iso2022 (struct encoding_stream *str, Emchar c,
236 unsigned_char_dynarr *dst, unsigned int *flags);
237 void char_finish_iso2022 (struct encoding_stream *str,
238 unsigned_char_dynarr *dst, unsigned int *flags);
240 static void decode_coding_no_conversion (Lstream *decoding, const Extbyte *src,
241 unsigned_char_dynarr *dst, size_t n);
242 static void encode_coding_no_conversion (Lstream *encoding, const Bufbyte *src,
243 unsigned_char_dynarr *dst, size_t n);
244 static void mule_decode (Lstream *decoding, const Extbyte *src,
245 unsigned_char_dynarr *dst, size_t n);
246 static void mule_encode (Lstream *encoding, const Bufbyte *src,
247 unsigned_char_dynarr *dst, size_t n);
249 typedef struct codesys_prop codesys_prop;
258 Dynarr_declare (codesys_prop);
259 } codesys_prop_dynarr;
261 static const struct lrecord_description codesys_prop_description_1[] = {
262 { XD_LISP_OBJECT, offsetof (codesys_prop, sym) },
266 static const struct struct_description codesys_prop_description = {
267 sizeof (codesys_prop),
268 codesys_prop_description_1
271 static const struct lrecord_description codesys_prop_dynarr_description_1[] = {
272 XD_DYNARR_DESC (codesys_prop_dynarr, &codesys_prop_description),
276 static const struct struct_description codesys_prop_dynarr_description = {
277 sizeof (codesys_prop_dynarr),
278 codesys_prop_dynarr_description_1
281 codesys_prop_dynarr *the_codesys_prop_dynarr;
283 enum codesys_prop_enum
286 CODESYS_PROP_ISO2022,
291 /************************************************************************/
292 /* Coding system functions */
293 /************************************************************************/
295 static Lisp_Object mark_coding_system (Lisp_Object);
296 static void print_coding_system (Lisp_Object, Lisp_Object, int);
297 static void finalize_coding_system (void *header, int for_disksave);
300 static const struct lrecord_description ccs_description_1[] = {
301 { XD_LISP_OBJECT, offsetof (charset_conversion_spec, from_charset) },
302 { XD_LISP_OBJECT, offsetof (charset_conversion_spec, to_charset) },
306 static const struct struct_description ccs_description = {
307 sizeof (charset_conversion_spec),
311 static const struct lrecord_description ccsd_description_1[] = {
312 XD_DYNARR_DESC (charset_conversion_spec_dynarr, &ccs_description),
316 static const struct struct_description ccsd_description = {
317 sizeof (charset_conversion_spec_dynarr),
322 static const struct lrecord_description coding_system_description[] = {
323 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, name) },
324 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, doc_string) },
325 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, mnemonic) },
326 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, post_read_conversion) },
327 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, pre_write_conversion) },
328 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_lf) },
329 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_crlf) },
330 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_cr) },
332 { XD_LISP_OBJECT_ARRAY, offsetof (Lisp_Coding_System, iso2022.initial_charset), 4 },
333 { XD_STRUCT_PTR, offsetof (Lisp_Coding_System, iso2022.input_conv), 1, &ccsd_description },
334 { XD_STRUCT_PTR, offsetof (Lisp_Coding_System, iso2022.output_conv), 1, &ccsd_description },
335 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccl.decode) },
336 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccl.encode) },
338 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccs_priority_list) },
344 DEFINE_LRECORD_IMPLEMENTATION ("coding-system", coding_system,
345 mark_coding_system, print_coding_system,
346 finalize_coding_system,
347 0, 0, coding_system_description,
351 mark_coding_system (Lisp_Object obj)
353 Lisp_Coding_System *codesys = XCODING_SYSTEM (obj);
355 mark_object (CODING_SYSTEM_NAME (codesys));
356 mark_object (CODING_SYSTEM_DOC_STRING (codesys));
357 mark_object (CODING_SYSTEM_MNEMONIC (codesys));
358 mark_object (CODING_SYSTEM_EOL_LF (codesys));
359 mark_object (CODING_SYSTEM_EOL_CRLF (codesys));
360 mark_object (CODING_SYSTEM_EOL_CR (codesys));
362 switch (CODING_SYSTEM_TYPE (codesys))
366 case CODESYS_ISO2022:
367 for (i = 0; i < 4; i++)
368 mark_object (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i));
369 if (codesys->iso2022.input_conv)
371 for (i = 0; i < Dynarr_length (codesys->iso2022.input_conv); i++)
373 struct charset_conversion_spec *ccs =
374 Dynarr_atp (codesys->iso2022.input_conv, i);
375 mark_object (ccs->from_charset);
376 mark_object (ccs->to_charset);
379 if (codesys->iso2022.output_conv)
381 for (i = 0; i < Dynarr_length (codesys->iso2022.output_conv); i++)
383 struct charset_conversion_spec *ccs =
384 Dynarr_atp (codesys->iso2022.output_conv, i);
385 mark_object (ccs->from_charset);
386 mark_object (ccs->to_charset);
393 mark_object (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 0));
394 mark_object (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 1));
399 mark_object (CODING_SYSTEM_CCL_DECODE (codesys));
400 mark_object (CODING_SYSTEM_CCL_ENCODE (codesys));
407 mark_object (CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys));
409 mark_object (CODING_SYSTEM_CCS_PRIORITY_LIST (codesys));
411 return CODING_SYSTEM_POST_READ_CONVERSION (codesys);
415 print_coding_system (Lisp_Object obj, Lisp_Object printcharfun,
418 Lisp_Coding_System *c = XCODING_SYSTEM (obj);
420 error ("printing unreadable object #<coding_system 0x%x>",
423 write_c_string ("#<coding_system ", printcharfun);
424 print_internal (c->name, printcharfun, 1);
425 write_c_string (">", printcharfun);
429 finalize_coding_system (void *header, int for_disksave)
431 Lisp_Coding_System *c = (Lisp_Coding_System *) header;
432 /* Since coding systems never go away, this function is not
433 necessary. But it would be necessary if we changed things
434 so that coding systems could go away. */
435 if (!for_disksave) /* see comment in lstream.c */
437 switch (CODING_SYSTEM_TYPE (c))
440 case CODESYS_ISO2022:
441 if (c->iso2022.input_conv)
443 Dynarr_free (c->iso2022.input_conv);
444 c->iso2022.input_conv = 0;
446 if (c->iso2022.output_conv)
448 Dynarr_free (c->iso2022.output_conv);
449 c->iso2022.output_conv = 0;
460 symbol_to_eol_type (Lisp_Object symbol)
462 CHECK_SYMBOL (symbol);
463 if (NILP (symbol)) return EOL_AUTODETECT;
464 if (EQ (symbol, Qlf)) return EOL_LF;
465 if (EQ (symbol, Qcrlf)) return EOL_CRLF;
466 if (EQ (symbol, Qcr)) return EOL_CR;
468 signal_simple_error ("Unrecognized eol type", symbol);
469 return EOL_AUTODETECT; /* not reached */
473 eol_type_to_symbol (eol_type_t type)
478 case EOL_LF: return Qlf;
479 case EOL_CRLF: return Qcrlf;
480 case EOL_CR: return Qcr;
481 case EOL_AUTODETECT: return Qnil;
486 setup_eol_coding_systems (Lisp_Coding_System *codesys)
488 Lisp_Object codesys_obj;
489 int len = string_length (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name);
490 char *codesys_name = (char *) alloca (len + 7);
492 char *codesys_mnemonic=0;
494 Lisp_Object codesys_name_sym, sub_codesys_obj;
498 XSETCODING_SYSTEM (codesys_obj, codesys);
500 memcpy (codesys_name,
501 string_data (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name), len);
503 if (STRINGP (CODING_SYSTEM_MNEMONIC (codesys)))
505 mlen = XSTRING_LENGTH (CODING_SYSTEM_MNEMONIC (codesys));
506 codesys_mnemonic = (char *) alloca (mlen + 7);
507 memcpy (codesys_mnemonic,
508 XSTRING_DATA (CODING_SYSTEM_MNEMONIC (codesys)), mlen);
511 #define DEFINE_SUB_CODESYS(op_sys, op_sys_abbr, Type) do { \
512 strcpy (codesys_name + len, "-" op_sys); \
514 strcpy (codesys_mnemonic + mlen, op_sys_abbr); \
515 codesys_name_sym = intern (codesys_name); \
516 sub_codesys_obj = Fcopy_coding_system (codesys_obj, codesys_name_sym); \
517 XCODING_SYSTEM_EOL_TYPE (sub_codesys_obj) = Type; \
519 XCODING_SYSTEM_MNEMONIC(sub_codesys_obj) = \
520 build_string (codesys_mnemonic); \
521 CODING_SYSTEM_##Type (codesys) = sub_codesys_obj; \
524 DEFINE_SUB_CODESYS("unix", "", EOL_LF);
525 DEFINE_SUB_CODESYS("dos", ":T", EOL_CRLF);
526 DEFINE_SUB_CODESYS("mac", ":t", EOL_CR);
529 DEFUN ("coding-system-p", Fcoding_system_p, 1, 1, 0, /*
530 Return t if OBJECT is a coding system.
531 A coding system is an object that defines how text containing multiple
532 character sets is encoded into a stream of (typically 8-bit) bytes.
533 The coding system is used to decode the stream into a series of
534 characters (which may be from multiple charsets) when the text is read
535 from a file or process, and is used to encode the text back into the
536 same format when it is written out to a file or process.
538 For example, many ISO2022-compliant coding systems (such as Compound
539 Text, which is used for inter-client data under the X Window System)
540 use escape sequences to switch between different charsets -- Japanese
541 Kanji, for example, is invoked with "ESC $ ( B"; ASCII is invoked
542 with "ESC ( B"; and Cyrillic is invoked with "ESC - L". See
543 `make-coding-system' for more information.
545 Coding systems are normally identified using a symbol, and the
546 symbol is accepted in place of the actual coding system object whenever
547 a coding system is called for. (This is similar to how faces work.)
551 return CODING_SYSTEMP (object) ? Qt : Qnil;
554 DEFUN ("find-coding-system", Ffind_coding_system, 1, 1, 0, /*
555 Retrieve the coding system of the given name.
557 If CODING-SYSTEM-OR-NAME is a coding-system object, it is simply
558 returned. Otherwise, CODING-SYSTEM-OR-NAME should be a symbol.
559 If there is no such coding system, nil is returned. Otherwise the
560 associated coding system object is returned.
562 (coding_system_or_name))
564 if (NILP (coding_system_or_name))
565 coding_system_or_name = Qbinary;
566 else if (CODING_SYSTEMP (coding_system_or_name))
567 return coding_system_or_name;
569 CHECK_SYMBOL (coding_system_or_name);
573 coding_system_or_name =
574 Fgethash (coding_system_or_name, Vcoding_system_hash_table, Qnil);
576 if (CODING_SYSTEMP (coding_system_or_name) || NILP (coding_system_or_name))
577 return coding_system_or_name;
581 DEFUN ("get-coding-system", Fget_coding_system, 1, 1, 0, /*
582 Retrieve the coding system of the given name.
583 Same as `find-coding-system' except that if there is no such
584 coding system, an error is signaled instead of returning nil.
588 Lisp_Object coding_system = Ffind_coding_system (name);
590 if (NILP (coding_system))
591 signal_simple_error ("No such coding system", name);
592 return coding_system;
595 /* We store the coding systems in hash tables with the names as the key and the
596 actual coding system object as the value. Occasionally we need to use them
597 in a list format. These routines provide us with that. */
598 struct coding_system_list_closure
600 Lisp_Object *coding_system_list;
604 add_coding_system_to_list_mapper (Lisp_Object key, Lisp_Object value,
605 void *coding_system_list_closure)
607 /* This function can GC */
608 struct coding_system_list_closure *cscl =
609 (struct coding_system_list_closure *) coding_system_list_closure;
610 Lisp_Object *coding_system_list = cscl->coding_system_list;
612 *coding_system_list = Fcons (key, *coding_system_list);
616 DEFUN ("coding-system-list", Fcoding_system_list, 0, 0, 0, /*
617 Return a list of the names of all defined coding systems.
621 Lisp_Object coding_system_list = Qnil;
623 struct coding_system_list_closure coding_system_list_closure;
625 GCPRO1 (coding_system_list);
626 coding_system_list_closure.coding_system_list = &coding_system_list;
627 elisp_maphash (add_coding_system_to_list_mapper, Vcoding_system_hash_table,
628 &coding_system_list_closure);
631 return coding_system_list;
634 DEFUN ("coding-system-name", Fcoding_system_name, 1, 1, 0, /*
635 Return the name of the given coding system.
639 coding_system = Fget_coding_system (coding_system);
640 return XCODING_SYSTEM_NAME (coding_system);
643 static Lisp_Coding_System *
644 allocate_coding_system (enum coding_system_type type, Lisp_Object name)
646 Lisp_Coding_System *codesys =
647 alloc_lcrecord_type (Lisp_Coding_System, &lrecord_coding_system);
649 zero_lcrecord (codesys);
650 CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = Qnil;
651 CODING_SYSTEM_POST_READ_CONVERSION (codesys) = Qnil;
652 CODING_SYSTEM_EOL_TYPE (codesys) = EOL_AUTODETECT;
653 CODING_SYSTEM_EOL_CRLF (codesys) = Qnil;
654 CODING_SYSTEM_EOL_CR (codesys) = Qnil;
655 CODING_SYSTEM_EOL_LF (codesys) = Qnil;
656 CODING_SYSTEM_TYPE (codesys) = type;
657 CODING_SYSTEM_MNEMONIC (codesys) = Qnil;
660 CODING_SYSTEM_CCS_PRIORITY_LIST (codesys) = Qnil;
662 if (type == CODESYS_ISO2022)
665 for (i = 0; i < 4; i++)
666 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i) = Qnil;
669 if (type == CODESYS_BIG5)
671 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 0)
673 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 1)
674 = Vcharset_chinese_big5;
675 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 2)
677 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 3)
681 else if (type == CODESYS_CCL)
683 CODING_SYSTEM_CCL_DECODE (codesys) = Qnil;
684 CODING_SYSTEM_CCL_ENCODE (codesys) = Qnil;
687 CODING_SYSTEM_NAME (codesys) = name;
693 /* Given a list of charset conversion specs as specified in a Lisp
694 program, parse it into STORE_HERE. */
697 parse_charset_conversion_specs (charset_conversion_spec_dynarr *store_here,
698 Lisp_Object spec_list)
702 EXTERNAL_LIST_LOOP (rest, spec_list)
704 Lisp_Object car = XCAR (rest);
705 Lisp_Object from, to;
706 struct charset_conversion_spec spec;
708 if (!CONSP (car) || !CONSP (XCDR (car)) || !NILP (XCDR (XCDR (car))))
709 signal_simple_error ("Invalid charset conversion spec", car);
710 from = Fget_charset (XCAR (car));
711 to = Fget_charset (XCAR (XCDR (car)));
712 if ( (XCHARSET_CHARS (from) != XCHARSET_CHARS (to)) ||
713 (XCHARSET_DIMENSION (from) != XCHARSET_DIMENSION (to)) )
714 signal_simple_error_2
715 ("Attempted conversion between different charset types",
717 spec.from_charset = from;
718 spec.to_charset = to;
720 Dynarr_add (store_here, spec);
724 /* Given a dynarr LOAD_HERE of internally-stored charset conversion
725 specs, return the equivalent as the Lisp programmer would see it.
727 If LOAD_HERE is 0, return Qnil. */
730 unparse_charset_conversion_specs (charset_conversion_spec_dynarr *load_here)
737 for (i = 0, result = Qnil; i < Dynarr_length (load_here); i++)
739 struct charset_conversion_spec *ccs = Dynarr_atp (load_here, i);
740 result = Fcons (list2 (ccs->from_charset, ccs->to_charset), result);
743 return Fnreverse (result);
748 DEFUN ("make-coding-system", Fmake_coding_system, 2, 4, 0, /*
749 Register symbol NAME as a coding system.
751 TYPE describes the conversion method used and should be one of
754 Automatic conversion. XEmacs attempts to detect the coding system
757 No conversion. Use this for binary files and such. On output,
758 graphic characters that are not in ASCII or Latin-1 will be
759 replaced by a ?. (For a no-conversion-encoded buffer, these
760 characters will only be present if you explicitly insert them.)
762 Shift-JIS (a Japanese encoding commonly used in PC operating systems).
764 ISO 10646 UCS-4 encoding.
766 ISO 10646 UTF-8 encoding.
768 Any ISO2022-compliant encoding. Among other things, this includes
769 JIS (the Japanese encoding commonly used for e-mail), EUC (the
770 standard Unix encoding for Japanese and other languages), and
771 Compound Text (the encoding used in X11). You can specify more
772 specific information about the conversion with the PROPS argument.
774 Big5 (the encoding commonly used for Taiwanese).
776 The conversion is performed using a user-written pseudo-code
777 program. CCL (Code Conversion Language) is the name of this
780 Write out or read in the raw contents of the memory representing
781 the buffer's text. This is primarily useful for debugging
782 purposes, and is only enabled when XEmacs has been compiled with
783 DEBUG_XEMACS defined (via the --debug configure option).
784 WARNING: Reading in a file using 'internal conversion can result
785 in an internal inconsistency in the memory representing a
786 buffer's text, which will produce unpredictable results and may
787 cause XEmacs to crash. Under normal circumstances you should
788 never use 'internal conversion.
790 DOC-STRING is a string describing the coding system.
792 PROPS is a property list, describing the specific nature of the
793 character set. Recognized properties are:
796 String to be displayed in the modeline when this coding system is
800 End-of-line conversion to be used. It should be one of
803 Automatically detect the end-of-line type (LF, CRLF,
804 or CR). Also generate subsidiary coding systems named
805 `NAME-unix', `NAME-dos', and `NAME-mac', that are
806 identical to this coding system but have an EOL-TYPE
807 value of 'lf, 'crlf, and 'cr, respectively.
809 The end of a line is marked externally using ASCII LF.
810 Since this is also the way that XEmacs represents an
811 end-of-line internally, specifying this option results
812 in no end-of-line conversion. This is the standard
813 format for Unix text files.
815 The end of a line is marked externally using ASCII
816 CRLF. This is the standard format for MS-DOS text
819 The end of a line is marked externally using ASCII CR.
820 This is the standard format for Macintosh text files.
822 Automatically detect the end-of-line type but do not
823 generate subsidiary coding systems. (This value is
824 converted to nil when stored internally, and
825 `coding-system-property' will return nil.)
828 If non-nil, composition/decomposition for combining characters
831 'use-entity-reference
832 If non-nil, SGML style entity-reference is used for non-system-characters.
834 'post-read-conversion
835 Function called after a file has been read in, to perform the
836 decoding. Called with two arguments, START and END, denoting
837 a region of the current buffer to be decoded.
839 'pre-write-conversion
840 Function called before a file is written out, to perform the
841 encoding. Called with two arguments, START and END, denoting
842 a region of the current buffer to be encoded.
845 The following additional properties are recognized if TYPE is 'iso2022:
851 The character set initially designated to the G0 - G3 registers.
852 The value should be one of
854 -- A charset object (designate that character set)
855 -- nil (do not ever use this register)
856 -- t (no character set is initially designated to
857 the register, but may be later on; this automatically
858 sets the corresponding `force-g*-on-output' property)
864 If non-nil, send an explicit designation sequence on output before
865 using the specified register.
868 If non-nil, use the short forms "ESC $ @", "ESC $ A", and
869 "ESC $ B" on output in place of the full designation sequences
870 "ESC $ ( @", "ESC $ ( A", and "ESC $ ( B".
873 If non-nil, don't designate ASCII to G0 at each end of line on output.
874 Setting this to non-nil also suppresses other state-resetting that
875 normally happens at the end of a line.
878 If non-nil, don't designate ASCII to G0 before control chars on output.
881 If non-nil, use 7-bit environment on output. Otherwise, use 8-bit
885 If non-nil, use locking-shift (SO/SI) instead of single-shift
886 or designation by escape sequence.
889 If non-nil, don't use ISO6429's direction specification.
892 If non-nil, literal control characters that are the same as
893 the beginning of a recognized ISO2022 or ISO6429 escape sequence
894 (in particular, ESC (0x1B), SO (0x0E), SI (0x0F), SS2 (0x8E),
895 SS3 (0x8F), and CSI (0x9B)) are "quoted" with an escape character
896 so that they can be properly distinguished from an escape sequence.
897 (Note that doing this results in a non-portable encoding.) This
898 encoding flag is used for byte-compiled files. Note that ESC
899 is a good choice for a quoting character because there are no
900 escape sequences whose second byte is a character from the Control-0
901 or Control-1 character sets; this is explicitly disallowed by the
904 'input-charset-conversion
905 A list of conversion specifications, specifying conversion of
906 characters in one charset to another when decoding is performed.
907 Each specification is a list of two elements: the source charset,
908 and the destination charset.
910 'output-charset-conversion
911 A list of conversion specifications, specifying conversion of
912 characters in one charset to another when encoding is performed.
913 The form of each specification is the same as for
914 'input-charset-conversion.
917 The following additional properties are recognized (and required)
921 CCL program used for decoding (converting to internal format).
924 CCL program used for encoding (converting to external format).
926 (name, type, doc_string, props))
928 Lisp_Coding_System *codesys;
929 enum coding_system_type ty;
930 int need_to_setup_eol_systems = 1;
932 /* Convert type to constant */
933 if (NILP (type) || EQ (type, Qundecided))
934 { ty = CODESYS_AUTODETECT; }
936 else if (EQ (type, Qshift_jis)) { ty = CODESYS_SHIFT_JIS; }
937 else if (EQ (type, Qiso2022)) { ty = CODESYS_ISO2022; }
938 else if (EQ (type, Qbig5)) { ty = CODESYS_BIG5; }
939 else if (EQ (type, Qucs4)) { ty = CODESYS_UCS4; }
940 else if (EQ (type, Qutf8)) { ty = CODESYS_UTF8; }
941 else if (EQ (type, Qccl)) { ty = CODESYS_CCL; }
943 else if (EQ (type, Qno_conversion)) { ty = CODESYS_NO_CONVERSION; }
945 else if (EQ (type, Qinternal)) { ty = CODESYS_INTERNAL; }
948 signal_simple_error ("Invalid coding system type", type);
952 codesys = allocate_coding_system (ty, name);
954 if (NILP (doc_string))
955 doc_string = build_string ("");
957 CHECK_STRING (doc_string);
958 CODING_SYSTEM_DOC_STRING (codesys) = doc_string;
961 EXTERNAL_PROPERTY_LIST_LOOP_3 (key, value, props)
963 if (EQ (key, Qmnemonic))
966 CHECK_STRING (value);
967 CODING_SYSTEM_MNEMONIC (codesys) = value;
970 else if (EQ (key, Qeol_type))
972 need_to_setup_eol_systems = NILP (value);
975 CODING_SYSTEM_EOL_TYPE (codesys) = symbol_to_eol_type (value);
978 else if (EQ (key, Qpost_read_conversion))
979 CODING_SYSTEM_POST_READ_CONVERSION (codesys) = value;
980 else if (EQ (key, Qpre_write_conversion))
981 CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = value;
983 else if (EQ (key, Qdisable_composition))
984 CODING_SYSTEM_DISABLE_COMPOSITION (codesys) = !NILP (value);
985 else if (EQ (key, Quse_entity_reference))
986 CODING_SYSTEM_USE_ENTITY_REFERENCE (codesys) = !NILP (value);
989 else if (ty == CODESYS_ISO2022)
991 #define FROB_INITIAL_CHARSET(charset_num) \
992 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, charset_num) = \
993 ((EQ (value, Qt) || EQ (value, Qnil)) ? value : Fget_charset (value))
995 if (EQ (key, Qcharset_g0)) FROB_INITIAL_CHARSET (0);
996 else if (EQ (key, Qcharset_g1)) FROB_INITIAL_CHARSET (1);
997 else if (EQ (key, Qcharset_g2)) FROB_INITIAL_CHARSET (2);
998 else if (EQ (key, Qcharset_g3)) FROB_INITIAL_CHARSET (3);
1000 #define FROB_FORCE_CHARSET(charset_num) \
1001 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (codesys, charset_num) = !NILP (value)
1003 else if (EQ (key, Qforce_g0_on_output)) FROB_FORCE_CHARSET (0);
1004 else if (EQ (key, Qforce_g1_on_output)) FROB_FORCE_CHARSET (1);
1005 else if (EQ (key, Qforce_g2_on_output)) FROB_FORCE_CHARSET (2);
1006 else if (EQ (key, Qforce_g3_on_output)) FROB_FORCE_CHARSET (3);
1008 #define FROB_BOOLEAN_PROPERTY(prop) \
1009 CODING_SYSTEM_ISO2022_##prop (codesys) = !NILP (value)
1011 else if (EQ (key, Qshort)) FROB_BOOLEAN_PROPERTY (SHORT);
1012 else if (EQ (key, Qno_ascii_eol)) FROB_BOOLEAN_PROPERTY (NO_ASCII_EOL);
1013 else if (EQ (key, Qno_ascii_cntl)) FROB_BOOLEAN_PROPERTY (NO_ASCII_CNTL);
1014 else if (EQ (key, Qseven)) FROB_BOOLEAN_PROPERTY (SEVEN);
1015 else if (EQ (key, Qlock_shift)) FROB_BOOLEAN_PROPERTY (LOCK_SHIFT);
1016 else if (EQ (key, Qno_iso6429)) FROB_BOOLEAN_PROPERTY (NO_ISO6429);
1017 else if (EQ (key, Qescape_quoted)) FROB_BOOLEAN_PROPERTY (ESCAPE_QUOTED);
1019 else if (EQ (key, Qinput_charset_conversion))
1021 codesys->iso2022.input_conv =
1022 Dynarr_new (charset_conversion_spec);
1023 parse_charset_conversion_specs (codesys->iso2022.input_conv,
1026 else if (EQ (key, Qoutput_charset_conversion))
1028 codesys->iso2022.output_conv =
1029 Dynarr_new (charset_conversion_spec);
1030 parse_charset_conversion_specs (codesys->iso2022.output_conv,
1034 signal_simple_error ("Unrecognized property", key);
1037 else if (ty == CODESYS_BIG5)
1039 if (EQ (key, Qcharset_g0)) FROB_INITIAL_CHARSET (0);
1040 else if (EQ (key, Qcharset_g1)) FROB_INITIAL_CHARSET (1);
1042 signal_simple_error ("Unrecognized property", key);
1045 else if (EQ (type, Qccl))
1048 struct ccl_program test_ccl;
1051 /* Check key first. */
1052 if (EQ (key, Qdecode))
1053 suffix = "-ccl-decode";
1054 else if (EQ (key, Qencode))
1055 suffix = "-ccl-encode";
1057 signal_simple_error ("Unrecognized property", key);
1059 /* If value is vector, register it as a ccl program
1060 associated with an newly created symbol for
1061 backward compatibility. */
1062 if (VECTORP (value))
1064 sym = Fintern (concat2 (Fsymbol_name (name),
1065 build_string (suffix)),
1067 Fregister_ccl_program (sym, value);
1071 CHECK_SYMBOL (value);
1074 /* check if the given ccl programs are valid. */
1075 if (setup_ccl_program (&test_ccl, sym) < 0)
1076 signal_simple_error ("Invalid CCL program", value);
1078 if (EQ (key, Qdecode))
1079 CODING_SYSTEM_CCL_DECODE (codesys) = sym;
1080 else if (EQ (key, Qencode))
1081 CODING_SYSTEM_CCL_ENCODE (codesys) = sym;
1086 signal_simple_error ("Unrecognized property", key);
1090 if (need_to_setup_eol_systems)
1091 setup_eol_coding_systems (codesys);
1094 Lisp_Object codesys_obj;
1095 XSETCODING_SYSTEM (codesys_obj, codesys);
1096 Fputhash (name, codesys_obj, Vcoding_system_hash_table);
1101 DEFUN ("copy-coding-system", Fcopy_coding_system, 2, 2, 0, /*
1102 Copy OLD-CODING-SYSTEM to NEW-NAME.
1103 If NEW-NAME does not name an existing coding system, a new one will
1106 (old_coding_system, new_name))
1108 Lisp_Object new_coding_system;
1109 old_coding_system = Fget_coding_system (old_coding_system);
1110 new_coding_system = Ffind_coding_system (new_name);
1111 if (NILP (new_coding_system))
1113 XSETCODING_SYSTEM (new_coding_system,
1114 allocate_coding_system
1115 (XCODING_SYSTEM_TYPE (old_coding_system),
1117 Fputhash (new_name, new_coding_system, Vcoding_system_hash_table);
1121 Lisp_Coding_System *to = XCODING_SYSTEM (new_coding_system);
1122 Lisp_Coding_System *from = XCODING_SYSTEM (old_coding_system);
1123 memcpy (((char *) to ) + sizeof (to->header),
1124 ((char *) from) + sizeof (from->header),
1125 sizeof (*from) - sizeof (from->header));
1126 to->name = new_name;
1128 return new_coding_system;
1131 DEFUN ("coding-system-canonical-name-p", Fcoding_system_canonical_name_p, 1, 1, 0, /*
1132 Return t if OBJECT names a coding system, and is not a coding system alias.
1136 return CODING_SYSTEMP (Fgethash (object, Vcoding_system_hash_table, Qnil))
1140 DEFUN ("coding-system-alias-p", Fcoding_system_alias_p, 1, 1, 0, /*
1141 Return t if OBJECT is a coding system alias.
1142 All coding system aliases are created by `define-coding-system-alias'.
1146 return SYMBOLP (Fgethash (object, Vcoding_system_hash_table, Qzero))
1150 DEFUN ("coding-system-aliasee", Fcoding_system_aliasee, 1, 1, 0, /*
1151 Return the coding-system symbol for which symbol ALIAS is an alias.
1155 Lisp_Object aliasee = Fgethash (alias, Vcoding_system_hash_table, Qnil);
1156 if (SYMBOLP (aliasee))
1159 signal_simple_error ("Symbol is not a coding system alias", alias);
1160 return Qnil; /* To keep the compiler happy */
1164 append_suffix_to_symbol (Lisp_Object symbol, const char *ascii_string)
1166 return Fintern (concat2 (Fsymbol_name (symbol), build_string (ascii_string)),
1170 /* A maphash function, for removing dangling coding system aliases. */
1172 dangling_coding_system_alias_p (Lisp_Object alias,
1173 Lisp_Object aliasee,
1174 void *dangling_aliases)
1176 if (SYMBOLP (aliasee)
1177 && NILP (Fgethash (aliasee, Vcoding_system_hash_table, Qnil)))
1179 (*(int *) dangling_aliases)++;
1186 DEFUN ("define-coding-system-alias", Fdefine_coding_system_alias, 2, 2, 0, /*
1187 Define symbol ALIAS as an alias for coding system ALIASEE.
1189 You can use this function to redefine an alias that has already been defined,
1190 but you cannot redefine a name which is the canonical name for a coding system.
1191 \(a canonical name of a coding system is what is returned when you call
1192 `coding-system-name' on a coding system).
1194 ALIASEE itself can be an alias, which allows you to define nested aliases.
1196 You are forbidden, however, from creating alias loops or `dangling' aliases.
1197 These will be detected, and an error will be signaled if you attempt to do so.
1199 If ALIASEE is nil, then ALIAS will simply be undefined.
1201 See also `coding-system-alias-p', `coding-system-aliasee',
1202 and `coding-system-canonical-name-p'.
1206 Lisp_Object real_coding_system, probe;
1208 CHECK_SYMBOL (alias);
1210 if (!NILP (Fcoding_system_canonical_name_p (alias)))
1212 ("Symbol is the canonical name of a coding system and cannot be redefined",
1217 Lisp_Object subsidiary_unix = append_suffix_to_symbol (alias, "-unix");
1218 Lisp_Object subsidiary_dos = append_suffix_to_symbol (alias, "-dos");
1219 Lisp_Object subsidiary_mac = append_suffix_to_symbol (alias, "-mac");
1221 Fremhash (alias, Vcoding_system_hash_table);
1223 /* Undefine subsidiary aliases,
1224 presumably created by a previous call to this function */
1225 if (! NILP (Fcoding_system_alias_p (subsidiary_unix)) &&
1226 ! NILP (Fcoding_system_alias_p (subsidiary_dos)) &&
1227 ! NILP (Fcoding_system_alias_p (subsidiary_mac)))
1229 Fdefine_coding_system_alias (subsidiary_unix, Qnil);
1230 Fdefine_coding_system_alias (subsidiary_dos, Qnil);
1231 Fdefine_coding_system_alias (subsidiary_mac, Qnil);
1234 /* Undefine dangling coding system aliases. */
1236 int dangling_aliases;
1239 dangling_aliases = 0;
1240 elisp_map_remhash (dangling_coding_system_alias_p,
1241 Vcoding_system_hash_table,
1243 } while (dangling_aliases > 0);
1249 if (CODING_SYSTEMP (aliasee))
1250 aliasee = XCODING_SYSTEM_NAME (aliasee);
1252 /* Checks that aliasee names a coding-system */
1253 real_coding_system = Fget_coding_system (aliasee);
1255 /* Check for coding system alias loops */
1256 if (EQ (alias, aliasee))
1257 alias_loop: signal_simple_error_2
1258 ("Attempt to create a coding system alias loop", alias, aliasee);
1260 for (probe = aliasee;
1262 probe = Fgethash (probe, Vcoding_system_hash_table, Qzero))
1264 if (EQ (probe, alias))
1268 Fputhash (alias, aliasee, Vcoding_system_hash_table);
1270 /* Set up aliases for subsidiaries.
1271 #### There must be a better way to handle subsidiary coding systems. */
1273 static const char *suffixes[] = { "-unix", "-dos", "-mac" };
1275 for (i = 0; i < countof (suffixes); i++)
1277 Lisp_Object alias_subsidiary =
1278 append_suffix_to_symbol (alias, suffixes[i]);
1279 Lisp_Object aliasee_subsidiary =
1280 append_suffix_to_symbol (aliasee, suffixes[i]);
1282 if (! NILP (Ffind_coding_system (aliasee_subsidiary)))
1283 Fdefine_coding_system_alias (alias_subsidiary, aliasee_subsidiary);
1286 /* FSF return value is a vector of [ALIAS-unix ALIAS-dos ALIAS-mac],
1287 but it doesn't look intentional, so I'd rather return something
1288 meaningful or nothing at all. */
1293 subsidiary_coding_system (Lisp_Object coding_system, eol_type_t type)
1295 Lisp_Coding_System *cs = XCODING_SYSTEM (coding_system);
1296 Lisp_Object new_coding_system;
1298 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
1299 return coding_system;
1303 case EOL_AUTODETECT: return coding_system;
1304 case EOL_LF: new_coding_system = CODING_SYSTEM_EOL_LF (cs); break;
1305 case EOL_CR: new_coding_system = CODING_SYSTEM_EOL_CR (cs); break;
1306 case EOL_CRLF: new_coding_system = CODING_SYSTEM_EOL_CRLF (cs); break;
1307 default: abort (); return Qnil;
1310 return NILP (new_coding_system) ? coding_system : new_coding_system;
1313 DEFUN ("subsidiary-coding-system", Fsubsidiary_coding_system, 2, 2, 0, /*
1314 Return the subsidiary coding system of CODING-SYSTEM with eol type EOL-TYPE.
1316 (coding_system, eol_type))
1318 coding_system = Fget_coding_system (coding_system);
1320 return subsidiary_coding_system (coding_system,
1321 symbol_to_eol_type (eol_type));
1325 /************************************************************************/
1326 /* Coding system accessors */
1327 /************************************************************************/
1329 DEFUN ("coding-system-doc-string", Fcoding_system_doc_string, 1, 1, 0, /*
1330 Return the doc string for CODING-SYSTEM.
1334 coding_system = Fget_coding_system (coding_system);
1335 return XCODING_SYSTEM_DOC_STRING (coding_system);
1338 DEFUN ("coding-system-type", Fcoding_system_type, 1, 1, 0, /*
1339 Return the type of CODING-SYSTEM.
1343 switch (XCODING_SYSTEM_TYPE (Fget_coding_system (coding_system)))
1346 case CODESYS_AUTODETECT: return Qundecided;
1348 case CODESYS_SHIFT_JIS: return Qshift_jis;
1349 case CODESYS_ISO2022: return Qiso2022;
1350 case CODESYS_BIG5: return Qbig5;
1351 case CODESYS_UCS4: return Qucs4;
1352 case CODESYS_UTF8: return Qutf8;
1353 case CODESYS_CCL: return Qccl;
1355 case CODESYS_NO_CONVERSION: return Qno_conversion;
1357 case CODESYS_INTERNAL: return Qinternal;
1364 Lisp_Object coding_system_charset (Lisp_Object coding_system, int gnum)
1367 = XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, gnum);
1369 return CHARSETP (cs) ? XCHARSET_NAME (cs) : Qnil;
1372 DEFUN ("coding-system-charset", Fcoding_system_charset, 2, 2, 0, /*
1373 Return initial charset of CODING-SYSTEM designated to GNUM.
1376 (coding_system, gnum))
1378 coding_system = Fget_coding_system (coding_system);
1381 return coding_system_charset (coding_system, XINT (gnum));
1385 DEFUN ("coding-system-property", Fcoding_system_property, 2, 2, 0, /*
1386 Return the PROP property of CODING-SYSTEM.
1388 (coding_system, prop))
1391 enum coding_system_type type;
1393 coding_system = Fget_coding_system (coding_system);
1394 CHECK_SYMBOL (prop);
1395 type = XCODING_SYSTEM_TYPE (coding_system);
1397 for (i = 0; !ok && i < Dynarr_length (the_codesys_prop_dynarr); i++)
1398 if (EQ (Dynarr_at (the_codesys_prop_dynarr, i).sym, prop))
1401 switch (Dynarr_at (the_codesys_prop_dynarr, i).prop_type)
1403 case CODESYS_PROP_ALL_OK:
1406 case CODESYS_PROP_ISO2022:
1407 if (type != CODESYS_ISO2022)
1409 ("Property only valid in ISO2022 coding systems",
1413 case CODESYS_PROP_CCL:
1414 if (type != CODESYS_CCL)
1416 ("Property only valid in CCL coding systems",
1426 signal_simple_error ("Unrecognized property", prop);
1428 if (EQ (prop, Qname))
1429 return XCODING_SYSTEM_NAME (coding_system);
1430 else if (EQ (prop, Qtype))
1431 return Fcoding_system_type (coding_system);
1432 else if (EQ (prop, Qdoc_string))
1433 return XCODING_SYSTEM_DOC_STRING (coding_system);
1434 else if (EQ (prop, Qmnemonic))
1435 return XCODING_SYSTEM_MNEMONIC (coding_system);
1436 else if (EQ (prop, Qeol_type))
1437 return eol_type_to_symbol (XCODING_SYSTEM_EOL_TYPE (coding_system));
1438 else if (EQ (prop, Qeol_lf))
1439 return XCODING_SYSTEM_EOL_LF (coding_system);
1440 else if (EQ (prop, Qeol_crlf))
1441 return XCODING_SYSTEM_EOL_CRLF (coding_system);
1442 else if (EQ (prop, Qeol_cr))
1443 return XCODING_SYSTEM_EOL_CR (coding_system);
1444 else if (EQ (prop, Qpost_read_conversion))
1445 return XCODING_SYSTEM_POST_READ_CONVERSION (coding_system);
1446 else if (EQ (prop, Qpre_write_conversion))
1447 return XCODING_SYSTEM_PRE_WRITE_CONVERSION (coding_system);
1450 else if (EQ (prop, Qdisable_composition))
1451 return XCODING_SYSTEM_DISABLE_COMPOSITION (coding_system) ? Qt : Qnil;
1452 else if (EQ (prop, Quse_entity_reference))
1453 return XCODING_SYSTEM_USE_ENTITY_REFERENCE (coding_system) ? Qt : Qnil;
1455 else if (type == CODESYS_ISO2022)
1457 if (EQ (prop, Qcharset_g0))
1458 return coding_system_charset (coding_system, 0);
1459 else if (EQ (prop, Qcharset_g1))
1460 return coding_system_charset (coding_system, 1);
1461 else if (EQ (prop, Qcharset_g2))
1462 return coding_system_charset (coding_system, 2);
1463 else if (EQ (prop, Qcharset_g3))
1464 return coding_system_charset (coding_system, 3);
1466 #define FORCE_CHARSET(charset_num) \
1467 (XCODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT \
1468 (coding_system, charset_num) ? Qt : Qnil)
1470 else if (EQ (prop, Qforce_g0_on_output)) return FORCE_CHARSET (0);
1471 else if (EQ (prop, Qforce_g1_on_output)) return FORCE_CHARSET (1);
1472 else if (EQ (prop, Qforce_g2_on_output)) return FORCE_CHARSET (2);
1473 else if (EQ (prop, Qforce_g3_on_output)) return FORCE_CHARSET (3);
1475 #define LISP_BOOLEAN(prop) \
1476 (XCODING_SYSTEM_ISO2022_##prop (coding_system) ? Qt : Qnil)
1478 else if (EQ (prop, Qshort)) return LISP_BOOLEAN (SHORT);
1479 else if (EQ (prop, Qno_ascii_eol)) return LISP_BOOLEAN (NO_ASCII_EOL);
1480 else if (EQ (prop, Qno_ascii_cntl)) return LISP_BOOLEAN (NO_ASCII_CNTL);
1481 else if (EQ (prop, Qseven)) return LISP_BOOLEAN (SEVEN);
1482 else if (EQ (prop, Qlock_shift)) return LISP_BOOLEAN (LOCK_SHIFT);
1483 else if (EQ (prop, Qno_iso6429)) return LISP_BOOLEAN (NO_ISO6429);
1484 else if (EQ (prop, Qescape_quoted)) return LISP_BOOLEAN (ESCAPE_QUOTED);
1486 else if (EQ (prop, Qinput_charset_conversion))
1488 unparse_charset_conversion_specs
1489 (XCODING_SYSTEM (coding_system)->iso2022.input_conv);
1490 else if (EQ (prop, Qoutput_charset_conversion))
1492 unparse_charset_conversion_specs
1493 (XCODING_SYSTEM (coding_system)->iso2022.output_conv);
1497 else if (type == CODESYS_CCL)
1499 if (EQ (prop, Qdecode))
1500 return XCODING_SYSTEM_CCL_DECODE (coding_system);
1501 else if (EQ (prop, Qencode))
1502 return XCODING_SYSTEM_CCL_ENCODE (coding_system);
1510 return Qnil; /* not reached */
1514 /************************************************************************/
1515 /* Coding category functions */
1516 /************************************************************************/
1519 decode_coding_category (Lisp_Object symbol)
1523 CHECK_SYMBOL (symbol);
1524 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1525 if (EQ (coding_category_symbol[i], symbol))
1528 signal_simple_error ("Unrecognized coding category", symbol);
1529 return 0; /* not reached */
1532 DEFUN ("coding-category-list", Fcoding_category_list, 0, 0, 0, /*
1533 Return a list of all recognized coding categories.
1538 Lisp_Object list = Qnil;
1540 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
1541 list = Fcons (coding_category_symbol[i], list);
1545 DEFUN ("set-coding-priority-list", Fset_coding_priority_list, 1, 1, 0, /*
1546 Change the priority order of the coding categories.
1547 LIST should be list of coding categories, in descending order of
1548 priority. Unspecified coding categories will be lower in priority
1549 than all specified ones, in the same relative order they were in
1554 int category_to_priority[CODING_CATEGORY_LAST];
1558 /* First generate a list that maps coding categories to priorities. */
1560 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1561 category_to_priority[i] = -1;
1563 /* Highest priority comes from the specified list. */
1565 EXTERNAL_LIST_LOOP (rest, list)
1567 int cat = decode_coding_category (XCAR (rest));
1569 if (category_to_priority[cat] >= 0)
1570 signal_simple_error ("Duplicate coding category in list", XCAR (rest));
1571 category_to_priority[cat] = i++;
1574 /* Now go through the existing categories by priority to retrieve
1575 the categories not yet specified and preserve their priority
1577 for (j = 0; j < CODING_CATEGORY_LAST; j++)
1579 int cat = fcd->coding_category_by_priority[j];
1580 if (category_to_priority[cat] < 0)
1581 category_to_priority[cat] = i++;
1584 /* Now we need to construct the inverse of the mapping we just
1587 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1588 fcd->coding_category_by_priority[category_to_priority[i]] = i;
1590 /* Phew! That was confusing. */
1594 DEFUN ("coding-priority-list", Fcoding_priority_list, 0, 0, 0, /*
1595 Return a list of coding categories in descending order of priority.
1600 Lisp_Object list = Qnil;
1602 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
1603 list = Fcons (coding_category_symbol[fcd->coding_category_by_priority[i]],
1608 DEFUN ("set-coding-category-system", Fset_coding_category_system, 2, 2, 0, /*
1609 Change the coding system associated with a coding category.
1611 (coding_category, coding_system))
1613 int cat = decode_coding_category (coding_category);
1615 coding_system = Fget_coding_system (coding_system);
1616 fcd->coding_category_system[cat] = coding_system;
1620 DEFUN ("coding-category-system", Fcoding_category_system, 1, 1, 0, /*
1621 Return the coding system associated with a coding category.
1625 int cat = decode_coding_category (coding_category);
1626 Lisp_Object sys = fcd->coding_category_system[cat];
1629 return XCODING_SYSTEM_NAME (sys);
1634 /************************************************************************/
1635 /* Detecting the encoding of data */
1636 /************************************************************************/
1638 struct detection_state
1640 eol_type_t eol_type;
1676 struct iso2022_decoder iso;
1678 int high_byte_count;
1679 unsigned int saw_single_shift:1;
1692 acceptable_control_char_p (int c)
1696 /* Allow and ignore control characters that you might
1697 reasonably see in a text file */
1702 case 8: /* backspace */
1703 case 11: /* vertical tab */
1704 case 12: /* form feed */
1705 case 26: /* MS-DOS C-z junk */
1706 case 31: /* '^_' -- for info */
1714 mask_has_at_most_one_bit_p (int mask)
1716 /* Perhaps the only thing useful you learn from intensive Microsoft
1717 technical interviews */
1718 return (mask & (mask - 1)) == 0;
1722 detect_eol_type (struct detection_state *st, const Extbyte *src,
1727 unsigned char c = *(unsigned char *)src++;
1730 if (st->eol.just_saw_cr)
1732 else if (st->eol.seen_anything)
1735 else if (st->eol.just_saw_cr)
1738 st->eol.just_saw_cr = 1;
1740 st->eol.just_saw_cr = 0;
1741 st->eol.seen_anything = 1;
1744 return EOL_AUTODETECT;
1747 /* Attempt to determine the encoding and EOL type of the given text.
1748 Before calling this function for the first type, you must initialize
1749 st->eol_type as appropriate and initialize st->mask to ~0.
1751 st->eol_type holds the determined EOL type, or EOL_AUTODETECT if
1754 st->mask holds the determined coding category mask, or ~0 if only
1755 ASCII has been seen so far.
1759 0 == st->eol_type is EOL_AUTODETECT and/or more than coding category
1760 is present in st->mask
1761 1 == definitive answers are here for both st->eol_type and st->mask
1765 detect_coding_type (struct detection_state *st, const Extbyte *src,
1766 size_t n, int just_do_eol)
1768 if (st->eol_type == EOL_AUTODETECT)
1769 st->eol_type = detect_eol_type (st, src, n);
1772 return st->eol_type != EOL_AUTODETECT;
1774 if (!st->seen_non_ascii)
1776 for (; n; n--, src++)
1778 unsigned char c = *(unsigned char *) src;
1779 if ((c < 0x20 && !acceptable_control_char_p (c)) || c >= 0x80)
1781 st->seen_non_ascii = 1;
1783 st->shift_jis.mask = ~0;
1787 st->iso2022.mask = ~0;
1797 if (!mask_has_at_most_one_bit_p (st->iso2022.mask))
1798 st->iso2022.mask = detect_coding_iso2022 (st, src, n);
1799 if (!mask_has_at_most_one_bit_p (st->shift_jis.mask))
1800 st->shift_jis.mask = detect_coding_sjis (st, src, n);
1801 if (!mask_has_at_most_one_bit_p (st->big5.mask))
1802 st->big5.mask = detect_coding_big5 (st, src, n);
1803 if (!mask_has_at_most_one_bit_p (st->utf8.mask))
1804 st->utf8.mask = detect_coding_utf8 (st, src, n);
1805 if (!mask_has_at_most_one_bit_p (st->ucs4.mask))
1806 st->ucs4.mask = detect_coding_ucs4 (st, src, n);
1809 = st->iso2022.mask | st->shift_jis.mask | st->big5.mask
1810 | st->utf8.mask | st->ucs4.mask;
1813 int retval = mask_has_at_most_one_bit_p (st->mask);
1814 st->mask |= CODING_CATEGORY_NO_CONVERSION_MASK;
1815 return retval && st->eol_type != EOL_AUTODETECT;
1820 coding_system_from_mask (int mask)
1824 /* If the file was entirely or basically ASCII, use the
1825 default value of `buffer-file-coding-system'. */
1826 Lisp_Object retval =
1827 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system;
1830 retval = Ffind_coding_system (retval);
1834 (Qbad_variable, Qwarning,
1835 "Invalid `default-buffer-file-coding-system', set to nil");
1836 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system = Qnil;
1840 retval = Fget_coding_system (Qraw_text);
1848 mask = postprocess_iso2022_mask (mask);
1850 /* Look through the coding categories by priority and find
1851 the first one that is allowed. */
1852 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1854 cat = fcd->coding_category_by_priority[i];
1855 if ((mask & (1 << cat)) &&
1856 !NILP (fcd->coding_category_system[cat]))
1860 return fcd->coding_category_system[cat];
1862 return Fget_coding_system (Qraw_text);
1866 /* Given a seekable read stream and potential coding system and EOL type
1867 as specified, do any autodetection that is called for. If the
1868 coding system and/or EOL type are not `autodetect', they will be left
1869 alone; but this function will never return an autodetect coding system
1872 This function does not automatically fetch subsidiary coding systems;
1873 that should be unnecessary with the explicit eol-type argument. */
1875 #define LENGTH(string_constant) (sizeof (string_constant) - 1)
1878 determine_real_coding_system (Lstream *stream, Lisp_Object *codesys_in_out,
1879 eol_type_t *eol_type_in_out)
1881 struct detection_state decst;
1883 if (*eol_type_in_out == EOL_AUTODETECT)
1884 *eol_type_in_out = XCODING_SYSTEM_EOL_TYPE (*codesys_in_out);
1887 decst.eol_type = *eol_type_in_out;
1890 /* If autodetection is called for, do it now. */
1891 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT
1892 || *eol_type_in_out == EOL_AUTODETECT)
1895 Lisp_Object coding_system = Qnil;
1897 ssize_t nread = Lstream_read (stream, buf, sizeof (buf));
1900 /* Look for initial "-*-"; mode line prefix */
1902 scan_end = buf + nread - LENGTH ("-*-coding:?-*-");
1907 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-')
1909 Extbyte *local_vars_beg = p + 3;
1910 /* Look for final "-*-"; mode line suffix */
1911 for (p = local_vars_beg,
1912 scan_end = buf + nread - LENGTH ("-*-");
1917 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-')
1919 Extbyte *suffix = p;
1920 /* Look for "coding:" */
1921 for (p = local_vars_beg,
1922 scan_end = suffix - LENGTH ("coding:?");
1925 if (memcmp ("coding:", p, LENGTH ("coding:")) == 0
1926 && (p == local_vars_beg
1927 || (*(p-1) == ' ' ||
1933 p += LENGTH ("coding:");
1934 while (*p == ' ' || *p == '\t') p++;
1936 /* Get coding system name */
1937 save = *suffix; *suffix = '\0';
1938 /* Characters valid in a MIME charset name (rfc 1521),
1939 and in a Lisp symbol name. */
1940 n = strspn ( (char *) p,
1941 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1942 "abcdefghijklmnopqrstuvwxyz"
1948 save = p[n]; p[n] = '\0';
1950 Ffind_coding_system (intern ((char *) p));
1960 if (NILP (coding_system))
1963 if (detect_coding_type (&decst, buf, nread,
1964 XCODING_SYSTEM_TYPE (*codesys_in_out)
1965 != CODESYS_AUTODETECT))
1967 nread = Lstream_read (stream, buf, sizeof (buf));
1973 else if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT
1974 && XCODING_SYSTEM_EOL_TYPE (coding_system) == EOL_AUTODETECT)
1977 if (detect_coding_type (&decst, buf, nread, 1))
1979 nread = Lstream_read (stream, buf, sizeof (buf));
1985 *eol_type_in_out = decst.eol_type;
1986 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT)
1988 if (NILP (coding_system))
1989 *codesys_in_out = coding_system_from_mask (decst.mask);
1991 *codesys_in_out = coding_system;
1995 /* If we absolutely can't determine the EOL type, just assume LF. */
1996 if (*eol_type_in_out == EOL_AUTODETECT)
1997 *eol_type_in_out = EOL_LF;
1999 Lstream_rewind (stream);
2002 DEFUN ("detect-coding-region", Fdetect_coding_region, 2, 3, 0, /*
2003 Detect coding system of the text in the region between START and END.
2004 Return a list of possible coding systems ordered by priority.
2005 If only ASCII characters are found, return 'undecided or one of
2006 its subsidiary coding systems according to a detected end-of-line
2007 type. Optional arg BUFFER defaults to the current buffer.
2009 (start, end, buffer))
2011 Lisp_Object val = Qnil;
2012 struct buffer *buf = decode_buffer (buffer, 0);
2014 Lisp_Object instream, lb_instream;
2015 Lstream *istr, *lb_istr;
2016 struct detection_state decst;
2017 struct gcpro gcpro1, gcpro2;
2019 get_buffer_range_char (buf, start, end, &b, &e, 0);
2020 lb_instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2021 lb_istr = XLSTREAM (lb_instream);
2022 instream = make_encoding_input_stream (lb_istr, Fget_coding_system (Qbinary));
2023 istr = XLSTREAM (instream);
2024 GCPRO2 (instream, lb_instream);
2026 decst.eol_type = EOL_AUTODETECT;
2030 Extbyte random_buffer[4096];
2031 ssize_t nread = Lstream_read (istr, random_buffer, sizeof (random_buffer));
2035 if (detect_coding_type (&decst, random_buffer, nread, 0))
2039 if (decst.mask == ~0)
2040 val = subsidiary_coding_system (Fget_coding_system (Qundecided),
2048 decst.mask = postprocess_iso2022_mask (decst.mask);
2050 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
2052 int sys = fcd->coding_category_by_priority[i];
2053 if (decst.mask & (1 << sys))
2055 Lisp_Object codesys = fcd->coding_category_system[sys];
2056 if (!NILP (codesys))
2057 codesys = subsidiary_coding_system (codesys, decst.eol_type);
2058 val = Fcons (codesys, val);
2062 Lstream_close (istr);
2064 Lstream_delete (istr);
2065 Lstream_delete (lb_istr);
2070 /************************************************************************/
2071 /* Converting to internal Mule format ("decoding") */
2072 /************************************************************************/
2074 /* A decoding stream is a stream used for decoding text (i.e.
2075 converting from some external format to internal format).
2076 The decoding-stream object keeps track of the actual coding
2077 stream, the stream that is at the other end, and data that
2078 needs to be persistent across the lifetime of the stream. */
2080 /* Handle the EOL stuff related to just-read-in character C.
2081 EOL_TYPE is the EOL type of the coding stream.
2082 FLAGS is the current value of FLAGS in the coding stream, and may
2083 be modified by this macro. (The macro only looks at the
2084 CODING_STATE_CR flag.) DST is the Dynarr to which the decoded
2085 bytes are to be written. You need to also define a local goto
2086 label "label_continue_loop" that is at the end of the main
2087 character-reading loop.
2089 If C is a CR character, then this macro handles it entirely and
2090 jumps to label_continue_loop. Otherwise, this macro does not add
2091 anything to DST, and continues normally. You should continue
2092 processing C normally after this macro. */
2094 #define DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst) \
2098 if (eol_type == EOL_CR) \
2099 Dynarr_add (dst, '\n'); \
2100 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR) \
2101 Dynarr_add (dst, c); \
2103 flags |= CODING_STATE_CR; \
2104 goto label_continue_loop; \
2106 else if (flags & CODING_STATE_CR) \
2107 { /* eol_type == CODING_SYSTEM_EOL_CRLF */ \
2109 Dynarr_add (dst, '\r'); \
2110 flags &= ~CODING_STATE_CR; \
2114 /* C should be a binary character in the range 0 - 255; convert
2115 to internal format and add to Dynarr DST. */
2118 #define DECODE_ADD_BINARY_CHAR(c, dst) \
2120 if (BYTE_ASCII_P (c)) \
2121 Dynarr_add (dst, c); \
2124 Dynarr_add (dst, (c >> 6) | 0xc0); \
2125 Dynarr_add (dst, (c & 0x3f) | 0x80); \
2129 INLINE_HEADER void DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst);
2131 DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst)
2135 Dynarr_add (dst, c);
2137 else if ( c <= 0x7ff )
2139 Dynarr_add (dst, (c >> 6) | 0xc0);
2140 Dynarr_add (dst, (c & 0x3f) | 0x80);
2142 else if ( c <= 0xffff )
2144 Dynarr_add (dst, (c >> 12) | 0xe0);
2145 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2146 Dynarr_add (dst, (c & 0x3f) | 0x80);
2148 else if ( c <= 0x1fffff )
2150 Dynarr_add (dst, (c >> 18) | 0xf0);
2151 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
2152 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2153 Dynarr_add (dst, (c & 0x3f) | 0x80);
2155 else if ( c <= 0x3ffffff )
2157 Dynarr_add (dst, (c >> 24) | 0xf8);
2158 Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
2159 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
2160 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2161 Dynarr_add (dst, (c & 0x3f) | 0x80);
2165 Dynarr_add (dst, (c >> 30) | 0xfc);
2166 Dynarr_add (dst, ((c >> 24) & 0x3f) | 0x80);
2167 Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
2168 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
2169 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2170 Dynarr_add (dst, (c & 0x3f) | 0x80);
2174 #define DECODE_ADD_BINARY_CHAR(c, dst) \
2176 if (BYTE_ASCII_P (c)) \
2177 Dynarr_add (dst, c); \
2178 else if (BYTE_C1_P (c)) \
2180 Dynarr_add (dst, LEADING_BYTE_CONTROL_1); \
2181 Dynarr_add (dst, c + 0x20); \
2185 Dynarr_add (dst, LEADING_BYTE_LATIN_ISO8859_1); \
2186 Dynarr_add (dst, c); \
2191 #define DECODE_OUTPUT_PARTIAL_CHAR(ch) \
2195 DECODE_ADD_BINARY_CHAR (ch, dst); \
2200 #define DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst) \
2202 if (flags & CODING_STATE_END) \
2204 DECODE_OUTPUT_PARTIAL_CHAR (ch); \
2205 if (flags & CODING_STATE_CR) \
2206 Dynarr_add (dst, '\r'); \
2210 #define DECODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, decoding)
2212 struct decoding_stream
2214 /* Coding system that governs the conversion. */
2215 Lisp_Coding_System *codesys;
2217 /* Stream that we read the encoded data from or
2218 write the decoded data to. */
2221 /* If we are reading, then we can return only a fixed amount of
2222 data, so if the conversion resulted in too much data, we store it
2223 here for retrieval the next time around. */
2224 unsigned_char_dynarr *runoff;
2226 /* FLAGS holds flags indicating the current state of the decoding.
2227 Some of these flags are dependent on the coding system. */
2230 /* CPOS holds a partially built-up code-point of character. */
2233 /* EOL_TYPE specifies the type of end-of-line conversion that
2234 currently applies. We need to keep this separate from the
2235 EOL type stored in CODESYS because the latter might indicate
2236 automatic EOL-type detection while the former will always
2237 indicate a particular EOL type. */
2238 eol_type_t eol_type;
2240 /* Additional ISO2022 information. We define the structure above
2241 because it's also needed by the detection routines. */
2242 struct iso2022_decoder iso2022;
2244 /* Additional information (the state of the running CCL program)
2245 used by the CCL decoder. */
2246 struct ccl_program ccl;
2248 /* counter for UTF-8 or UCS-4 */
2249 unsigned char counter;
2252 unsigned char er_counter;
2253 unsigned char er_buf[16];
2255 unsigned combined_char_count;
2256 Emchar combined_chars[16];
2257 Lisp_Object combining_table;
2259 struct detection_state decst;
2263 extern Lisp_Object Qcomposition;
2266 decode_flush_er_chars (struct decoding_stream *str, unsigned_char_dynarr* dst);
2268 decode_flush_er_chars (struct decoding_stream *str, unsigned_char_dynarr* dst)
2270 if ( str->er_counter > 0)
2272 Dynarr_add_many (dst, str->er_buf, str->er_counter);
2273 str->er_counter = 0;
2277 void decode_add_er_char (struct decoding_stream *str, Emchar character,
2278 unsigned_char_dynarr* dst);
2280 decode_add_er_char (struct decoding_stream *str, Emchar c,
2281 unsigned_char_dynarr* dst)
2283 if (str->er_counter == 0)
2285 if (CODING_SYSTEM_USE_ENTITY_REFERENCE (str->codesys)
2288 str->er_buf[0] = '&';
2292 DECODE_ADD_UCS_CHAR (c, dst);
2296 Lisp_Object string = make_string (str->er_buf,
2298 Lisp_Object rest = Vcoded_charset_entity_reference_alist;
2305 while (!NILP (rest))
2309 if (NILP (ccs = Ffind_charset (ccs)))
2324 pat = concat3 (build_string ("^&"),
2325 pat, build_string ("\\([0-9]+\\)$"));
2328 else if (EQ (ret, Qx))
2330 pat = concat3 (build_string ("^&"),
2331 pat, build_string ("\\([0-9a-f]+\\)$"));
2334 else if (EQ (ret, QX))
2336 pat = concat3 (build_string ("^&"),
2337 pat, build_string ("\\([0-9A-F]+\\)$"));
2343 if (!NILP (Fstring_match (pat, string, Qnil, Qnil)))
2346 = XINT (Fstring_to_number
2347 (Fsubstring (string,
2348 Fmatch_beginning (make_int (1)),
2349 Fmatch_end (make_int (1))),
2352 DECODE_ADD_UCS_CHAR (DECODE_CHAR (ccs, code), dst);
2357 if (!NILP (Fstring_match (build_string ("^&MCS-\\([0-9A-F]+\\)$"),
2358 string, Qnil, Qnil)))
2361 = XINT (Fstring_to_number
2362 (Fsubstring (string,
2363 Fmatch_beginning (make_int (1)),
2364 Fmatch_end (make_int (1))),
2367 DECODE_ADD_UCS_CHAR (code, dst);
2371 Dynarr_add_many (dst, str->er_buf, str->er_counter);
2372 Dynarr_add (dst, ';');
2375 str->er_counter = 0;
2377 else if ( (str->er_counter >= 16) || (c >= 0x7F) )
2379 Dynarr_add_many (dst, str->er_buf, str->er_counter);
2380 str->er_counter = 0;
2381 DECODE_ADD_UCS_CHAR (c, dst);
2384 str->er_buf[str->er_counter++] = c;
2388 COMPOSE_FLUSH_CHARS (struct decoding_stream *str, unsigned_char_dynarr* dst);
2390 COMPOSE_FLUSH_CHARS (struct decoding_stream *str, unsigned_char_dynarr* dst)
2394 for (i = 0; i < str->combined_char_count; i++)
2395 decode_add_er_char (str, str->combined_chars[i], dst);
2396 str->combined_char_count = 0;
2397 str->combining_table = Qnil;
2400 void COMPOSE_ADD_CHAR (struct decoding_stream *str, Emchar character,
2401 unsigned_char_dynarr* dst);
2403 COMPOSE_ADD_CHAR (struct decoding_stream *str,
2404 Emchar character, unsigned_char_dynarr* dst)
2406 if (CODING_SYSTEM_DISABLE_COMPOSITION (str->codesys))
2407 decode_add_er_char (str, character, dst);
2408 else if (!CONSP (str->combining_table))
2411 = Fget_char_attribute (make_char (character), Qcomposition, Qnil);
2414 decode_add_er_char (str, character, dst);
2417 str->combined_chars[0] = character;
2418 str->combined_char_count = 1;
2419 str->combining_table = ret;
2424 Lisp_Object ret = Fcdr (Fassq (make_char (character), str->combining_table));
2428 Emchar char2 = XCHARVAL (ret);
2429 ret = Fget_char_attribute (make_char (character), Qcomposition, Qnil);
2432 decode_add_er_char (str, character, dst);
2433 str->combined_char_count = 0;
2434 str->combining_table = Qnil;
2438 str->combined_chars[0] = char2;
2439 str->combined_char_count = 1;
2440 str->combining_table = ret;
2445 COMPOSE_FLUSH_CHARS (str, dst);
2446 decode_add_er_char (str, character, dst);
2450 #else /* not UTF2000 */
2451 #define COMPOSE_FLUSH_CHARS(str, dst)
2452 #define COMPOSE_ADD_CHAR(str, ch, dst) DECODE_ADD_UCS_CHAR (ch, dst)
2453 #endif /* UTF2000 */
2455 static ssize_t decoding_reader (Lstream *stream,
2456 unsigned char *data, size_t size);
2457 static ssize_t decoding_writer (Lstream *stream,
2458 const unsigned char *data, size_t size);
2459 static int decoding_rewinder (Lstream *stream);
2460 static int decoding_seekable_p (Lstream *stream);
2461 static int decoding_flusher (Lstream *stream);
2462 static int decoding_closer (Lstream *stream);
2464 static Lisp_Object decoding_marker (Lisp_Object stream);
2466 DEFINE_LSTREAM_IMPLEMENTATION ("decoding", lstream_decoding,
2467 sizeof (struct decoding_stream));
2470 decoding_marker (Lisp_Object stream)
2472 Lstream *str = DECODING_STREAM_DATA (XLSTREAM (stream))->other_end;
2473 Lisp_Object str_obj;
2475 /* We do not need to mark the coding systems or charsets stored
2476 within the stream because they are stored in a global list
2477 and automatically marked. */
2479 XSETLSTREAM (str_obj, str);
2480 mark_object (str_obj);
2481 if (str->imp->marker)
2482 return (str->imp->marker) (str_obj);
2487 /* Read SIZE bytes of data and store it into DATA. We are a decoding stream
2488 so we read data from the other end, decode it, and store it into DATA. */
2491 decoding_reader (Lstream *stream, unsigned char *data, size_t size)
2493 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2494 unsigned char *orig_data = data;
2496 int error_occurred = 0;
2498 /* We need to interface to mule_decode(), which expects to take some
2499 amount of data and store the result into a Dynarr. We have
2500 mule_decode() store into str->runoff, and take data from there
2503 /* We loop until we have enough data, reading chunks from the other
2504 end and decoding it. */
2507 /* Take data from the runoff if we can. Make sure to take at
2508 most SIZE bytes, and delete the data from the runoff. */
2509 if (Dynarr_length (str->runoff) > 0)
2511 size_t chunk = min (size, (size_t) Dynarr_length (str->runoff));
2512 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2513 Dynarr_delete_many (str->runoff, 0, chunk);
2519 break; /* No more room for data */
2521 if (str->flags & CODING_STATE_END)
2522 /* This means that on the previous iteration, we hit the EOF on
2523 the other end. We loop once more so that mule_decode() can
2524 output any final stuff it may be holding, or any "go back
2525 to a sane state" escape sequences. (This latter makes sense
2526 during encoding.) */
2529 /* Exhausted the runoff, so get some more. DATA has at least
2530 SIZE bytes left of storage in it, so it's OK to read directly
2531 into it. (We'll be overwriting above, after we've decoded it
2532 into the runoff.) */
2533 read_size = Lstream_read (str->other_end, data, size);
2540 /* There might be some more end data produced in the translation.
2541 See the comment above. */
2542 str->flags |= CODING_STATE_END;
2543 mule_decode (stream, (Extbyte *) data, str->runoff, read_size);
2546 if (data - orig_data == 0)
2547 return error_occurred ? -1 : 0;
2549 return data - orig_data;
2553 decoding_writer (Lstream *stream, const unsigned char *data, size_t size)
2555 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2558 /* Decode all our data into the runoff, and then attempt to write
2559 it all out to the other end. Remove whatever chunk we succeeded
2561 mule_decode (stream, (Extbyte *) data, str->runoff, size);
2562 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
2563 Dynarr_length (str->runoff));
2565 Dynarr_delete_many (str->runoff, 0, retval);
2566 /* Do NOT return retval. The return value indicates how much
2567 of the incoming data was written, not how many bytes were
2573 reset_decoding_stream (struct decoding_stream *str)
2576 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_ISO2022)
2578 Lisp_Object coding_system;
2579 XSETCODING_SYSTEM (coding_system, str->codesys);
2580 reset_iso2022 (coding_system, &str->iso2022);
2582 else if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_CCL)
2584 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_DECODE (str->codesys));
2589 str->er_counter = 0;
2590 str->combined_char_count = 0;
2591 str->combining_table = Qnil;
2593 str->flags = str->cpos = 0;
2597 decoding_rewinder (Lstream *stream)
2599 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2600 reset_decoding_stream (str);
2601 Dynarr_reset (str->runoff);
2602 return Lstream_rewind (str->other_end);
2606 decoding_seekable_p (Lstream *stream)
2608 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2609 return Lstream_seekable_p (str->other_end);
2613 decoding_flusher (Lstream *stream)
2615 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2616 return Lstream_flush (str->other_end);
2620 decoding_closer (Lstream *stream)
2622 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2623 if (stream->flags & LSTREAM_FL_WRITE)
2625 str->flags |= CODING_STATE_END;
2626 decoding_writer (stream, 0, 0);
2628 Dynarr_free (str->runoff);
2630 #ifdef ENABLE_COMPOSITE_CHARS
2631 if (str->iso2022.composite_chars)
2632 Dynarr_free (str->iso2022.composite_chars);
2635 return Lstream_close (str->other_end);
2639 decoding_stream_coding_system (Lstream *stream)
2641 Lisp_Object coding_system;
2642 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2644 XSETCODING_SYSTEM (coding_system, str->codesys);
2645 return subsidiary_coding_system (coding_system, str->eol_type);
2649 set_decoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
2651 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
2652 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2654 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
2655 str->eol_type = CODING_SYSTEM_EOL_TYPE (cs);
2656 reset_decoding_stream (str);
2659 /* WARNING WARNING WARNING WARNING!!!!! If you open up a decoding
2660 stream for writing, no automatic code detection will be performed.
2661 The reason for this is that automatic code detection requires a
2662 seekable input. Things will also fail if you open a decoding
2663 stream for reading using a non-fully-specified coding system and
2664 a non-seekable input stream. */
2667 make_decoding_stream_1 (Lstream *stream, Lisp_Object codesys,
2670 Lstream *lstr = Lstream_new (lstream_decoding, mode);
2671 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2675 str->other_end = stream;
2676 str->runoff = (unsigned_char_dynarr *) Dynarr_new (unsigned_char);
2677 str->eol_type = EOL_AUTODETECT;
2678 if (!strcmp (mode, "r")
2679 && Lstream_seekable_p (stream))
2680 /* We can determine the coding system now. */
2681 determine_real_coding_system (stream, &codesys, &str->eol_type);
2682 set_decoding_stream_coding_system (lstr, codesys);
2683 str->decst.eol_type = str->eol_type;
2684 str->decst.mask = ~0;
2685 XSETLSTREAM (obj, lstr);
2690 make_decoding_input_stream (Lstream *stream, Lisp_Object codesys)
2692 return make_decoding_stream_1 (stream, codesys, "r");
2696 make_decoding_output_stream (Lstream *stream, Lisp_Object codesys)
2698 return make_decoding_stream_1 (stream, codesys, "w");
2701 /* Note: the decode_coding_* functions all take the same
2702 arguments as mule_decode(), which is to say some SRC data of
2703 size N, which is to be stored into dynamic array DST.
2704 DECODING is the stream within which the decoding is
2705 taking place, but no data is actually read from or
2706 written to that stream; that is handled in decoding_reader()
2707 or decoding_writer(). This allows the same functions to
2708 be used for both reading and writing. */
2711 mule_decode (Lstream *decoding, const Extbyte *src,
2712 unsigned_char_dynarr *dst, size_t n)
2714 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
2716 /* If necessary, do encoding-detection now. We do this when
2717 we're a writing stream or a non-seekable reading stream,
2718 meaning that we can't just process the whole input,
2719 rewind, and start over. */
2721 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT ||
2722 str->eol_type == EOL_AUTODETECT)
2724 Lisp_Object codesys;
2726 XSETCODING_SYSTEM (codesys, str->codesys);
2727 detect_coding_type (&str->decst, src, n,
2728 CODING_SYSTEM_TYPE (str->codesys) !=
2729 CODESYS_AUTODETECT);
2730 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT &&
2731 str->decst.mask != ~0)
2732 /* #### This is cheesy. What we really ought to do is
2733 buffer up a certain amount of data so as to get a
2734 less random result. */
2735 codesys = coding_system_from_mask (str->decst.mask);
2736 str->eol_type = str->decst.eol_type;
2737 if (XCODING_SYSTEM (codesys) != str->codesys)
2739 /* Preserve the CODING_STATE_END flag in case it was set.
2740 If we erase it, bad things might happen. */
2741 int was_end = str->flags & CODING_STATE_END;
2742 set_decoding_stream_coding_system (decoding, codesys);
2744 str->flags |= CODING_STATE_END;
2748 switch (CODING_SYSTEM_TYPE (str->codesys))
2751 case CODESYS_INTERNAL:
2752 Dynarr_add_many (dst, src, n);
2755 case CODESYS_AUTODETECT:
2756 /* If we got this far and still haven't decided on the coding
2757 system, then do no conversion. */
2758 case CODESYS_NO_CONVERSION:
2759 decode_coding_no_conversion (decoding, src, dst, n);
2762 case CODESYS_SHIFT_JIS:
2763 decode_coding_sjis (decoding, src, dst, n);
2766 decode_coding_big5 (decoding, src, dst, n);
2769 decode_coding_ucs4 (decoding, src, dst, n);
2772 decode_coding_utf8 (decoding, src, dst, n);
2775 str->ccl.last_block = str->flags & CODING_STATE_END;
2776 /* When applying ccl program to stream, MUST NOT set NULL
2778 ccl_driver (&str->ccl, (src ? (unsigned char *)src : (unsigned char*)""),
2779 dst, n, 0, CCL_MODE_DECODING);
2781 case CODESYS_ISO2022:
2782 decode_coding_iso2022 (decoding, src, dst, n);
2790 DEFUN ("decode-coding-region", Fdecode_coding_region, 3, 4, 0, /*
2791 Decode the text between START and END which is encoded in CODING-SYSTEM.
2792 This is useful if you've read in encoded text from a file without decoding
2793 it (e.g. you read in a JIS-formatted file but used the `binary' or
2794 `no-conversion' coding system, so that it shows up as "^[$B!<!+^[(B").
2795 Return length of decoded text.
2796 BUFFER defaults to the current buffer if unspecified.
2798 (start, end, coding_system, buffer))
2801 struct buffer *buf = decode_buffer (buffer, 0);
2802 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2803 Lstream *istr, *ostr;
2804 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2806 get_buffer_range_char (buf, start, end, &b, &e, 0);
2808 barf_if_buffer_read_only (buf, b, e);
2810 coding_system = Fget_coding_system (coding_system);
2811 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2812 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
2813 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
2815 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
2816 Fget_coding_system (Qbinary));
2817 istr = XLSTREAM (instream);
2818 ostr = XLSTREAM (outstream);
2819 GCPRO4 (instream, lb_outstream, de_outstream, outstream);
2821 /* The chain of streams looks like this:
2823 [BUFFER] <----- send through
2824 ------> [ENCODE AS BINARY]
2825 ------> [DECODE AS SPECIFIED]
2831 char tempbuf[1024]; /* some random amount */
2832 Bufpos newpos, even_newer_pos;
2833 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
2834 ssize_t size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
2838 newpos = lisp_buffer_stream_startpos (istr);
2839 Lstream_write (ostr, tempbuf, size_in_bytes);
2840 even_newer_pos = lisp_buffer_stream_startpos (istr);
2841 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
2844 Lstream_close (istr);
2845 Lstream_close (ostr);
2847 Lstream_delete (istr);
2848 Lstream_delete (ostr);
2849 Lstream_delete (XLSTREAM (de_outstream));
2850 Lstream_delete (XLSTREAM (lb_outstream));
2855 /************************************************************************/
2856 /* Converting to an external encoding ("encoding") */
2857 /************************************************************************/
2859 /* An encoding stream is an output stream. When you create the
2860 stream, you specify the coding system that governs the encoding
2861 and another stream that the resulting encoded data is to be
2862 sent to, and then start sending data to it. */
2864 #define ENCODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, encoding)
2866 struct encoding_stream
2868 /* Coding system that governs the conversion. */
2869 Lisp_Coding_System *codesys;
2871 /* Stream that we read the encoded data from or
2872 write the decoded data to. */
2875 /* If we are reading, then we can return only a fixed amount of
2876 data, so if the conversion resulted in too much data, we store it
2877 here for retrieval the next time around. */
2878 unsigned_char_dynarr *runoff;
2880 /* FLAGS holds flags indicating the current state of the encoding.
2881 Some of these flags are dependent on the coding system. */
2884 /* CH holds a partially built-up character. Since we only deal
2885 with one- and two-byte characters at the moment, we only use
2886 this to store the first byte of a two-byte character. */
2889 /* Additional information used by the ISO2022 encoder. */
2892 /* CHARSET holds the character sets currently assigned to the G0
2893 through G3 registers. It is initialized from the array
2894 INITIAL_CHARSET in CODESYS. */
2895 Lisp_Object charset[4];
2897 /* Which registers are currently invoked into the left (GL) and
2898 right (GR) halves of the 8-bit encoding space? */
2899 int register_left, register_right;
2901 /* Whether we need to explicitly designate the charset in the
2902 G? register before using it. It is initialized from the
2903 array FORCE_CHARSET_ON_OUTPUT in CODESYS. */
2904 unsigned char force_charset_on_output[4];
2906 /* Other state variables that need to be preserved across
2908 Lisp_Object current_charset;
2910 int current_char_boundary;
2913 void (*encode_char) (struct encoding_stream *str, Emchar c,
2914 unsigned_char_dynarr *dst, unsigned int *flags);
2915 void (*finish) (struct encoding_stream *str,
2916 unsigned_char_dynarr *dst, unsigned int *flags);
2918 /* Additional information (the state of the running CCL program)
2919 used by the CCL encoder. */
2920 struct ccl_program ccl;
2924 static ssize_t encoding_reader (Lstream *stream, unsigned char *data, size_t size);
2925 static ssize_t encoding_writer (Lstream *stream, const unsigned char *data,
2927 static int encoding_rewinder (Lstream *stream);
2928 static int encoding_seekable_p (Lstream *stream);
2929 static int encoding_flusher (Lstream *stream);
2930 static int encoding_closer (Lstream *stream);
2932 static Lisp_Object encoding_marker (Lisp_Object stream);
2934 DEFINE_LSTREAM_IMPLEMENTATION ("encoding", lstream_encoding,
2935 sizeof (struct encoding_stream));
2938 encoding_marker (Lisp_Object stream)
2940 Lstream *str = ENCODING_STREAM_DATA (XLSTREAM (stream))->other_end;
2941 Lisp_Object str_obj;
2943 /* We do not need to mark the coding systems or charsets stored
2944 within the stream because they are stored in a global list
2945 and automatically marked. */
2947 XSETLSTREAM (str_obj, str);
2948 mark_object (str_obj);
2949 if (str->imp->marker)
2950 return (str->imp->marker) (str_obj);
2955 /* Read SIZE bytes of data and store it into DATA. We are a encoding stream
2956 so we read data from the other end, encode it, and store it into DATA. */
2959 encoding_reader (Lstream *stream, unsigned char *data, size_t size)
2961 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2962 unsigned char *orig_data = data;
2964 int error_occurred = 0;
2966 /* We need to interface to mule_encode(), which expects to take some
2967 amount of data and store the result into a Dynarr. We have
2968 mule_encode() store into str->runoff, and take data from there
2971 /* We loop until we have enough data, reading chunks from the other
2972 end and encoding it. */
2975 /* Take data from the runoff if we can. Make sure to take at
2976 most SIZE bytes, and delete the data from the runoff. */
2977 if (Dynarr_length (str->runoff) > 0)
2979 int chunk = min ((int) size, Dynarr_length (str->runoff));
2980 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2981 Dynarr_delete_many (str->runoff, 0, chunk);
2987 break; /* No more room for data */
2989 if (str->flags & CODING_STATE_END)
2990 /* This means that on the previous iteration, we hit the EOF on
2991 the other end. We loop once more so that mule_encode() can
2992 output any final stuff it may be holding, or any "go back
2993 to a sane state" escape sequences. (This latter makes sense
2994 during encoding.) */
2997 /* Exhausted the runoff, so get some more. DATA at least SIZE bytes
2998 left of storage in it, so it's OK to read directly into it.
2999 (We'll be overwriting above, after we've encoded it into the
3001 read_size = Lstream_read (str->other_end, data, size);
3008 /* There might be some more end data produced in the translation.
3009 See the comment above. */
3010 str->flags |= CODING_STATE_END;
3011 mule_encode (stream, data, str->runoff, read_size);
3014 if (data == orig_data)
3015 return error_occurred ? -1 : 0;
3017 return data - orig_data;
3021 encoding_writer (Lstream *stream, const unsigned char *data, size_t size)
3023 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3026 /* Encode all our data into the runoff, and then attempt to write
3027 it all out to the other end. Remove whatever chunk we succeeded
3029 mule_encode (stream, data, str->runoff, size);
3030 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
3031 Dynarr_length (str->runoff));
3033 Dynarr_delete_many (str->runoff, 0, retval);
3034 /* Do NOT return retval. The return value indicates how much
3035 of the incoming data was written, not how many bytes were
3041 reset_encoding_stream (struct encoding_stream *str)
3044 switch (CODING_SYSTEM_TYPE (str->codesys))
3046 case CODESYS_ISO2022:
3050 str->encode_char = &char_encode_iso2022;
3051 str->finish = &char_finish_iso2022;
3052 for (i = 0; i < 4; i++)
3054 str->iso2022.charset[i] =
3055 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, i);
3056 str->iso2022.force_charset_on_output[i] =
3057 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (str->codesys, i);
3059 str->iso2022.register_left = 0;
3060 str->iso2022.register_right = 1;
3061 str->iso2022.current_charset = Qnil;
3062 str->iso2022.current_half = 0;
3066 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_ENCODE (str->codesys));
3069 str->encode_char = &char_encode_utf8;
3070 str->finish = &char_finish_utf8;
3073 str->encode_char = &char_encode_ucs4;
3074 str->finish = &char_finish_ucs4;
3076 case CODESYS_SHIFT_JIS:
3077 str->encode_char = &char_encode_shift_jis;
3078 str->finish = &char_finish_shift_jis;
3081 str->encode_char = &char_encode_big5;
3082 str->finish = &char_finish_big5;
3088 str->iso2022.current_char_boundary = 0;
3089 str->flags = str->ch = 0;
3093 encoding_rewinder (Lstream *stream)
3095 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3096 reset_encoding_stream (str);
3097 Dynarr_reset (str->runoff);
3098 return Lstream_rewind (str->other_end);
3102 encoding_seekable_p (Lstream *stream)
3104 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3105 return Lstream_seekable_p (str->other_end);
3109 encoding_flusher (Lstream *stream)
3111 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3112 return Lstream_flush (str->other_end);
3116 encoding_closer (Lstream *stream)
3118 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3119 if (stream->flags & LSTREAM_FL_WRITE)
3121 str->flags |= CODING_STATE_END;
3122 encoding_writer (stream, 0, 0);
3124 Dynarr_free (str->runoff);
3125 return Lstream_close (str->other_end);
3129 encoding_stream_coding_system (Lstream *stream)
3131 Lisp_Object coding_system;
3132 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3134 XSETCODING_SYSTEM (coding_system, str->codesys);
3135 return coding_system;
3139 set_encoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
3141 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
3142 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
3144 reset_encoding_stream (str);
3148 make_encoding_stream_1 (Lstream *stream, Lisp_Object codesys,
3151 Lstream *lstr = Lstream_new (lstream_encoding, mode);
3152 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
3156 str->runoff = Dynarr_new (unsigned_char);
3157 str->other_end = stream;
3158 set_encoding_stream_coding_system (lstr, codesys);
3159 XSETLSTREAM (obj, lstr);
3164 make_encoding_input_stream (Lstream *stream, Lisp_Object codesys)
3166 return make_encoding_stream_1 (stream, codesys, "r");
3170 make_encoding_output_stream (Lstream *stream, Lisp_Object codesys)
3172 return make_encoding_stream_1 (stream, codesys, "w");
3175 /* Convert N bytes of internally-formatted data stored in SRC to an
3176 external format, according to the encoding stream ENCODING.
3177 Store the encoded data into DST. */
3180 mule_encode (Lstream *encoding, const Bufbyte *src,
3181 unsigned_char_dynarr *dst, size_t n)
3183 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3185 switch (CODING_SYSTEM_TYPE (str->codesys))
3188 case CODESYS_INTERNAL:
3189 Dynarr_add_many (dst, src, n);
3192 case CODESYS_AUTODETECT:
3193 /* If we got this far and still haven't decided on the coding
3194 system, then do no conversion. */
3195 case CODESYS_NO_CONVERSION:
3196 encode_coding_no_conversion (encoding, src, dst, n);
3200 str->ccl.last_block = str->flags & CODING_STATE_END;
3201 /* When applying ccl program to stream, MUST NOT set NULL
3203 ccl_driver (&str->ccl, ((src) ? src : (unsigned char*)""),
3204 dst, n, 0, CCL_MODE_ENCODING);
3208 text_encode_generic (encoding, src, dst, n);
3212 DEFUN ("encode-coding-region", Fencode_coding_region, 3, 4, 0, /*
3213 Encode the text between START and END using CODING-SYSTEM.
3214 This will, for example, convert Japanese characters into stuff such as
3215 "^[$B!<!+^[(B" if you use the JIS encoding. Return length of encoded
3216 text. BUFFER defaults to the current buffer if unspecified.
3218 (start, end, coding_system, buffer))
3221 struct buffer *buf = decode_buffer (buffer, 0);
3222 Lisp_Object instream, lb_outstream, de_outstream, outstream;
3223 Lstream *istr, *ostr;
3224 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
3226 get_buffer_range_char (buf, start, end, &b, &e, 0);
3228 barf_if_buffer_read_only (buf, b, e);
3230 coding_system = Fget_coding_system (coding_system);
3231 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
3232 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
3233 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
3234 Fget_coding_system (Qbinary));
3235 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
3237 istr = XLSTREAM (instream);
3238 ostr = XLSTREAM (outstream);
3239 GCPRO4 (instream, outstream, de_outstream, lb_outstream);
3240 /* The chain of streams looks like this:
3242 [BUFFER] <----- send through
3243 ------> [ENCODE AS SPECIFIED]
3244 ------> [DECODE AS BINARY]
3249 char tempbuf[1024]; /* some random amount */
3250 Bufpos newpos, even_newer_pos;
3251 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
3252 ssize_t size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
3256 newpos = lisp_buffer_stream_startpos (istr);
3257 Lstream_write (ostr, tempbuf, size_in_bytes);
3258 even_newer_pos = lisp_buffer_stream_startpos (istr);
3259 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
3265 lisp_buffer_stream_startpos (XLSTREAM (instream)) - b;
3266 Lstream_close (istr);
3267 Lstream_close (ostr);
3269 Lstream_delete (istr);
3270 Lstream_delete (ostr);
3271 Lstream_delete (XLSTREAM (de_outstream));
3272 Lstream_delete (XLSTREAM (lb_outstream));
3273 return make_int (retlen);
3280 text_encode_generic (Lstream *encoding, const Bufbyte *src,
3281 unsigned_char_dynarr *dst, size_t n)
3284 unsigned char char_boundary;
3285 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3286 unsigned int flags = str->flags;
3287 Emchar ch = str->ch;
3289 char_boundary = str->iso2022.current_char_boundary;
3295 if (char_boundary == 0)
3323 (*str->encode_char) (str, c, dst, &flags);
3325 else if (char_boundary == 1)
3327 (*str->encode_char) (str, (ch << 6) | (c & 0x3f), dst, &flags);
3333 ch = (ch << 6) | (c & 0x3f);
3338 if ((char_boundary == 0) && (flags & CODING_STATE_END))
3340 (*str->finish) (str, dst, &flags);
3345 str->iso2022.current_char_boundary = char_boundary;
3349 /************************************************************************/
3350 /* Shift-JIS methods */
3351 /************************************************************************/
3353 /* Shift-JIS is a coding system encoding three character sets: ASCII, right
3354 half of JISX0201-Kana, and JISX0208. An ASCII character is encoded
3355 as is. A character of JISX0201-Kana (DIMENSION1_CHARS94 character set) is
3356 encoded by "position-code + 0x80". A character of JISX0208
3357 (DIMENSION2_CHARS94 character set) is encoded in 2-byte but two
3358 position-codes are divided and shifted so that it fit in the range
3361 --- CODE RANGE of Shift-JIS ---
3362 (character set) (range)
3364 JISX0201-Kana 0xA0 .. 0xDF
3365 JISX0208 (1st byte) 0x80 .. 0x9F and 0xE0 .. 0xEF
3366 (2nd byte) 0x40 .. 0x7E and 0x80 .. 0xFC
3367 -------------------------------
3371 /* Is this the first byte of a Shift-JIS two-byte char? */
3373 #define BYTE_SJIS_TWO_BYTE_1_P(c) \
3374 (((c) >= 0x81 && (c) <= 0x9F) || ((c) >= 0xE0 && (c) <= 0xEF))
3376 /* Is this the second byte of a Shift-JIS two-byte char? */
3378 #define BYTE_SJIS_TWO_BYTE_2_P(c) \
3379 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0x80 && (c) <= 0xFC))
3381 #define BYTE_SJIS_KATAKANA_P(c) \
3382 ((c) >= 0xA1 && (c) <= 0xDF)
3385 detect_coding_sjis (struct detection_state *st, const Extbyte *src, size_t n)
3389 unsigned char c = *(unsigned char *)src++;
3390 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
3392 if (st->shift_jis.in_second_byte)
3394 st->shift_jis.in_second_byte = 0;
3398 else if ((c >= 0x80 && c < 0xA0) || c >= 0xE0)
3399 st->shift_jis.in_second_byte = 1;
3401 return CODING_CATEGORY_SHIFT_JIS_MASK;
3404 /* Convert Shift-JIS data to internal format. */
3407 decode_coding_sjis (Lstream *decoding, const Extbyte *src,
3408 unsigned_char_dynarr *dst, size_t n)
3410 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3411 unsigned int flags = str->flags;
3412 unsigned int cpos = str->cpos;
3413 eol_type_t eol_type = str->eol_type;
3417 unsigned char c = *(unsigned char *)src++;
3421 /* Previous character was first byte of Shift-JIS Kanji char. */
3422 if (BYTE_SJIS_TWO_BYTE_2_P (c))
3424 unsigned char e1, e2;
3426 DECODE_SJIS (cpos, c, e1, e2);
3428 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_japanese_jisx0208,
3432 Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208);
3433 Dynarr_add (dst, e1);
3434 Dynarr_add (dst, e2);
3439 DECODE_ADD_BINARY_CHAR (cpos, dst);
3440 DECODE_ADD_BINARY_CHAR (c, dst);
3446 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3447 if (BYTE_SJIS_TWO_BYTE_1_P (c))
3449 else if (BYTE_SJIS_KATAKANA_P (c))
3452 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_katakana_jisx0201,
3455 Dynarr_add (dst, LEADING_BYTE_KATAKANA_JISX0201);
3456 Dynarr_add (dst, c);
3461 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_latin_jisx0201,
3465 DECODE_ADD_BINARY_CHAR (c, dst);
3467 label_continue_loop:;
3470 DECODE_HANDLE_END_OF_CONVERSION (flags, cpos, dst);
3476 /* Convert internal character representation to Shift_JIS. */
3479 char_encode_shift_jis (struct encoding_stream *str, Emchar ch,
3480 unsigned_char_dynarr *dst, unsigned int *flags)
3482 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3486 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3487 Dynarr_add (dst, '\r');
3488 if (eol_type != EOL_CR)
3489 Dynarr_add (dst, ch);
3493 unsigned int s1, s2;
3495 int code_point = charset_code_point (Vcharset_latin_jisx0201, ch);
3497 if (code_point >= 0)
3498 Dynarr_add (dst, code_point);
3499 else if ((code_point
3500 = charset_code_point (Vcharset_japanese_jisx0208_1990, ch))
3503 ENCODE_SJIS ((code_point >> 8) | 0x80,
3504 (code_point & 0xFF) | 0x80, s1, s2);
3505 Dynarr_add (dst, s1);
3506 Dynarr_add (dst, s2);
3508 else if ((code_point
3509 = charset_code_point (Vcharset_katakana_jisx0201, ch))
3511 Dynarr_add (dst, code_point | 0x80);
3512 else if ((code_point
3513 = charset_code_point (Vcharset_japanese_jisx0208, ch))
3516 ENCODE_SJIS ((code_point >> 8) | 0x80,
3517 (code_point & 0xFF) | 0x80, s1, s2);
3518 Dynarr_add (dst, s1);
3519 Dynarr_add (dst, s2);
3521 else if ((code_point = charset_code_point (Vcharset_ascii, ch))
3523 Dynarr_add (dst, code_point);
3525 Dynarr_add (dst, '?');
3527 Lisp_Object charset;
3528 unsigned int c1, c2;
3530 BREAKUP_CHAR (ch, charset, c1, c2);
3532 if (EQ(charset, Vcharset_katakana_jisx0201))
3534 Dynarr_add (dst, c1 | 0x80);
3538 Dynarr_add (dst, c1);
3540 else if (EQ(charset, Vcharset_japanese_jisx0208))
3542 ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
3543 Dynarr_add (dst, s1);
3544 Dynarr_add (dst, s2);
3547 Dynarr_add (dst, '?');
3553 char_finish_shift_jis (struct encoding_stream *str, unsigned_char_dynarr *dst,
3554 unsigned int *flags)
3558 DEFUN ("decode-shift-jis-char", Fdecode_shift_jis_char, 1, 1, 0, /*
3559 Decode a JISX0208 character of Shift-JIS coding-system.
3560 CODE is the character code in Shift-JIS as a cons of type bytes.
3561 Return the corresponding character.
3565 unsigned char c1, c2, s1, s2;
3568 CHECK_INT (XCAR (code));
3569 CHECK_INT (XCDR (code));
3570 s1 = XINT (XCAR (code));
3571 s2 = XINT (XCDR (code));
3572 if (BYTE_SJIS_TWO_BYTE_1_P (s1) &&
3573 BYTE_SJIS_TWO_BYTE_2_P (s2))
3575 DECODE_SJIS (s1, s2, c1, c2);
3576 return make_char (MAKE_CHAR (Vcharset_japanese_jisx0208,
3577 c1 & 0x7F, c2 & 0x7F));
3583 DEFUN ("encode-shift-jis-char", Fencode_shift_jis_char, 1, 1, 0, /*
3584 Encode a JISX0208 character CHARACTER to SHIFT-JIS coding-system.
3585 Return the corresponding character code in SHIFT-JIS as a cons of two bytes.
3589 Lisp_Object charset;
3592 CHECK_CHAR_COERCE_INT (character);
3593 BREAKUP_CHAR (XCHAR (character), charset, c1, c2);
3594 if (EQ (charset, Vcharset_japanese_jisx0208))
3596 ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
3597 return Fcons (make_int (s1), make_int (s2));
3604 /************************************************************************/
3606 /************************************************************************/
3608 /* BIG5 is a coding system encoding two character sets: ASCII and
3609 Big5. An ASCII character is encoded as is. Big5 is a two-byte
3610 character set and is encoded in two-byte.
3612 --- CODE RANGE of BIG5 ---
3613 (character set) (range)
3615 Big5 (1st byte) 0xA1 .. 0xFE
3616 (2nd byte) 0x40 .. 0x7E and 0xA1 .. 0xFE
3617 --------------------------
3619 Since the number of characters in Big5 is larger than maximum
3620 characters in Emacs' charset (96x96), it can't be handled as one
3621 charset. So, in Emacs, Big5 is divided into two: `charset-big5-1'
3622 and `charset-big5-2'. Both <type>s are DIMENSION2_CHARS94. The former
3623 contains frequently used characters and the latter contains less
3624 frequently used characters. */
3627 #define BYTE_BIG5_TWO_BYTE_1_P(c) \
3628 ((c) >= 0x81 && (c) <= 0xFE)
3630 #define BYTE_BIG5_TWO_BYTE_1_P(c) \
3631 ((c) >= 0xA1 && (c) <= 0xFE)
3634 /* Is this the second byte of a Shift-JIS two-byte char? */
3636 #define BYTE_BIG5_TWO_BYTE_2_P(c) \
3637 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0xA1 && (c) <= 0xFE))
3639 /* Number of Big5 characters which have the same code in 1st byte. */
3641 #define BIG5_SAME_ROW (0xFF - 0xA1 + 0x7F - 0x40)
3643 /* Code conversion macros. These are macros because they are used in
3644 inner loops during code conversion.
3646 Note that temporary variables in macros introduce the classic
3647 dynamic-scoping problems with variable names. We use capital-
3648 lettered variables in the assumption that XEmacs does not use
3649 capital letters in variables except in a very formalized way
3652 /* Convert Big5 code (b1, b2) into its internal string representation
3655 /* There is a much simpler way to split the Big5 charset into two.
3656 For the moment I'm going to leave the algorithm as-is because it
3657 claims to separate out the most-used characters into a single
3658 charset, which perhaps will lead to optimizations in various
3661 The way the algorithm works is something like this:
3663 Big5 can be viewed as a 94x157 charset, where the row is
3664 encoded into the bytes 0xA1 .. 0xFE and the column is encoded
3665 into the bytes 0x40 .. 0x7E and 0xA1 .. 0xFE. As for frequency,
3666 the split between low and high column numbers is apparently
3667 meaningless; ascending rows produce less and less frequent chars.
3668 Therefore, we assign the lower half of rows (0xA1 .. 0xC8) to
3669 the first charset, and the upper half (0xC9 .. 0xFE) to the
3670 second. To do the conversion, we convert the character into
3671 a single number where 0 .. 156 is the first row, 157 .. 313
3672 is the second, etc. That way, the characters are ordered by
3673 decreasing frequency. Then we just chop the space in two
3674 and coerce the result into a 94x94 space.
3677 #define DECODE_BIG5(b1, b2, lb, c1, c2) do \
3679 int B1 = b1, B2 = b2; \
3681 = (B1 - 0xA1) * BIG5_SAME_ROW + B2 - (B2 < 0x7F ? 0x40 : 0x62); \
3685 lb = LEADING_BYTE_CHINESE_BIG5_1; \
3689 lb = LEADING_BYTE_CHINESE_BIG5_2; \
3690 I -= (BIG5_SAME_ROW) * (0xC9 - 0xA1); \
3692 c1 = I / (0xFF - 0xA1) + 0xA1; \
3693 c2 = I % (0xFF - 0xA1) + 0xA1; \
3696 /* Convert the internal string representation of a Big5 character
3697 (lb, c1, c2) into Big5 code (b1, b2). */
3699 #define ENCODE_BIG5(lb, c1, c2, b1, b2) do \
3701 unsigned int I = ((c1) - 0xA1) * (0xFF - 0xA1) + ((c2) - 0xA1); \
3703 if (lb == LEADING_BYTE_CHINESE_BIG5_2) \
3705 I += BIG5_SAME_ROW * (0xC9 - 0xA1); \
3707 b1 = I / BIG5_SAME_ROW + 0xA1; \
3708 b2 = I % BIG5_SAME_ROW; \
3709 b2 += b2 < 0x3F ? 0x40 : 0x62; \
3713 detect_coding_big5 (struct detection_state *st, const Extbyte *src, size_t n)
3717 unsigned char c = *(unsigned char *)src++;
3718 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO
3720 || (c >= 0x80 && c <= 0xA0)
3724 if (st->big5.in_second_byte)
3726 st->big5.in_second_byte = 0;
3727 if (c < 0x40 || (c >= 0x80 && c <= 0xA0))
3737 st->big5.in_second_byte = 1;
3739 return CODING_CATEGORY_BIG5_MASK;
3742 /* Convert Big5 data to internal format. */
3745 decode_coding_big5 (Lstream *decoding, const Extbyte *src,
3746 unsigned_char_dynarr *dst, size_t n)
3748 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3749 unsigned int flags = str->flags;
3750 unsigned int cpos = str->cpos;
3751 eol_type_t eol_type = str->eol_type;
3754 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (DECODING_STREAM_DATA
3755 (decoding)->codesys, 1);
3760 unsigned char c = *(unsigned char *)src++;
3763 /* Previous character was first byte of Big5 char. */
3764 if (BYTE_BIG5_TWO_BYTE_2_P (c))
3767 int code_point = (cpos << 8) | c;
3768 Emchar char_id = decode_defined_char (ccs, code_point);
3771 char_id = DECODE_CHAR (Vcharset_chinese_big5, code_point);
3772 DECODE_ADD_UCS_CHAR (char_id, dst);
3774 unsigned char b1, b2, b3;
3775 DECODE_BIG5 (cpos, c, b1, b2, b3);
3776 Dynarr_add (dst, b1);
3777 Dynarr_add (dst, b2);
3778 Dynarr_add (dst, b3);
3783 DECODE_ADD_BINARY_CHAR (cpos, dst);
3784 DECODE_ADD_BINARY_CHAR (c, dst);
3790 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3791 if (BYTE_BIG5_TWO_BYTE_1_P (c))
3794 DECODE_ADD_BINARY_CHAR (c, dst);
3796 label_continue_loop:;
3799 DECODE_HANDLE_END_OF_CONVERSION (flags, cpos, dst);
3805 /* Convert internally-formatted data to Big5. */
3808 char_encode_big5 (struct encoding_stream *str, Emchar ch,
3809 unsigned_char_dynarr *dst, unsigned int *flags)
3811 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3815 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3816 Dynarr_add (dst, '\r');
3817 if (eol_type != EOL_CR)
3818 Dynarr_add (dst, ch);
3825 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, 1);
3827 if ((code_point = charset_code_point (Vcharset_ascii, ch)) >= 0)
3828 Dynarr_add (dst, code_point);
3829 else if ((code_point = charset_code_point (ccs, ch)) >= 0)
3831 Dynarr_add (dst, code_point >> 8);
3832 Dynarr_add (dst, code_point & 0xFF);
3834 else if ((code_point
3835 = charset_code_point (Vcharset_chinese_big5, ch)) >= 0)
3837 Dynarr_add (dst, code_point >> 8);
3838 Dynarr_add (dst, code_point & 0xFF);
3840 else if ((code_point
3841 = charset_code_point (Vcharset_chinese_big5_1, ch)) >= 0)
3844 = ((code_point >> 8) - 33) * (0xFF - 0xA1)
3845 + ((code_point & 0xFF) - 33);
3846 unsigned char b1 = I / BIG5_SAME_ROW + 0xA1;
3847 unsigned char b2 = I % BIG5_SAME_ROW;
3849 b2 += b2 < 0x3F ? 0x40 : 0x62;
3850 Dynarr_add (dst, b1);
3851 Dynarr_add (dst, b2);
3853 else if ((code_point
3854 = charset_code_point (Vcharset_chinese_big5_2, ch)) >= 0)
3857 = ((code_point >> 8) - 33) * (0xFF - 0xA1)
3858 + ((code_point & 0xFF) - 33);
3859 unsigned char b1, b2;
3861 I += BIG5_SAME_ROW * (0xC9 - 0xA1);
3862 b1 = I / BIG5_SAME_ROW + 0xA1;
3863 b2 = I % BIG5_SAME_ROW;
3864 b2 += b2 < 0x3F ? 0x40 : 0x62;
3865 Dynarr_add (dst, b1);
3866 Dynarr_add (dst, b2);
3869 Dynarr_add (dst, '?');
3876 char_finish_big5 (struct encoding_stream *str, unsigned_char_dynarr *dst,
3877 unsigned int *flags)
3882 DEFUN ("decode-big5-char", Fdecode_big5_char, 1, 1, 0, /*
3883 Decode a Big5 character CODE of BIG5 coding-system.
3884 CODE is the character code in BIG5, a cons of two integers.
3885 Return the corresponding character.
3889 unsigned char c1, c2, b1, b2;
3892 CHECK_INT (XCAR (code));
3893 CHECK_INT (XCDR (code));
3894 b1 = XINT (XCAR (code));
3895 b2 = XINT (XCDR (code));
3896 if (BYTE_BIG5_TWO_BYTE_1_P (b1) &&
3897 BYTE_BIG5_TWO_BYTE_2_P (b2))
3899 Charset_ID leading_byte;
3900 Lisp_Object charset;
3901 DECODE_BIG5 (b1, b2, leading_byte, c1, c2);
3902 charset = CHARSET_BY_LEADING_BYTE (leading_byte);
3903 return make_char (MAKE_CHAR (charset, c1 & 0x7F, c2 & 0x7F));
3909 DEFUN ("encode-big5-char", Fencode_big5_char, 1, 1, 0, /*
3910 Encode the Big5 character CHARACTER in the BIG5 coding-system.
3911 Return the corresponding character code in Big5.
3915 Lisp_Object charset;
3918 CHECK_CHAR_COERCE_INT (character);
3919 BREAKUP_CHAR (XCHAR (character), charset, c1, c2);
3920 if (EQ (charset, Vcharset_chinese_big5_1) ||
3921 EQ (charset, Vcharset_chinese_big5_2))
3923 ENCODE_BIG5 (XCHARSET_LEADING_BYTE (charset), c1 | 0x80, c2 | 0x80,
3925 return Fcons (make_int (b1), make_int (b2));
3932 /************************************************************************/
3934 /************************************************************************/
3937 detect_coding_ucs4 (struct detection_state *st, const Extbyte *src, size_t n)
3941 unsigned char c = *(unsigned char *)src++;
3942 switch (st->ucs4.in_byte)
3951 st->ucs4.in_byte = 0;
3957 return CODING_CATEGORY_UCS4_MASK;
3961 decode_coding_ucs4 (Lstream *decoding, const Extbyte *src,
3962 unsigned_char_dynarr *dst, size_t n)
3964 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3965 unsigned int flags = str->flags;
3966 unsigned int cpos = str->cpos;
3967 unsigned char counter = str->counter;
3971 unsigned char c = *(unsigned char *)src++;
3979 DECODE_ADD_UCS_CHAR ((cpos << 8) | c, dst);
3984 cpos = ( cpos << 8 ) | c;
3988 if (counter & CODING_STATE_END)
3989 DECODE_OUTPUT_PARTIAL_CHAR (cpos);
3993 str->counter = counter;
3997 char_encode_ucs4 (struct encoding_stream *str, Emchar ch,
3998 unsigned_char_dynarr *dst, unsigned int *flags)
4000 Dynarr_add (dst, ch >> 24);
4001 Dynarr_add (dst, ch >> 16);
4002 Dynarr_add (dst, ch >> 8);
4003 Dynarr_add (dst, ch );
4007 char_finish_ucs4 (struct encoding_stream *str, unsigned_char_dynarr *dst,
4008 unsigned int *flags)
4013 /************************************************************************/
4015 /************************************************************************/
4018 detect_coding_utf8 (struct detection_state *st, const Extbyte *src, size_t n)
4022 unsigned char c = *(unsigned char *)src++;
4023 switch (st->utf8.in_byte)
4026 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
4029 st->utf8.in_byte = 5;
4031 st->utf8.in_byte = 4;
4033 st->utf8.in_byte = 3;
4035 st->utf8.in_byte = 2;
4037 st->utf8.in_byte = 1;
4042 if ((c & 0xc0) != 0x80)
4048 return CODING_CATEGORY_UTF8_MASK;
4052 decode_output_utf8_partial_char (unsigned char counter,
4054 unsigned_char_dynarr *dst)
4057 DECODE_ADD_BINARY_CHAR ( (cpos|0xFC), dst);
4058 else if (counter == 4)
4060 if (cpos < (1 << 6))
4061 DECODE_ADD_BINARY_CHAR ( (cpos|0xF8), dst);
4064 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xFC), dst);
4065 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4068 else if (counter == 3)
4070 if (cpos < (1 << 6))
4071 DECODE_ADD_BINARY_CHAR ( (cpos|0xF0), dst);
4072 else if (cpos < (1 << 12))
4074 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xF8), dst);
4075 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4079 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 12)|0xFC), dst);
4080 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4081 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4084 else if (counter == 2)
4086 if (cpos < (1 << 6))
4087 DECODE_ADD_BINARY_CHAR ( (cpos|0xE0), dst);
4088 else if (cpos < (1 << 12))
4090 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xF0), dst);
4091 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4093 else if (cpos < (1 << 18))
4095 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 12)|0xF8), dst);
4096 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4097 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4101 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 18)|0xFC), dst);
4102 DECODE_ADD_BINARY_CHAR ( (((cpos >> 12)&0x3F)|0x80), dst);
4103 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4104 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4109 if (cpos < (1 << 6))
4110 DECODE_ADD_BINARY_CHAR ( (cpos|0xC0), dst);
4111 else if (cpos < (1 << 12))
4113 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xE0), dst);
4114 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4116 else if (cpos < (1 << 18))
4118 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 12)|0xF0), dst);
4119 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4120 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4122 else if (cpos < (1 << 24))
4124 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 18)|0xF8), dst);
4125 DECODE_ADD_BINARY_CHAR ( (((cpos >> 12)&0x3F)|0x80), dst);
4126 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4127 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4131 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 24)|0xFC), dst);
4132 DECODE_ADD_BINARY_CHAR ( (((cpos >> 18)&0x3F)|0x80), dst);
4133 DECODE_ADD_BINARY_CHAR ( (((cpos >> 12)&0x3F)|0x80), dst);
4134 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4135 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4141 decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
4142 unsigned_char_dynarr *dst, size_t n)
4144 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
4145 unsigned int flags = str->flags;
4146 unsigned int cpos = str->cpos;
4147 eol_type_t eol_type = str->eol_type;
4148 unsigned char counter = str->counter;
4152 unsigned char c = *(unsigned char *)src++;
4157 COMPOSE_FLUSH_CHARS (str, dst);
4158 decode_flush_er_chars (str, dst);
4159 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
4160 DECODE_ADD_UCS_CHAR (c, dst);
4162 else if ( c < 0xC0 )
4163 /* decode_add_er_char (str, c, dst); */
4164 COMPOSE_ADD_CHAR (str, c, dst);
4167 /* decode_flush_er_chars (str, dst); */
4173 else if ( c < 0xF0 )
4178 else if ( c < 0xF8 )
4183 else if ( c < 0xFC )
4195 else if ( (c & 0xC0) == 0x80 )
4197 cpos = ( cpos << 6 ) | ( c & 0x3f );
4200 /* DECODE_ADD_UCS_CHAR (cpos, dst); */
4201 COMPOSE_ADD_CHAR (str, cpos, dst);
4210 COMPOSE_FLUSH_CHARS (str, dst);
4211 decode_flush_er_chars (str, dst);
4212 decode_output_utf8_partial_char (counter, cpos, dst);
4213 DECODE_ADD_BINARY_CHAR (c, dst);
4217 label_continue_loop:;
4220 if (flags & CODING_STATE_END)
4222 COMPOSE_FLUSH_CHARS (str, dst);
4223 decode_flush_er_chars (str, dst);
4226 decode_output_utf8_partial_char (counter, cpos, dst);
4233 str->counter = counter;
4237 char_encode_utf8 (struct encoding_stream *str, Emchar ch,
4238 unsigned_char_dynarr *dst, unsigned int *flags)
4240 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
4244 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
4245 Dynarr_add (dst, '\r');
4246 if (eol_type != EOL_CR)
4247 Dynarr_add (dst, ch);
4249 else if (ch <= 0x7f)
4251 Dynarr_add (dst, ch);
4255 int code_point = charset_code_point (Vcharset_ucs, ch);
4257 if ( (code_point < 0) || (code_point > 0x10FFFF) )
4259 if (CODING_SYSTEM_USE_ENTITY_REFERENCE (str->codesys))
4261 Lisp_Object rest = Vcoded_charset_entity_reference_alist;
4265 int format_columns, idx;
4266 char buf[16], format[16];
4268 while (!NILP (rest))
4272 if (!NILP (ccs = Ffind_charset (ccs)))
4275 = charset_code_point (ccs, ch)) >= 0 )
4280 && ((idx =XSTRING_LENGTH (ret)) <= 6))
4282 strncpy (format, XSTRING_DATA (ret), idx);
4291 format [idx++] = '%';
4292 format_columns = XINT (ret);
4293 if ( (2 <= format_columns)
4294 && (format_columns <= 8) )
4296 format [idx++] = '0';
4297 format [idx++] = '0' + format_columns;
4304 format [idx++] = 'd';
4305 else if (EQ (ret, Qx))
4306 format [idx++] = 'x';
4307 else if (EQ (ret, QX))
4308 format [idx++] = 'X';
4313 sprintf (buf, format, code_point);
4314 Dynarr_add (dst, '&');
4315 Dynarr_add_many (dst, buf, strlen (buf));
4316 Dynarr_add (dst, ';');
4322 sprintf (buf, "&MCS-%08X;", ch);
4323 Dynarr_add_many (dst, buf, strlen (buf));
4329 if (code_point <= 0x7ff)
4331 Dynarr_add (dst, (code_point >> 6) | 0xc0);
4332 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4334 else if (code_point <= 0xffff)
4336 Dynarr_add (dst, (code_point >> 12) | 0xe0);
4337 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4338 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4340 else if (code_point <= 0x1fffff)
4342 Dynarr_add (dst, (code_point >> 18) | 0xf0);
4343 Dynarr_add (dst, ((code_point >> 12) & 0x3f) | 0x80);
4344 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4345 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4347 else if (code_point <= 0x3ffffff)
4349 Dynarr_add (dst, (code_point >> 24) | 0xf8);
4350 Dynarr_add (dst, ((code_point >> 18) & 0x3f) | 0x80);
4351 Dynarr_add (dst, ((code_point >> 12) & 0x3f) | 0x80);
4352 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4353 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4357 Dynarr_add (dst, (code_point >> 30) | 0xfc);
4358 Dynarr_add (dst, ((code_point >> 24) & 0x3f) | 0x80);
4359 Dynarr_add (dst, ((code_point >> 18) & 0x3f) | 0x80);
4360 Dynarr_add (dst, ((code_point >> 12) & 0x3f) | 0x80);
4361 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4362 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4368 char_finish_utf8 (struct encoding_stream *str, unsigned_char_dynarr *dst,
4369 unsigned int *flags)
4374 /************************************************************************/
4375 /* ISO2022 methods */
4376 /************************************************************************/
4378 /* The following note describes the coding system ISO2022 briefly.
4379 Since the intention of this note is to help understand the
4380 functions in this file, some parts are NOT ACCURATE or OVERLY
4381 SIMPLIFIED. For thorough understanding, please refer to the
4382 original document of ISO2022.
4384 ISO2022 provides many mechanisms to encode several character sets
4385 in 7-bit and 8-bit environments. For 7-bit environments, all text
4386 is encoded using bytes less than 128. This may make the encoded
4387 text a little bit longer, but the text passes more easily through
4388 several gateways, some of which strip off MSB (Most Signigant Bit).
4390 There are two kinds of character sets: control character set and
4391 graphic character set. The former contains control characters such
4392 as `newline' and `escape' to provide control functions (control
4393 functions are also provided by escape sequences). The latter
4394 contains graphic characters such as 'A' and '-'. Emacs recognizes
4395 two control character sets and many graphic character sets.
4397 Graphic character sets are classified into one of the following
4398 four classes, according to the number of bytes (DIMENSION) and
4399 number of characters in one dimension (CHARS) of the set:
4400 - DIMENSION1_CHARS94
4401 - DIMENSION1_CHARS96
4402 - DIMENSION2_CHARS94
4403 - DIMENSION2_CHARS96
4405 In addition, each character set is assigned an identification tag,
4406 unique for each set, called "final character" (denoted as <F>
4407 hereafter). The <F> of each character set is decided by ECMA(*)
4408 when it is registered in ISO. The code range of <F> is 0x30..0x7F
4409 (0x30..0x3F are for private use only).
4411 Note (*): ECMA = European Computer Manufacturers Association
4413 Here are examples of graphic character set [NAME(<F>)]:
4414 o DIMENSION1_CHARS94 -- ASCII('B'), right-half-of-JISX0201('I'), ...
4415 o DIMENSION1_CHARS96 -- right-half-of-ISO8859-1('A'), ...
4416 o DIMENSION2_CHARS94 -- GB2312('A'), JISX0208('B'), ...
4417 o DIMENSION2_CHARS96 -- none for the moment
4419 A code area (1 byte = 8 bits) is divided into 4 areas, C0, GL, C1, and GR.
4420 C0 [0x00..0x1F] -- control character plane 0
4421 GL [0x20..0x7F] -- graphic character plane 0
4422 C1 [0x80..0x9F] -- control character plane 1
4423 GR [0xA0..0xFF] -- graphic character plane 1
4425 A control character set is directly designated and invoked to C0 or
4426 C1 by an escape sequence. The most common case is that:
4427 - ISO646's control character set is designated/invoked to C0, and
4428 - ISO6429's control character set is designated/invoked to C1,
4429 and usually these designations/invocations are omitted in encoded
4430 text. In a 7-bit environment, only C0 can be used, and a control
4431 character for C1 is encoded by an appropriate escape sequence to
4432 fit into the environment. All control characters for C1 are
4433 defined to have corresponding escape sequences.
4435 A graphic character set is at first designated to one of four
4436 graphic registers (G0 through G3), then these graphic registers are
4437 invoked to GL or GR. These designations and invocations can be
4438 done independently. The most common case is that G0 is invoked to
4439 GL, G1 is invoked to GR, and ASCII is designated to G0. Usually
4440 these invocations and designations are omitted in encoded text.
4441 In a 7-bit environment, only GL can be used.
4443 When a graphic character set of CHARS94 is invoked to GL, codes
4444 0x20 and 0x7F of the GL area work as control characters SPACE and
4445 DEL respectively, and codes 0xA0 and 0xFF of the GR area should not
4448 There are two ways of invocation: locking-shift and single-shift.
4449 With locking-shift, the invocation lasts until the next different
4450 invocation, whereas with single-shift, the invocation affects the
4451 following character only and doesn't affect the locking-shift
4452 state. Invocations are done by the following control characters or
4455 ----------------------------------------------------------------------
4456 abbrev function cntrl escape seq description
4457 ----------------------------------------------------------------------
4458 SI/LS0 (shift-in) 0x0F none invoke G0 into GL
4459 SO/LS1 (shift-out) 0x0E none invoke G1 into GL
4460 LS2 (locking-shift-2) none ESC 'n' invoke G2 into GL
4461 LS3 (locking-shift-3) none ESC 'o' invoke G3 into GL
4462 LS1R (locking-shift-1 right) none ESC '~' invoke G1 into GR (*)
4463 LS2R (locking-shift-2 right) none ESC '}' invoke G2 into GR (*)
4464 LS3R (locking-shift 3 right) none ESC '|' invoke G3 into GR (*)
4465 SS2 (single-shift-2) 0x8E ESC 'N' invoke G2 for one char
4466 SS3 (single-shift-3) 0x8F ESC 'O' invoke G3 for one char
4467 ----------------------------------------------------------------------
4468 (*) These are not used by any known coding system.
4470 Control characters for these functions are defined by macros
4471 ISO_CODE_XXX in `coding.h'.
4473 Designations are done by the following escape sequences:
4474 ----------------------------------------------------------------------
4475 escape sequence description
4476 ----------------------------------------------------------------------
4477 ESC '(' <F> designate DIMENSION1_CHARS94<F> to G0
4478 ESC ')' <F> designate DIMENSION1_CHARS94<F> to G1
4479 ESC '*' <F> designate DIMENSION1_CHARS94<F> to G2
4480 ESC '+' <F> designate DIMENSION1_CHARS94<F> to G3
4481 ESC ',' <F> designate DIMENSION1_CHARS96<F> to G0 (*)
4482 ESC '-' <F> designate DIMENSION1_CHARS96<F> to G1
4483 ESC '.' <F> designate DIMENSION1_CHARS96<F> to G2
4484 ESC '/' <F> designate DIMENSION1_CHARS96<F> to G3
4485 ESC '$' '(' <F> designate DIMENSION2_CHARS94<F> to G0 (**)
4486 ESC '$' ')' <F> designate DIMENSION2_CHARS94<F> to G1
4487 ESC '$' '*' <F> designate DIMENSION2_CHARS94<F> to G2
4488 ESC '$' '+' <F> designate DIMENSION2_CHARS94<F> to G3
4489 ESC '$' ',' <F> designate DIMENSION2_CHARS96<F> to G0 (*)
4490 ESC '$' '-' <F> designate DIMENSION2_CHARS96<F> to G1
4491 ESC '$' '.' <F> designate DIMENSION2_CHARS96<F> to G2
4492 ESC '$' '/' <F> designate DIMENSION2_CHARS96<F> to G3
4493 ----------------------------------------------------------------------
4495 In this list, "DIMENSION1_CHARS94<F>" means a graphic character set
4496 of dimension 1, chars 94, and final character <F>, etc...
4498 Note (*): Although these designations are not allowed in ISO2022,
4499 Emacs accepts them on decoding, and produces them on encoding
4500 CHARS96 character sets in a coding system which is characterized as
4501 7-bit environment, non-locking-shift, and non-single-shift.
4503 Note (**): If <F> is '@', 'A', or 'B', the intermediate character
4504 '(' can be omitted. We refer to this as "short-form" hereafter.
4506 Now you may notice that there are a lot of ways for encoding the
4507 same multilingual text in ISO2022. Actually, there exist many
4508 coding systems such as Compound Text (used in X11's inter client
4509 communication, ISO-2022-JP (used in Japanese internet), ISO-2022-KR
4510 (used in Korean internet), EUC (Extended UNIX Code, used in Asian
4511 localized platforms), and all of these are variants of ISO2022.
4513 In addition to the above, Emacs handles two more kinds of escape
4514 sequences: ISO6429's direction specification and Emacs' private
4515 sequence for specifying character composition.
4517 ISO6429's direction specification takes the following form:
4518 o CSI ']' -- end of the current direction
4519 o CSI '0' ']' -- end of the current direction
4520 o CSI '1' ']' -- start of left-to-right text
4521 o CSI '2' ']' -- start of right-to-left text
4522 The control character CSI (0x9B: control sequence introducer) is
4523 abbreviated to the escape sequence ESC '[' in a 7-bit environment.
4525 Character composition specification takes the following form:
4526 o ESC '0' -- start character composition
4527 o ESC '1' -- end character composition
4528 Since these are not standard escape sequences of any ISO standard,
4529 their use with these meanings is restricted to Emacs only. */
4532 reset_iso2022 (Lisp_Object coding_system, struct iso2022_decoder *iso)
4536 for (i = 0; i < 4; i++)
4538 if (!NILP (coding_system))
4540 XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, i);
4542 iso->charset[i] = Qt;
4543 iso->invalid_designated[i] = 0;
4545 iso->esc = ISO_ESC_NOTHING;
4546 iso->esc_bytes_index = 0;
4547 iso->register_left = 0;
4548 iso->register_right = 1;
4549 iso->switched_dir_and_no_valid_charset_yet = 0;
4550 iso->invalid_switch_dir = 0;
4551 iso->output_direction_sequence = 0;
4552 iso->output_literally = 0;
4553 #ifdef ENABLE_COMPOSITE_CHARS
4554 if (iso->composite_chars)
4555 Dynarr_reset (iso->composite_chars);
4560 fit_to_be_escape_quoted (unsigned char c)
4577 /* Parse one byte of an ISO2022 escape sequence.
4578 If the result is an invalid escape sequence, return 0 and
4579 do not change anything in STR. Otherwise, if the result is
4580 an incomplete escape sequence, update ISO2022.ESC and
4581 ISO2022.ESC_BYTES and return -1. Otherwise, update
4582 all the state variables (but not ISO2022.ESC_BYTES) and
4585 If CHECK_INVALID_CHARSETS is non-zero, check for designation
4586 or invocation of an invalid character set and treat that as
4587 an unrecognized escape sequence. */
4590 parse_iso2022_esc (Lisp_Object codesys, struct iso2022_decoder *iso,
4591 unsigned char c, unsigned int *flags,
4592 int check_invalid_charsets)
4594 /* (1) If we're at the end of a designation sequence, CS is the
4595 charset being designated and REG is the register to designate
4598 (2) If we're at the end of a locking-shift sequence, REG is
4599 the register to invoke and HALF (0 == left, 1 == right) is
4600 the half to invoke it into.
4602 (3) If we're at the end of a single-shift sequence, REG is
4603 the register to invoke. */
4604 Lisp_Object cs = Qnil;
4607 /* NOTE: This code does goto's all over the fucking place.
4608 The reason for this is that we're basically implementing
4609 a state machine here, and hierarchical languages like C
4610 don't really provide a clean way of doing this. */
4612 if (! (*flags & CODING_STATE_ESCAPE))
4613 /* At beginning of escape sequence; we need to reset our
4614 escape-state variables. */
4615 iso->esc = ISO_ESC_NOTHING;
4617 iso->output_literally = 0;
4618 iso->output_direction_sequence = 0;
4622 case ISO_ESC_NOTHING:
4623 iso->esc_bytes_index = 0;
4626 case ISO_CODE_ESC: /* Start escape sequence */
4627 *flags |= CODING_STATE_ESCAPE;
4631 case ISO_CODE_CSI: /* ISO6429 (specifying directionality) */
4632 *flags |= CODING_STATE_ESCAPE;
4633 iso->esc = ISO_ESC_5_11;
4636 case ISO_CODE_SO: /* locking shift 1 */
4639 case ISO_CODE_SI: /* locking shift 0 */
4643 case ISO_CODE_SS2: /* single shift */
4646 case ISO_CODE_SS3: /* single shift */
4650 default: /* Other control characters */
4657 /**** single shift ****/
4659 case 'N': /* single shift 2 */
4662 case 'O': /* single shift 3 */
4666 /**** locking shift ****/
4668 case '~': /* locking shift 1 right */
4671 case 'n': /* locking shift 2 */
4674 case '}': /* locking shift 2 right */
4677 case 'o': /* locking shift 3 */
4680 case '|': /* locking shift 3 right */
4684 #ifdef ENABLE_COMPOSITE_CHARS
4685 /**** composite ****/
4688 iso->esc = ISO_ESC_START_COMPOSITE;
4689 *flags = (*flags & CODING_STATE_ISO2022_LOCK) |
4690 CODING_STATE_COMPOSITE;
4694 iso->esc = ISO_ESC_END_COMPOSITE;
4695 *flags = (*flags & CODING_STATE_ISO2022_LOCK) &
4696 ~CODING_STATE_COMPOSITE;
4698 #endif /* ENABLE_COMPOSITE_CHARS */
4700 /**** directionality ****/
4703 iso->esc = ISO_ESC_5_11;
4706 /**** designation ****/
4708 case '$': /* multibyte charset prefix */
4709 iso->esc = ISO_ESC_2_4;
4713 if (0x28 <= c && c <= 0x2F)
4715 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_8);
4719 /* This function is called with CODESYS equal to nil when
4720 doing coding-system detection. */
4722 && XCODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
4723 && fit_to_be_escape_quoted (c))
4725 iso->esc = ISO_ESC_LITERAL;
4726 *flags &= CODING_STATE_ISO2022_LOCK;
4736 /**** directionality ****/
4738 case ISO_ESC_5_11: /* ISO6429 direction control */
4741 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4742 goto directionality;
4744 if (c == '0') iso->esc = ISO_ESC_5_11_0;
4745 else if (c == '1') iso->esc = ISO_ESC_5_11_1;
4746 else if (c == '2') iso->esc = ISO_ESC_5_11_2;
4750 case ISO_ESC_5_11_0:
4753 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4754 goto directionality;
4758 case ISO_ESC_5_11_1:
4761 *flags = (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4762 goto directionality;
4766 case ISO_ESC_5_11_2:
4769 *flags = (*flags & CODING_STATE_ISO2022_LOCK) | CODING_STATE_R2L;
4770 goto directionality;
4775 iso->esc = ISO_ESC_DIRECTIONALITY;
4776 /* Various junk here to attempt to preserve the direction sequences
4777 literally in the text if they would otherwise be swallowed due
4778 to invalid designations that don't show up as actual charset
4779 changes in the text. */
4780 if (iso->invalid_switch_dir)
4782 /* We already inserted a direction switch literally into the
4783 text. We assume (#### this may not be right) that the
4784 next direction switch is the one going the other way,
4785 and we need to output that literally as well. */
4786 iso->output_literally = 1;
4787 iso->invalid_switch_dir = 0;
4793 /* If we are in the thrall of an invalid designation,
4794 then stick the directionality sequence literally into the
4795 output stream so it ends up in the original text again. */
4796 for (jj = 0; jj < 4; jj++)
4797 if (iso->invalid_designated[jj])
4801 iso->output_literally = 1;
4802 iso->invalid_switch_dir = 1;
4805 /* Indicate that we haven't yet seen a valid designation,
4806 so that if a switch-dir is directly followed by an
4807 invalid designation, both get inserted literally. */
4808 iso->switched_dir_and_no_valid_charset_yet = 1;
4813 /**** designation ****/
4816 if (0x28 <= c && c <= 0x2F)
4818 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_4_8);
4821 if (0x40 <= c && c <= 0x42)
4824 cs = CHARSET_BY_ATTRIBUTES (94, -1, c,
4825 *flags & CODING_STATE_R2L ?
4826 CHARSET_RIGHT_TO_LEFT :
4827 CHARSET_LEFT_TO_RIGHT);
4838 if (c < '0' || c > '~')
4839 return 0; /* bad final byte */
4841 if (iso->esc >= ISO_ESC_2_8 &&
4842 iso->esc <= ISO_ESC_2_15)
4844 chars = (iso->esc >= ISO_ESC_2_12) ? 96 : 94;
4845 single = 1; /* single-byte */
4846 reg = (iso->esc - ISO_ESC_2_8) & 3;
4848 else if (iso->esc >= ISO_ESC_2_4_8 &&
4849 iso->esc <= ISO_ESC_2_4_15)
4851 chars = (iso->esc >= ISO_ESC_2_4_12) ? 96 : 94;
4852 single = -1; /* multi-byte */
4853 reg = (iso->esc - ISO_ESC_2_4_8) & 3;
4857 /* Can this ever be reached? -slb */
4861 cs = CHARSET_BY_ATTRIBUTES (chars, single, c,
4862 *flags & CODING_STATE_R2L ?
4863 CHARSET_RIGHT_TO_LEFT :
4864 CHARSET_LEFT_TO_RIGHT);
4870 iso->esc_bytes[iso->esc_bytes_index++] = (unsigned char) c;
4874 if (check_invalid_charsets && !CHARSETP (iso->charset[reg]))
4875 /* can't invoke something that ain't there. */
4877 iso->esc = ISO_ESC_SINGLE_SHIFT;
4878 *flags &= CODING_STATE_ISO2022_LOCK;
4880 *flags |= CODING_STATE_SS2;
4882 *flags |= CODING_STATE_SS3;
4886 if (check_invalid_charsets &&
4887 !CHARSETP (iso->charset[reg]))
4888 /* can't invoke something that ain't there. */
4891 iso->register_right = reg;
4893 iso->register_left = reg;
4894 *flags &= CODING_STATE_ISO2022_LOCK;
4895 iso->esc = ISO_ESC_LOCKING_SHIFT;
4899 if (NILP (cs) && check_invalid_charsets)
4901 iso->invalid_designated[reg] = 1;
4902 iso->charset[reg] = Vcharset_ascii;
4903 iso->esc = ISO_ESC_DESIGNATE;
4904 *flags &= CODING_STATE_ISO2022_LOCK;
4905 iso->output_literally = 1;
4906 if (iso->switched_dir_and_no_valid_charset_yet)
4908 /* We encountered a switch-direction followed by an
4909 invalid designation. Ensure that the switch-direction
4910 gets outputted; otherwise it will probably get eaten
4911 when the text is written out again. */
4912 iso->switched_dir_and_no_valid_charset_yet = 0;
4913 iso->output_direction_sequence = 1;
4914 /* And make sure that the switch-dir going the other
4915 way gets outputted, as well. */
4916 iso->invalid_switch_dir = 1;
4920 /* This function is called with CODESYS equal to nil when
4921 doing coding-system detection. */
4922 if (!NILP (codesys))
4924 charset_conversion_spec_dynarr *dyn =
4925 XCODING_SYSTEM (codesys)->iso2022.input_conv;
4931 for (i = 0; i < Dynarr_length (dyn); i++)
4933 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
4934 if (EQ (cs, spec->from_charset))
4935 cs = spec->to_charset;
4940 iso->charset[reg] = cs;
4941 iso->esc = ISO_ESC_DESIGNATE;
4942 *flags &= CODING_STATE_ISO2022_LOCK;
4943 if (iso->invalid_designated[reg])
4945 iso->invalid_designated[reg] = 0;
4946 iso->output_literally = 1;
4948 if (iso->switched_dir_and_no_valid_charset_yet)
4949 iso->switched_dir_and_no_valid_charset_yet = 0;
4954 detect_coding_iso2022 (struct detection_state *st, const Extbyte *src, size_t n)
4958 /* #### There are serious deficiencies in the recognition mechanism
4959 here. This needs to be much smarter if it's going to cut it.
4960 The sequence "\xff\x0f" is currently detected as LOCK_SHIFT while
4961 it should be detected as Latin-1.
4962 All the ISO2022 stuff in this file should be synced up with the
4963 code from FSF Emacs-20.4, in which Mule should be more or less stable.
4964 Perhaps we should wait till R2L works in FSF Emacs? */
4966 if (!st->iso2022.initted)
4968 reset_iso2022 (Qnil, &st->iso2022.iso);
4969 st->iso2022.mask = (CODING_CATEGORY_ISO_7_MASK |
4970 CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
4971 CODING_CATEGORY_ISO_8_1_MASK |
4972 CODING_CATEGORY_ISO_8_2_MASK |
4973 CODING_CATEGORY_ISO_LOCK_SHIFT_MASK);
4974 st->iso2022.flags = 0;
4975 st->iso2022.high_byte_count = 0;
4976 st->iso2022.saw_single_shift = 0;
4977 st->iso2022.initted = 1;
4980 mask = st->iso2022.mask;
4984 unsigned char c = *(unsigned char *)src++;
4987 mask &= ~CODING_CATEGORY_ISO_7_MASK;
4988 st->iso2022.high_byte_count++;
4992 if (st->iso2022.high_byte_count && !st->iso2022.saw_single_shift)
4994 if (st->iso2022.high_byte_count & 1)
4995 /* odd number of high bytes; assume not iso-8-2 */
4996 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
4998 st->iso2022.high_byte_count = 0;
4999 st->iso2022.saw_single_shift = 0;
5001 mask &= ~CODING_CATEGORY_ISO_7_MASK;
5003 if (!(st->iso2022.flags & CODING_STATE_ESCAPE)
5004 && (BYTE_C0_P (c) || BYTE_C1_P (c)))
5005 { /* control chars */
5008 /* Allow and ignore control characters that you might
5009 reasonably see in a text file */
5014 case 8: /* backspace */
5015 case 11: /* vertical tab */
5016 case 12: /* form feed */
5017 case 26: /* MS-DOS C-z junk */
5018 case 31: /* '^_' -- for info */
5019 goto label_continue_loop;
5026 if ((st->iso2022.flags & CODING_STATE_ESCAPE) || BYTE_C0_P (c)
5029 if (parse_iso2022_esc (Qnil, &st->iso2022.iso, c,
5030 &st->iso2022.flags, 0))
5032 switch (st->iso2022.iso.esc)
5034 case ISO_ESC_DESIGNATE:
5035 mask &= ~CODING_CATEGORY_ISO_8_1_MASK;
5036 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
5038 case ISO_ESC_LOCKING_SHIFT:
5039 mask = CODING_CATEGORY_ISO_LOCK_SHIFT_MASK;
5040 goto ran_out_of_chars;
5041 case ISO_ESC_SINGLE_SHIFT:
5042 mask &= ~CODING_CATEGORY_ISO_8_DESIGNATE_MASK;
5043 st->iso2022.saw_single_shift = 1;
5052 goto ran_out_of_chars;
5055 label_continue_loop:;
5064 postprocess_iso2022_mask (int mask)
5066 /* #### kind of cheesy */
5067 /* If seven-bit ISO is allowed, then assume that the encoding is
5068 entirely seven-bit and turn off the eight-bit ones. */
5069 if (mask & CODING_CATEGORY_ISO_7_MASK)
5070 mask &= ~ (CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
5071 CODING_CATEGORY_ISO_8_1_MASK |
5072 CODING_CATEGORY_ISO_8_2_MASK);
5076 /* If FLAGS is a null pointer or specifies right-to-left motion,
5077 output a switch-dir-to-left-to-right sequence to DST.
5078 Also update FLAGS if it is not a null pointer.
5079 If INTERNAL_P is set, we are outputting in internal format and
5080 need to handle the CSI differently. */
5083 restore_left_to_right_direction (Lisp_Coding_System *codesys,
5084 unsigned_char_dynarr *dst,
5085 unsigned int *flags,
5088 if (!flags || (*flags & CODING_STATE_R2L))
5090 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
5092 Dynarr_add (dst, ISO_CODE_ESC);
5093 Dynarr_add (dst, '[');
5095 else if (internal_p)
5096 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
5098 Dynarr_add (dst, ISO_CODE_CSI);
5099 Dynarr_add (dst, '0');
5100 Dynarr_add (dst, ']');
5102 *flags &= ~CODING_STATE_R2L;
5106 /* If FLAGS is a null pointer or specifies a direction different from
5107 DIRECTION (which should be either CHARSET_RIGHT_TO_LEFT or
5108 CHARSET_LEFT_TO_RIGHT), output the appropriate switch-dir escape
5109 sequence to DST. Also update FLAGS if it is not a null pointer.
5110 If INTERNAL_P is set, we are outputting in internal format and
5111 need to handle the CSI differently. */
5114 ensure_correct_direction (int direction, Lisp_Coding_System *codesys,
5115 unsigned_char_dynarr *dst, unsigned int *flags,
5118 if ((!flags || (*flags & CODING_STATE_R2L)) &&
5119 direction == CHARSET_LEFT_TO_RIGHT)
5120 restore_left_to_right_direction (codesys, dst, flags, internal_p);
5121 else if (!CODING_SYSTEM_ISO2022_NO_ISO6429 (codesys)
5122 && (!flags || !(*flags & CODING_STATE_R2L)) &&
5123 direction == CHARSET_RIGHT_TO_LEFT)
5125 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
5127 Dynarr_add (dst, ISO_CODE_ESC);
5128 Dynarr_add (dst, '[');
5130 else if (internal_p)
5131 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
5133 Dynarr_add (dst, ISO_CODE_CSI);
5134 Dynarr_add (dst, '2');
5135 Dynarr_add (dst, ']');
5137 *flags |= CODING_STATE_R2L;
5141 /* Convert ISO2022-format data to internal format. */
5144 decode_coding_iso2022 (Lstream *decoding, const Extbyte *src,
5145 unsigned_char_dynarr *dst, size_t n)
5147 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
5148 unsigned int flags = str->flags;
5149 unsigned int cpos = str->cpos;
5150 unsigned char counter = str->counter;
5151 eol_type_t eol_type = str->eol_type;
5152 #ifdef ENABLE_COMPOSITE_CHARS
5153 unsigned_char_dynarr *real_dst = dst;
5155 Lisp_Object coding_system;
5157 XSETCODING_SYSTEM (coding_system, str->codesys);
5159 #ifdef ENABLE_COMPOSITE_CHARS
5160 if (flags & CODING_STATE_COMPOSITE)
5161 dst = str->iso2022.composite_chars;
5162 #endif /* ENABLE_COMPOSITE_CHARS */
5166 unsigned char c = *(unsigned char *)src++;
5167 if (flags & CODING_STATE_ESCAPE)
5168 { /* Within ESC sequence */
5169 int retval = parse_iso2022_esc (coding_system, &str->iso2022,
5174 switch (str->iso2022.esc)
5176 #ifdef ENABLE_COMPOSITE_CHARS
5177 case ISO_ESC_START_COMPOSITE:
5178 if (str->iso2022.composite_chars)
5179 Dynarr_reset (str->iso2022.composite_chars);
5181 str->iso2022.composite_chars = Dynarr_new (unsigned_char);
5182 dst = str->iso2022.composite_chars;
5184 case ISO_ESC_END_COMPOSITE:
5186 Bufbyte comstr[MAX_EMCHAR_LEN];
5188 Emchar emch = lookup_composite_char (Dynarr_atp (dst, 0),
5189 Dynarr_length (dst));
5191 len = set_charptr_emchar (comstr, emch);
5192 Dynarr_add_many (dst, comstr, len);
5195 #endif /* ENABLE_COMPOSITE_CHARS */
5197 case ISO_ESC_LITERAL:
5198 COMPOSE_FLUSH_CHARS (str, dst);
5199 decode_flush_er_chars (str, dst);
5200 DECODE_ADD_BINARY_CHAR (c, dst);
5204 /* Everything else handled already */
5209 /* Attempted error recovery. */
5210 if (str->iso2022.output_direction_sequence)
5211 ensure_correct_direction (flags & CODING_STATE_R2L ?
5212 CHARSET_RIGHT_TO_LEFT :
5213 CHARSET_LEFT_TO_RIGHT,
5214 str->codesys, dst, 0, 1);
5215 /* More error recovery. */
5216 if (!retval || str->iso2022.output_literally)
5218 /* Output the (possibly invalid) sequence */
5220 COMPOSE_FLUSH_CHARS (str, dst);
5221 decode_flush_er_chars (str, dst);
5222 for (i = 0; i < str->iso2022.esc_bytes_index; i++)
5223 DECODE_ADD_BINARY_CHAR (str->iso2022.esc_bytes[i], dst);
5224 flags &= CODING_STATE_ISO2022_LOCK;
5226 n++, src--;/* Repeat the loop with the same character. */
5229 /* No sense in reprocessing the final byte of the
5230 escape sequence; it could mess things up anyway.
5232 COMPOSE_FLUSH_CHARS (str, dst);
5233 decode_flush_er_chars (str, dst);
5234 DECODE_ADD_BINARY_CHAR (c, dst);
5240 else if (BYTE_C0_P (c) || BYTE_C1_P (c))
5241 { /* Control characters */
5243 /***** Error-handling *****/
5245 /* If we were in the middle of a character, dump out the
5246 partial character. */
5249 COMPOSE_FLUSH_CHARS (str, dst);
5250 decode_flush_er_chars (str, dst);
5254 DECODE_ADD_BINARY_CHAR
5255 ((unsigned char)(cpos >> (counter * 8)), dst);
5260 /* If we just saw a single-shift character, dump it out.
5261 This may dump out the wrong sort of single-shift character,
5262 but least it will give an indication that something went
5264 if (flags & CODING_STATE_SS2)
5266 COMPOSE_FLUSH_CHARS (str, dst);
5267 decode_flush_er_chars (str, dst);
5268 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS2, dst);
5269 flags &= ~CODING_STATE_SS2;
5271 if (flags & CODING_STATE_SS3)
5273 COMPOSE_FLUSH_CHARS (str, dst);
5274 decode_flush_er_chars (str, dst);
5275 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS3, dst);
5276 flags &= ~CODING_STATE_SS3;
5279 /***** Now handle the control characters. *****/
5285 COMPOSE_FLUSH_CHARS (str, dst);
5286 decode_flush_er_chars (str, dst);
5287 if (eol_type == EOL_CR)
5288 Dynarr_add (dst, '\n');
5289 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR)
5290 Dynarr_add (dst, c);
5292 flags |= CODING_STATE_CR;
5293 goto label_continue_loop;
5295 else if (flags & CODING_STATE_CR)
5296 { /* eol_type == CODING_SYSTEM_EOL_CRLF */
5298 Dynarr_add (dst, '\r');
5299 flags &= ~CODING_STATE_CR;
5302 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
5305 flags &= CODING_STATE_ISO2022_LOCK;
5307 if (!parse_iso2022_esc (coding_system, &str->iso2022, c, &flags, 1))
5309 COMPOSE_FLUSH_CHARS (str, dst);
5310 decode_flush_er_chars (str, dst);
5311 DECODE_ADD_BINARY_CHAR (c, dst);
5315 { /* Graphic characters */
5316 Lisp_Object charset;
5325 COMPOSE_FLUSH_CHARS (str, dst);
5326 decode_flush_er_chars (str, dst);
5327 if (eol_type == EOL_CR)
5328 Dynarr_add (dst, '\n');
5329 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR)
5330 Dynarr_add (dst, c);
5332 flags |= CODING_STATE_CR;
5333 goto label_continue_loop;
5335 else if (flags & CODING_STATE_CR)
5336 { /* eol_type == CODING_SYSTEM_EOL_CRLF */
5338 Dynarr_add (dst, '\r');
5339 flags &= ~CODING_STATE_CR;
5342 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
5345 /* Now determine the charset. */
5346 reg = ((flags & CODING_STATE_SS2) ? 2
5347 : (flags & CODING_STATE_SS3) ? 3
5348 : !BYTE_ASCII_P (c) ? str->iso2022.register_right
5349 : str->iso2022.register_left);
5350 charset = str->iso2022.charset[reg];
5352 /* Error checking: */
5353 if (! CHARSETP (charset)
5354 || str->iso2022.invalid_designated[reg]
5355 || (((c & 0x7F) == ' ' || (c & 0x7F) == ISO_CODE_DEL)
5356 && XCHARSET_CHARS (charset) == 94))
5357 /* Mrmph. We are trying to invoke a register that has no
5358 or an invalid charset in it, or trying to add a character
5359 outside the range of the charset. Insert that char literally
5360 to preserve it for the output. */
5362 COMPOSE_FLUSH_CHARS (str, dst);
5363 decode_flush_er_chars (str, dst);
5367 DECODE_ADD_BINARY_CHAR
5368 ((unsigned char)(cpos >> (counter * 8)), dst);
5371 DECODE_ADD_BINARY_CHAR (c, dst);
5376 /* Things are probably hunky-dorey. */
5378 /* Fetch reverse charset, maybe. */
5379 if (((flags & CODING_STATE_R2L) &&
5380 XCHARSET_DIRECTION (charset) == CHARSET_LEFT_TO_RIGHT)
5382 (!(flags & CODING_STATE_R2L) &&
5383 XCHARSET_DIRECTION (charset) == CHARSET_RIGHT_TO_LEFT))
5385 Lisp_Object new_charset =
5386 XCHARSET_REVERSE_DIRECTION_CHARSET (charset);
5387 if (!NILP (new_charset))
5388 charset = new_charset;
5393 if (XCHARSET_DIMENSION (charset) == counter)
5395 COMPOSE_ADD_CHAR (str,
5396 DECODE_CHAR (charset,
5397 ((cpos & 0x7F7F7F) << 8)
5404 cpos = (cpos << 8) | c;
5406 lb = XCHARSET_LEADING_BYTE (charset);
5407 switch (XCHARSET_REP_BYTES (charset))
5410 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5411 Dynarr_add (dst, c & 0x7F);
5414 case 2: /* one-byte official */
5415 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5416 Dynarr_add (dst, lb);
5417 Dynarr_add (dst, c | 0x80);
5420 case 3: /* one-byte private or two-byte official */
5421 if (XCHARSET_PRIVATE_P (charset))
5423 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5424 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_1);
5425 Dynarr_add (dst, lb);
5426 Dynarr_add (dst, c | 0x80);
5432 Dynarr_add (dst, lb);
5433 Dynarr_add (dst, ch | 0x80);
5434 Dynarr_add (dst, c | 0x80);
5442 default: /* two-byte private */
5445 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_2);
5446 Dynarr_add (dst, lb);
5447 Dynarr_add (dst, ch | 0x80);
5448 Dynarr_add (dst, c | 0x80);
5458 flags &= CODING_STATE_ISO2022_LOCK;
5461 label_continue_loop:;
5464 if (flags & CODING_STATE_END)
5466 COMPOSE_FLUSH_CHARS (str, dst);
5467 decode_flush_er_chars (str, dst);
5468 DECODE_OUTPUT_PARTIAL_CHAR (cpos);
5472 str->counter = counter;
5476 /***** ISO2022 encoder *****/
5478 /* Designate CHARSET into register REG. */
5481 iso2022_designate (Lisp_Object charset, unsigned char reg,
5482 struct encoding_stream *str, unsigned_char_dynarr *dst)
5484 static const char inter94[] = "()*+";
5485 static const char inter96[] = ",-./";
5486 unsigned short chars;
5487 unsigned char dimension;
5488 unsigned char final;
5489 Lisp_Object old_charset = str->iso2022.charset[reg];
5491 str->iso2022.charset[reg] = charset;
5492 if (!CHARSETP (charset))
5493 /* charset might be an initial nil or t. */
5495 chars = XCHARSET_CHARS (charset);
5496 dimension = XCHARSET_DIMENSION (charset);
5497 final = XCHARSET_FINAL (charset);
5498 if (!str->iso2022.force_charset_on_output[reg] &&
5499 CHARSETP (old_charset) &&
5500 XCHARSET_CHARS (old_charset) == chars &&
5501 XCHARSET_DIMENSION (old_charset) == dimension &&
5502 XCHARSET_FINAL (old_charset) == final)
5505 str->iso2022.force_charset_on_output[reg] = 0;
5508 charset_conversion_spec_dynarr *dyn =
5509 str->codesys->iso2022.output_conv;
5515 for (i = 0; i < Dynarr_length (dyn); i++)
5517 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
5518 if (EQ (charset, spec->from_charset))
5519 charset = spec->to_charset;
5524 Dynarr_add (dst, ISO_CODE_ESC);
5529 Dynarr_add (dst, inter94[reg]);
5532 Dynarr_add (dst, '$');
5534 || !(CODING_SYSTEM_ISO2022_SHORT (str->codesys))
5537 Dynarr_add (dst, inter94[reg]);
5542 Dynarr_add (dst, inter96[reg]);
5545 Dynarr_add (dst, '$');
5546 Dynarr_add (dst, inter96[reg]);
5550 Dynarr_add (dst, final);
5554 ensure_normal_shift (struct encoding_stream *str, unsigned_char_dynarr *dst)
5556 if (str->iso2022.register_left != 0)
5558 Dynarr_add (dst, ISO_CODE_SI);
5559 str->iso2022.register_left = 0;
5564 ensure_shift_out (struct encoding_stream *str, unsigned_char_dynarr *dst)
5566 if (str->iso2022.register_left != 1)
5568 Dynarr_add (dst, ISO_CODE_SO);
5569 str->iso2022.register_left = 1;
5574 char_encode_iso2022 (struct encoding_stream *str, Emchar ch,
5575 unsigned_char_dynarr *dst, unsigned int *flags)
5577 unsigned char charmask;
5578 Lisp_Coding_System* codesys = str->codesys;
5579 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
5581 Lisp_Object charset = str->iso2022.current_charset;
5582 int half = str->iso2022.current_half;
5583 int code_point = -1;
5587 restore_left_to_right_direction (codesys, dst, flags, 0);
5589 /* Make sure G0 contains ASCII */
5590 if ((ch > ' ' && ch < ISO_CODE_DEL)
5591 || !CODING_SYSTEM_ISO2022_NO_ASCII_CNTL (codesys))
5593 ensure_normal_shift (str, dst);
5594 iso2022_designate (Vcharset_ascii, 0, str, dst);
5597 /* If necessary, restore everything to the default state
5599 if (ch == '\n' && !(CODING_SYSTEM_ISO2022_NO_ASCII_EOL (codesys)))
5601 restore_left_to_right_direction (codesys, dst, flags, 0);
5603 ensure_normal_shift (str, dst);
5605 for (i = 0; i < 4; i++)
5607 Lisp_Object initial_charset =
5608 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
5609 iso2022_designate (initial_charset, i, str, dst);
5614 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5615 Dynarr_add (dst, '\r');
5616 if (eol_type != EOL_CR)
5617 Dynarr_add (dst, ch);
5621 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
5622 && fit_to_be_escape_quoted (ch))
5623 Dynarr_add (dst, ISO_CODE_ESC);
5624 Dynarr_add (dst, ch);
5627 else if ( (0x80 <= ch) && (ch <= 0x9f) )
5629 charmask = (half == 0 ? 0x00 : 0x80);
5631 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
5632 && fit_to_be_escape_quoted (ch))
5633 Dynarr_add (dst, ISO_CODE_ESC);
5634 /* you asked for it ... */
5635 Dynarr_add (dst, ch);
5641 /* Now determine which register to use. */
5643 for (i = 0; i < 4; i++)
5645 if ((CHARSETP (charset = str->iso2022.charset[i])
5646 && ((code_point = charset_code_point (charset, ch)) >= 0))
5650 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i))
5651 && ((code_point = charset_code_point (charset, ch)) >= 0)))
5659 Lisp_Object original_default_coded_charset_priority_list
5660 = Vdefault_coded_charset_priority_list;
5662 while (!EQ (Vdefault_coded_charset_priority_list, Qnil))
5664 code_point = ENCODE_CHAR (ch, charset);
5665 if (XCHARSET_FINAL (charset))
5667 Vdefault_coded_charset_priority_list
5668 = Fcdr (Fmemq (XCHARSET_NAME (charset),
5669 Vdefault_coded_charset_priority_list));
5671 code_point = ENCODE_CHAR (ch, charset);
5672 if (!XCHARSET_FINAL (charset))
5674 charset = Vcharset_ascii;
5678 Vdefault_coded_charset_priority_list
5679 = original_default_coded_charset_priority_list;
5681 ensure_correct_direction (XCHARSET_DIRECTION (charset),
5682 codesys, dst, flags, 0);
5686 if (XCHARSET_GRAPHIC (charset) != 0)
5688 if (!NILP (str->iso2022.charset[1]) &&
5689 (!CODING_SYSTEM_ISO2022_SEVEN (codesys)
5690 || CODING_SYSTEM_ISO2022_LOCK_SHIFT (codesys)))
5692 else if (!NILP (str->iso2022.charset[2]))
5694 else if (!NILP (str->iso2022.charset[3]))
5703 iso2022_designate (charset, reg, str, dst);
5705 /* Now invoke that register. */
5709 ensure_normal_shift (str, dst);
5713 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
5715 ensure_shift_out (str, dst);
5722 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
5724 Dynarr_add (dst, ISO_CODE_ESC);
5725 Dynarr_add (dst, 'N');
5730 Dynarr_add (dst, ISO_CODE_SS2);
5735 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
5737 Dynarr_add (dst, ISO_CODE_ESC);
5738 Dynarr_add (dst, 'O');
5743 Dynarr_add (dst, ISO_CODE_SS3);
5751 charmask = (half == 0 ? 0x00 : 0x80);
5753 switch (XCHARSET_DIMENSION (charset))
5756 Dynarr_add (dst, (code_point & 0xFF) | charmask);
5759 Dynarr_add (dst, ((code_point >> 8) & 0xFF) | charmask);
5760 Dynarr_add (dst, ( code_point & 0xFF) | charmask);
5763 Dynarr_add (dst, ((code_point >> 16) & 0xFF) | charmask);
5764 Dynarr_add (dst, ((code_point >> 8) & 0xFF) | charmask);
5765 Dynarr_add (dst, ( code_point & 0xFF) | charmask);
5768 Dynarr_add (dst, ((code_point >> 24) & 0xFF) | charmask);
5769 Dynarr_add (dst, ((code_point >> 16) & 0xFF) | charmask);
5770 Dynarr_add (dst, ((code_point >> 8) & 0xFF) | charmask);
5771 Dynarr_add (dst, ( code_point & 0xFF) | charmask);
5777 str->iso2022.current_charset = charset;
5778 str->iso2022.current_half = half;
5782 char_finish_iso2022 (struct encoding_stream *str, unsigned_char_dynarr *dst,
5783 unsigned int *flags)
5785 Lisp_Coding_System* codesys = str->codesys;
5788 restore_left_to_right_direction (codesys, dst, flags, 0);
5789 ensure_normal_shift (str, dst);
5790 for (i = 0; i < 4; i++)
5792 Lisp_Object initial_charset
5793 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
5794 iso2022_designate (initial_charset, i, str, dst);
5799 /************************************************************************/
5800 /* No-conversion methods */
5801 /************************************************************************/
5803 /* This is used when reading in "binary" files -- i.e. files that may
5804 contain all 256 possible byte values and that are not to be
5805 interpreted as being in any particular decoding. */
5807 decode_coding_no_conversion (Lstream *decoding, const Extbyte *src,
5808 unsigned_char_dynarr *dst, size_t n)
5810 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
5811 unsigned int flags = str->flags;
5812 unsigned int cpos = str->cpos;
5813 eol_type_t eol_type = str->eol_type;
5817 unsigned char c = *(unsigned char *)src++;
5819 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
5820 DECODE_ADD_BINARY_CHAR (c, dst);
5821 label_continue_loop:;
5824 DECODE_HANDLE_END_OF_CONVERSION (flags, cpos, dst);
5831 encode_coding_no_conversion (Lstream *encoding, const Bufbyte *src,
5832 unsigned_char_dynarr *dst, size_t n)
5835 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
5836 unsigned int flags = str->flags;
5837 unsigned int ch = str->ch;
5838 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
5840 unsigned char char_boundary = str->iso2022.current_char_boundary;
5847 if (char_boundary == 0)
5853 else if ( c >= 0xf8 )
5858 else if ( c >= 0xf0 )
5863 else if ( c >= 0xe0 )
5868 else if ( c >= 0xc0 )
5878 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5879 Dynarr_add (dst, '\r');
5880 if (eol_type != EOL_CR)
5881 Dynarr_add (dst, c);
5884 Dynarr_add (dst, c);
5887 else if (char_boundary == 1)
5889 ch = ( ch << 6 ) | ( c & 0x3f );
5890 Dynarr_add (dst, ch & 0xff);
5895 ch = ( ch << 6 ) | ( c & 0x3f );
5898 #else /* not UTF2000 */
5901 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5902 Dynarr_add (dst, '\r');
5903 if (eol_type != EOL_CR)
5904 Dynarr_add (dst, '\n');
5907 else if (BYTE_ASCII_P (c))
5910 Dynarr_add (dst, c);
5912 else if (BUFBYTE_LEADING_BYTE_P (c))
5915 if (c == LEADING_BYTE_LATIN_ISO8859_1 ||
5916 c == LEADING_BYTE_CONTROL_1)
5919 Dynarr_add (dst, '~'); /* untranslatable character */
5923 if (ch == LEADING_BYTE_LATIN_ISO8859_1)
5924 Dynarr_add (dst, c);
5925 else if (ch == LEADING_BYTE_CONTROL_1)
5928 Dynarr_add (dst, c - 0x20);
5930 /* else it should be the second or third byte of an
5931 untranslatable character, so ignore it */
5934 #endif /* not UTF2000 */
5940 str->iso2022.current_char_boundary = char_boundary;
5946 /************************************************************************/
5947 /* Initialization */
5948 /************************************************************************/
5951 syms_of_file_coding (void)
5953 INIT_LRECORD_IMPLEMENTATION (coding_system);
5955 deferror (&Qcoding_system_error, "coding-system-error",
5956 "Coding-system error", Qio_error);
5958 DEFSUBR (Fcoding_system_p);
5959 DEFSUBR (Ffind_coding_system);
5960 DEFSUBR (Fget_coding_system);
5961 DEFSUBR (Fcoding_system_list);
5962 DEFSUBR (Fcoding_system_name);
5963 DEFSUBR (Fmake_coding_system);
5964 DEFSUBR (Fcopy_coding_system);
5965 DEFSUBR (Fcoding_system_canonical_name_p);
5966 DEFSUBR (Fcoding_system_alias_p);
5967 DEFSUBR (Fcoding_system_aliasee);
5968 DEFSUBR (Fdefine_coding_system_alias);
5969 DEFSUBR (Fsubsidiary_coding_system);
5971 DEFSUBR (Fcoding_system_type);
5972 DEFSUBR (Fcoding_system_doc_string);
5974 DEFSUBR (Fcoding_system_charset);
5976 DEFSUBR (Fcoding_system_property);
5978 DEFSUBR (Fcoding_category_list);
5979 DEFSUBR (Fset_coding_priority_list);
5980 DEFSUBR (Fcoding_priority_list);
5981 DEFSUBR (Fset_coding_category_system);
5982 DEFSUBR (Fcoding_category_system);
5984 DEFSUBR (Fdetect_coding_region);
5985 DEFSUBR (Fdecode_coding_region);
5986 DEFSUBR (Fencode_coding_region);
5988 DEFSUBR (Fdecode_shift_jis_char);
5989 DEFSUBR (Fencode_shift_jis_char);
5990 DEFSUBR (Fdecode_big5_char);
5991 DEFSUBR (Fencode_big5_char);
5993 defsymbol (&Qcoding_systemp, "coding-system-p");
5994 defsymbol (&Qno_conversion, "no-conversion");
5995 defsymbol (&Qraw_text, "raw-text");
5997 defsymbol (&Qbig5, "big5");
5998 defsymbol (&Qshift_jis, "shift-jis");
5999 defsymbol (&Qucs4, "ucs-4");
6000 defsymbol (&Qutf8, "utf-8");
6001 defsymbol (&Qccl, "ccl");
6002 defsymbol (&Qiso2022, "iso2022");
6004 defsymbol (&Qmnemonic, "mnemonic");
6005 defsymbol (&Qeol_type, "eol-type");
6006 defsymbol (&Qpost_read_conversion, "post-read-conversion");
6007 defsymbol (&Qpre_write_conversion, "pre-write-conversion");
6009 defsymbol (&Qcr, "cr");
6010 defsymbol (&Qlf, "lf");
6011 defsymbol (&Qcrlf, "crlf");
6012 defsymbol (&Qeol_cr, "eol-cr");
6013 defsymbol (&Qeol_lf, "eol-lf");
6014 defsymbol (&Qeol_crlf, "eol-crlf");
6016 defsymbol (&Qcharset_g0, "charset-g0");
6017 defsymbol (&Qcharset_g1, "charset-g1");
6018 defsymbol (&Qcharset_g2, "charset-g2");
6019 defsymbol (&Qcharset_g3, "charset-g3");
6020 defsymbol (&Qforce_g0_on_output, "force-g0-on-output");
6021 defsymbol (&Qforce_g1_on_output, "force-g1-on-output");
6022 defsymbol (&Qforce_g2_on_output, "force-g2-on-output");
6023 defsymbol (&Qforce_g3_on_output, "force-g3-on-output");
6024 defsymbol (&Qno_iso6429, "no-iso6429");
6025 defsymbol (&Qinput_charset_conversion, "input-charset-conversion");
6026 defsymbol (&Qoutput_charset_conversion, "output-charset-conversion");
6028 defsymbol (&Qshort, "short");
6029 defsymbol (&Qno_ascii_eol, "no-ascii-eol");
6030 defsymbol (&Qno_ascii_cntl, "no-ascii-cntl");
6031 defsymbol (&Qseven, "seven");
6032 defsymbol (&Qlock_shift, "lock-shift");
6033 defsymbol (&Qescape_quoted, "escape-quoted");
6036 defsymbol (&Qdisable_composition, "disable-composition");
6037 defsymbol (&Quse_entity_reference, "use-entity-reference");
6038 defsymbol (&Qd, "d");
6039 defsymbol (&Qx, "x");
6040 defsymbol (&QX, "X");
6042 defsymbol (&Qencode, "encode");
6043 defsymbol (&Qdecode, "decode");
6046 defsymbol (&coding_category_symbol[CODING_CATEGORY_SHIFT_JIS],
6048 defsymbol (&coding_category_symbol[CODING_CATEGORY_BIG5],
6050 defsymbol (&coding_category_symbol[CODING_CATEGORY_UCS4],
6052 defsymbol (&coding_category_symbol[CODING_CATEGORY_UTF8],
6054 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_7],
6056 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_DESIGNATE],
6058 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_1],
6060 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_2],
6062 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_LOCK_SHIFT],
6065 defsymbol (&coding_category_symbol[CODING_CATEGORY_NO_CONVERSION],
6070 lstream_type_create_file_coding (void)
6072 LSTREAM_HAS_METHOD (decoding, reader);
6073 LSTREAM_HAS_METHOD (decoding, writer);
6074 LSTREAM_HAS_METHOD (decoding, rewinder);
6075 LSTREAM_HAS_METHOD (decoding, seekable_p);
6076 LSTREAM_HAS_METHOD (decoding, flusher);
6077 LSTREAM_HAS_METHOD (decoding, closer);
6078 LSTREAM_HAS_METHOD (decoding, marker);
6080 LSTREAM_HAS_METHOD (encoding, reader);
6081 LSTREAM_HAS_METHOD (encoding, writer);
6082 LSTREAM_HAS_METHOD (encoding, rewinder);
6083 LSTREAM_HAS_METHOD (encoding, seekable_p);
6084 LSTREAM_HAS_METHOD (encoding, flusher);
6085 LSTREAM_HAS_METHOD (encoding, closer);
6086 LSTREAM_HAS_METHOD (encoding, marker);
6090 vars_of_file_coding (void)
6094 fcd = xnew (struct file_coding_dump);
6095 dump_add_root_struct_ptr (&fcd, &fcd_description);
6097 /* Initialize to something reasonable ... */
6098 for (i = 0; i < CODING_CATEGORY_LAST; i++)
6100 fcd->coding_category_system[i] = Qnil;
6101 fcd->coding_category_by_priority[i] = i;
6104 Fprovide (intern ("file-coding"));
6106 DEFVAR_LISP ("keyboard-coding-system", &Vkeyboard_coding_system /*
6107 Coding system used for TTY keyboard input.
6108 Not used under a windowing system.
6110 Vkeyboard_coding_system = Qnil;
6112 DEFVAR_LISP ("terminal-coding-system", &Vterminal_coding_system /*
6113 Coding system used for TTY display output.
6114 Not used under a windowing system.
6116 Vterminal_coding_system = Qnil;
6118 DEFVAR_LISP ("coding-system-for-read", &Vcoding_system_for_read /*
6119 Overriding coding system used when reading from a file or process.
6120 You should bind this variable with `let', but do not set it globally.
6121 If this is non-nil, it specifies the coding system that will be used
6122 to decode input on read operations, such as from a file or process.
6123 It overrides `buffer-file-coding-system-for-read',
6124 `insert-file-contents-pre-hook', etc. Use those variables instead of
6125 this one for permanent changes to the environment. */ );
6126 Vcoding_system_for_read = Qnil;
6128 DEFVAR_LISP ("coding-system-for-write",
6129 &Vcoding_system_for_write /*
6130 Overriding coding system used when writing to a file or process.
6131 You should bind this variable with `let', but do not set it globally.
6132 If this is non-nil, it specifies the coding system that will be used
6133 to encode output for write operations, such as to a file or process.
6134 It overrides `buffer-file-coding-system', `write-region-pre-hook', etc.
6135 Use those variables instead of this one for permanent changes to the
6137 Vcoding_system_for_write = Qnil;
6139 DEFVAR_LISP ("file-name-coding-system", &Vfile_name_coding_system /*
6140 Coding system used to convert pathnames when accessing files.
6142 Vfile_name_coding_system = Qnil;
6144 DEFVAR_LISP ("coded-charset-entity-reference-alist",
6145 &Vcoded_charset_entity_reference_alist /*
6146 Alist of coded-charset vs corresponding entity-reference.
6147 Each element looks like (CCS PREFIX CODE-COLUMNS CODE-TYPE).
6148 CCS is coded-charset.
6149 CODE-COLUMNS is columns of code-point of entity-reference.
6150 CODE-TYPE is format type of code-point of entity-reference.
6151 `d' means decimal value and `x' means hexadecimal value.
6153 Vcoded_charset_entity_reference_alist = Qnil;
6155 DEFVAR_BOOL ("enable-multibyte-characters", &enable_multibyte_characters /*
6156 Non-nil means the buffer contents are regarded as multi-byte form
6157 of characters, not a binary code. This affects the display, file I/O,
6158 and behaviors of various editing commands.
6160 Setting this to nil does not do anything.
6162 enable_multibyte_characters = 1;
6166 complex_vars_of_file_coding (void)
6168 staticpro (&Vcoding_system_hash_table);
6169 Vcoding_system_hash_table =
6170 make_lisp_hash_table (50, HASH_TABLE_NON_WEAK, HASH_TABLE_EQ);
6172 the_codesys_prop_dynarr = Dynarr_new (codesys_prop);
6173 dump_add_root_struct_ptr (&the_codesys_prop_dynarr, &codesys_prop_dynarr_description);
6175 #define DEFINE_CODESYS_PROP(Prop_Type, Sym) do \
6177 struct codesys_prop csp; \
6179 csp.prop_type = (Prop_Type); \
6180 Dynarr_add (the_codesys_prop_dynarr, csp); \
6183 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qmnemonic);
6184 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_type);
6185 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_cr);
6186 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_crlf);
6187 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_lf);
6188 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpost_read_conversion);
6189 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpre_write_conversion);
6191 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g0);
6192 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g1);
6193 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g2);
6194 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g3);
6195 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g0_on_output);
6196 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g1_on_output);
6197 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g2_on_output);
6198 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g3_on_output);
6199 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qshort);
6200 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_eol);
6201 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_cntl);
6202 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qseven);
6203 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qlock_shift);
6204 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_iso6429);
6205 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qescape_quoted);
6206 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qinput_charset_conversion);
6207 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qoutput_charset_conversion);
6209 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qencode);
6210 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qdecode);
6212 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qdisable_composition);
6213 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Quse_entity_reference);
6216 /* Need to create this here or we're really screwed. */
6218 (Qraw_text, Qno_conversion,
6219 build_string ("Raw text, which means it converts only line-break-codes."),
6220 list2 (Qmnemonic, build_string ("Raw")));
6223 (Qbinary, Qno_conversion,
6224 build_string ("Binary, which means it does not convert anything."),
6225 list4 (Qeol_type, Qlf,
6226 Qmnemonic, build_string ("Binary")));
6231 build_string ("Coding-system of ISO/IEC 10646 UTF-8."),
6232 list2 (Qmnemonic, build_string ("UTF8")));
6235 Fdefine_coding_system_alias (Qno_conversion, Qraw_text);
6237 Fdefine_coding_system_alias (Qfile_name, Qbinary);
6239 Fdefine_coding_system_alias (Qterminal, Qbinary);
6240 Fdefine_coding_system_alias (Qkeyboard, Qbinary);
6242 /* Need this for bootstrapping */
6243 fcd->coding_category_system[CODING_CATEGORY_NO_CONVERSION] =
6244 Fget_coding_system (Qraw_text);
6247 fcd->coding_category_system[CODING_CATEGORY_UTF8]
6248 = Fget_coding_system (Qutf8);
6251 #if defined(MULE) && !defined(UTF2000)
6255 for (i = 0; i < countof (fcd->ucs_to_mule_table); i++)
6256 fcd->ucs_to_mule_table[i] = Qnil;
6258 staticpro (&mule_to_ucs_table);
6259 mule_to_ucs_table = Fmake_char_table(Qgeneric);
6260 #endif /* defined(MULE) && !defined(UTF2000) */