1 /* Code conversion functions.
2 Copyright (C) 1991, 1995 Free Software Foundation, Inc.
3 Copyright (C) 1995 Sun Microsystems, Inc.
5 This file is part of SXEmacs
7 SXEmacs is free software: you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation, either version 3 of the License, or
10 (at your option) any later version.
12 SXEmacs is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with this program. If not, see <http://www.gnu.org/licenses/>. */
21 /* Synched up with: Mule 2.3. Not in FSF. */
23 /* Rewritten by Ben Wing <ben@xemacs.org>. */
30 #include "ui/insdel.h"
37 #include "file-coding.h"
39 Lisp_Object Qcoding_system_error;
41 Lisp_Object Vkeyboard_coding_system;
42 Lisp_Object Vterminal_coding_system;
43 Lisp_Object Vcoding_system_for_read;
44 Lisp_Object Vcoding_system_for_write;
45 Lisp_Object Vfile_name_coding_system;
47 /* Table of symbols identifying each coding category. */
48 Lisp_Object coding_category_symbol[CODING_CATEGORY_LAST];
50 struct file_coding_dump {
51 /* Coding system currently associated with each coding category. */
52 Lisp_Object coding_category_system[CODING_CATEGORY_LAST];
54 /* Table of all coding categories in decreasing order of priority.
55 This describes a permutation of the possible coding categories. */
56 int coding_category_by_priority[CODING_CATEGORY_LAST];
59 Lisp_Object ucs_to_mule_table[65536];
63 static const struct lrecord_description fcd_description_1[] = {
64 {XD_LISP_OBJECT_ARRAY,
65 offsetof(struct file_coding_dump, coding_category_system),
66 CODING_CATEGORY_LAST},
68 {XD_LISP_OBJECT_ARRAY,
69 offsetof(struct file_coding_dump, ucs_to_mule_table),
70 countof(fcd->ucs_to_mule_table)},
75 static const struct struct_description fcd_description = {
76 sizeof(struct file_coding_dump),
80 Lisp_Object mule_to_ucs_table;
82 Lisp_Object Qcoding_systemp;
84 Lisp_Object Qraw_text, Qno_conversion, Qccl, Qiso2022;
85 /* Qinternal in general.c */
87 Lisp_Object Qmnemonic, Qeol_type;
88 Lisp_Object Qcr, Qcrlf, Qlf;
89 Lisp_Object Qeol_cr, Qeol_crlf, Qeol_lf;
90 Lisp_Object Qpost_read_conversion;
91 Lisp_Object Qpre_write_conversion;
94 Lisp_Object Qucs4, Qutf8;
95 Lisp_Object Qbig5, Qshift_jis;
96 Lisp_Object Qcharset_g0, Qcharset_g1, Qcharset_g2, Qcharset_g3;
97 Lisp_Object Qforce_g0_on_output, Qforce_g1_on_output;
98 Lisp_Object Qforce_g2_on_output, Qforce_g3_on_output;
99 Lisp_Object Qno_iso6429;
100 Lisp_Object Qinput_charset_conversion, Qoutput_charset_conversion;
101 Lisp_Object Qescape_quoted;
102 Lisp_Object Qno_ascii_eol, Qno_ascii_cntl, Qseven, Qlock_shift;
104 Lisp_Object Qencode, Qdecode;
106 Lisp_Object Vcoding_system_hash_table;
108 int enable_multibyte_characters;
111 /* Additional information used by the ISO2022 decoder and detector. */
112 struct iso2022_decoder {
113 /* CHARSET holds the character sets currently assigned to the G0
114 through G3 variables. It is initialized from the array
115 INITIAL_CHARSET in CODESYS. */
116 Lisp_Object charset[4];
118 /* Which registers are currently invoked into the left (GL) and
119 right (GR) halves of the 8-bit encoding space? */
120 int register_left, register_right;
122 /* ISO_ESC holds a value indicating part of an escape sequence
123 that has already been seen. */
124 enum iso_esc_flag esc;
126 /* This records the bytes we've seen so far in an escape sequence,
127 in case the sequence is invalid (we spit out the bytes unchanged). */
128 unsigned char esc_bytes[8];
130 /* Index for next byte to store in ISO escape sequence. */
133 #ifdef ENABLE_COMPOSITE_CHARS
134 /* Stuff seen so far when composing a string. */
135 unsigned_char_dynarr *composite_chars;
138 /* If we saw an invalid designation sequence for a particular
139 register, we flag it here and switch to ASCII. The next time we
140 see a valid designation for this register, we turn off the flag
141 and do the designation normally, but pretend the sequence was
142 invalid. The effect of all this is that (most of the time) the
143 escape sequences for both the switch to the unknown charset, and
144 the switch back to the known charset, get inserted literally into
145 the buffer and saved out as such. The hope is that we can
146 preserve the escape sequences so that the resulting written out
147 file makes sense. If we don't do any of this, the designation
148 to the invalid charset will be preserved but that switch back
149 to the known charset will probably get eaten because it was
150 the same charset that was already present in the register. */
151 unsigned char invalid_designated[4];
153 /* We try to do similar things as above for direction-switching
154 sequences. If we encountered a direction switch while an
155 invalid designation was present, or an invalid designation
156 just after a direction switch (i.e. no valid designation
157 encountered yet), we insert the direction-switch escape
158 sequence literally into the output stream, and later on
159 insert the corresponding direction-restoring escape sequence
161 unsigned int switched_dir_and_no_valid_charset_yet:1;
162 unsigned int invalid_switch_dir:1;
164 /* Tells the decoder to output the escape sequence literally
165 even though it was valid. Used in the games we play to
166 avoid lossage when we encounter invalid designations. */
167 unsigned int output_literally:1;
168 /* We encountered a direction switch followed by an invalid
169 designation. We didn't output the direction switch
170 literally because we didn't know about the invalid designation;
171 but we have to do so now. */
172 unsigned int output_direction_sequence:1;
175 EXFUN(Fcopy_coding_system, 2);
177 struct detection_state;
178 static int detect_coding_sjis(struct detection_state *st,
179 const Extbyte * src, Lstream_data_count n);
180 static void decode_coding_sjis(lstream_t decoding, const Extbyte * src,
181 unsigned_char_dynarr * dst,
182 Lstream_data_count n);
183 static void encode_coding_sjis(lstream_t encoding, const Bufbyte * src,
184 unsigned_char_dynarr * dst,
185 Lstream_data_count n);
186 static int detect_coding_big5(struct detection_state *st, const Extbyte * src,
187 Lstream_data_count n);
188 static void decode_coding_big5(lstream_t decoding, const Extbyte * src,
189 unsigned_char_dynarr * dst,
190 Lstream_data_count n);
191 static void encode_coding_big5(lstream_t encoding, const Bufbyte * src,
192 unsigned_char_dynarr * dst,
193 Lstream_data_count n);
194 static int detect_coding_ucs4(struct detection_state *st, const Extbyte * src,
195 Lstream_data_count n);
196 static void decode_coding_ucs4(lstream_t decoding, const Extbyte * src,
197 unsigned_char_dynarr * dst,
198 Lstream_data_count n);
199 static void encode_coding_ucs4(lstream_t encoding, const Bufbyte * src,
200 unsigned_char_dynarr * dst,
201 Lstream_data_count n);
202 static int detect_coding_utf8(struct detection_state *st, const Extbyte * src,
203 Lstream_data_count n);
204 static void decode_coding_utf8(lstream_t decoding, const Extbyte * src,
205 unsigned_char_dynarr * dst,
206 Lstream_data_count n);
207 static void encode_coding_utf8(lstream_t encoding, const Bufbyte * src,
208 unsigned_char_dynarr * dst,
209 Lstream_data_count n);
210 static int postprocess_iso2022_mask(int mask);
211 static void reset_iso2022(Lisp_Object coding_system,
212 struct iso2022_decoder *iso);
213 static int detect_coding_iso2022(struct detection_state *st,
214 const Extbyte * src, Lstream_data_count n);
215 static void decode_coding_iso2022(lstream_t decoding, const Extbyte * src,
216 unsigned_char_dynarr * dst,
217 Lstream_data_count n);
218 static void encode_coding_iso2022(lstream_t encoding, const Bufbyte * src,
219 unsigned_char_dynarr * dst,
220 Lstream_data_count n);
222 static void decode_coding_no_conversion(lstream_t decoding, const Extbyte * src,
223 unsigned_char_dynarr * dst,
224 Lstream_data_count n);
225 static void encode_coding_no_conversion(lstream_t encoding, const Bufbyte * src,
226 unsigned_char_dynarr * dst,
227 Lstream_data_count n);
228 static void mule_decode(lstream_t decoding, const Extbyte * src,
229 unsigned_char_dynarr * dst, Lstream_data_count n);
230 static void mule_encode(lstream_t encoding, const Bufbyte * src,
231 unsigned_char_dynarr * dst, Lstream_data_count n);
233 typedef struct codesys_prop codesys_prop;
234 struct codesys_prop {
240 Dynarr_declare(codesys_prop);
241 } codesys_prop_dynarr;
243 static const struct lrecord_description codesys_prop_description_1[] = {
244 {XD_LISP_OBJECT, offsetof(codesys_prop, sym)},
248 static const struct struct_description codesys_prop_description = {
249 sizeof(codesys_prop),
250 codesys_prop_description_1
253 static const struct lrecord_description codesys_prop_dynarr_description_1[] = {
254 XD_DYNARR_DESC(codesys_prop_dynarr, &codesys_prop_description),
258 static const struct struct_description codesys_prop_dynarr_description = {
259 sizeof(codesys_prop_dynarr),
260 codesys_prop_dynarr_description_1
263 codesys_prop_dynarr *the_codesys_prop_dynarr;
265 enum codesys_prop_enum {
267 CODESYS_PROP_ISO2022,
271 /************************************************************************/
272 /* Coding system functions */
273 /************************************************************************/
275 static Lisp_Object mark_coding_system(Lisp_Object);
276 static void print_coding_system(Lisp_Object, Lisp_Object, int);
277 static void finalize_coding_system(void *header, int for_disksave);
280 static const struct lrecord_description ccs_description_1[] = {
281 {XD_LISP_OBJECT, offsetof(charset_conversion_spec, from_charset)},
282 {XD_LISP_OBJECT, offsetof(charset_conversion_spec, to_charset)},
286 static const struct struct_description ccs_description = {
287 sizeof(charset_conversion_spec),
291 static const struct lrecord_description ccsd_description_1[] = {
292 XD_DYNARR_DESC(charset_conversion_spec_dynarr, &ccs_description),
296 static const struct struct_description ccsd_description = {
297 sizeof(charset_conversion_spec_dynarr),
302 static const struct lrecord_description coding_system_description[] = {
303 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, name)},
304 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, doc_string)},
305 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, mnemonic)},
306 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, post_read_conversion)},
307 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, pre_write_conversion)},
308 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, eol_lf)},
309 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, eol_crlf)},
310 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, eol_cr)},
312 {XD_LISP_OBJECT_ARRAY,
313 offsetof(Lisp_Coding_System, iso2022.initial_charset), 4},
314 {XD_STRUCT_PTR, offsetof(Lisp_Coding_System, iso2022.input_conv), 1,
316 {XD_STRUCT_PTR, offsetof(Lisp_Coding_System, iso2022.output_conv), 1,
318 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, ccl.decode)},
319 {XD_LISP_OBJECT, offsetof(Lisp_Coding_System, ccl.encode)},
324 DEFINE_LRECORD_IMPLEMENTATION("coding-system", coding_system,
325 mark_coding_system, print_coding_system,
326 finalize_coding_system,
327 0, 0, coding_system_description,
330 static Lisp_Object mark_coding_system(Lisp_Object obj)
332 Lisp_Coding_System *codesys = XCODING_SYSTEM(obj);
334 mark_object(CODING_SYSTEM_NAME(codesys));
335 mark_object(CODING_SYSTEM_DOC_STRING(codesys));
336 mark_object(CODING_SYSTEM_MNEMONIC(codesys));
337 mark_object(CODING_SYSTEM_EOL_LF(codesys));
338 mark_object(CODING_SYSTEM_EOL_CRLF(codesys));
339 mark_object(CODING_SYSTEM_EOL_CR(codesys));
341 switch (CODING_SYSTEM_TYPE(codesys)) {
344 case CODESYS_ISO2022:
345 for (i = 0; i < 4; i++)
346 mark_object(CODING_SYSTEM_ISO2022_INITIAL_CHARSET
348 if (codesys->iso2022.input_conv) {
350 i < Dynarr_length(codesys->iso2022.input_conv);
352 struct charset_conversion_spec *ccs =
353 Dynarr_atp(codesys->iso2022.input_conv, i);
354 mark_object(ccs->from_charset);
355 mark_object(ccs->to_charset);
358 if (codesys->iso2022.output_conv) {
360 i < Dynarr_length(codesys->iso2022.output_conv);
362 struct charset_conversion_spec *ccs =
363 Dynarr_atp(codesys->iso2022.output_conv, i);
364 mark_object(ccs->from_charset);
365 mark_object(ccs->to_charset);
371 mark_object(CODING_SYSTEM_CCL_DECODE(codesys));
372 mark_object(CODING_SYSTEM_CCL_ENCODE(codesys));
375 /* list the rest of them lot explicitly */
376 case CODESYS_AUTODETECT:
377 case CODESYS_SHIFT_JIS:
381 case CODESYS_NO_CONVERSION:
383 case CODESYS_INTERNAL:
390 mark_object(CODING_SYSTEM_PRE_WRITE_CONVERSION(codesys));
391 return CODING_SYSTEM_POST_READ_CONVERSION(codesys);
395 print_coding_system(Lisp_Object obj, Lisp_Object printcharfun, int escapeflag)
397 Lisp_Coding_System *c = XCODING_SYSTEM(obj);
399 error("printing unreadable object #<coding-system 0x%x>",
402 write_c_string("#<coding-system ", printcharfun);
403 print_internal(c->name, printcharfun, 1);
404 write_c_string(">", printcharfun);
407 static void finalize_coding_system(void *header, int for_disksave)
409 Lisp_Coding_System *c = (Lisp_Coding_System *) header;
410 /* Since coding systems never go away, this function is not
411 necessary. But it would be necessary if we changed things
412 so that coding systems could go away. */
413 if (!for_disksave) { /* see comment in lstream.c */
414 switch (CODING_SYSTEM_TYPE(c)) {
416 case CODESYS_ISO2022:
417 if (c->iso2022.input_conv) {
418 Dynarr_free(c->iso2022.input_conv);
419 c->iso2022.input_conv = 0;
421 if (c->iso2022.output_conv) {
422 Dynarr_free(c->iso2022.output_conv);
423 c->iso2022.output_conv = 0;
427 /* list the rest of them lot explicitly */
428 case CODESYS_AUTODETECT:
429 case CODESYS_SHIFT_JIS:
434 case CODESYS_NO_CONVERSION:
436 case CODESYS_INTERNAL:
445 static eol_type_t symbol_to_eol_type(Lisp_Object symbol)
447 CHECK_SYMBOL(symbol);
449 return EOL_AUTODETECT;
452 if (EQ(symbol, Qcrlf))
457 signal_simple_error("Unrecognized eol type", symbol);
458 return EOL_AUTODETECT; /* not reached */
461 static Lisp_Object eol_type_to_symbol(eol_type_t type)
478 static void setup_eol_coding_systems(Lisp_Coding_System * codesys)
480 Lisp_Object codesys_obj;
481 int len = string_length(XSYMBOL(CODING_SYSTEM_NAME(codesys))->name);
482 char *codesys_name = (char *)alloca(len + 7);
484 char *codesys_mnemonic = 0;
486 Lisp_Object codesys_name_sym, sub_codesys_obj;
490 XSETCODING_SYSTEM(codesys_obj, codesys);
493 string_data(XSYMBOL(CODING_SYSTEM_NAME(codesys))->name), len);
495 if (STRINGP(CODING_SYSTEM_MNEMONIC(codesys))) {
496 mlen = XSTRING_LENGTH(CODING_SYSTEM_MNEMONIC(codesys));
497 codesys_mnemonic = (char *)alloca(mlen + 7);
498 memcpy(codesys_mnemonic,
499 XSTRING_DATA(CODING_SYSTEM_MNEMONIC(codesys)), mlen);
501 #define DEFINE_SUB_CODESYS(op_sys, op_sys_abbr, Type) do { \
502 strcpy (codesys_name + len, "-" op_sys); \
504 strcpy (codesys_mnemonic + mlen, op_sys_abbr); \
505 codesys_name_sym = intern (codesys_name); \
506 sub_codesys_obj = Fcopy_coding_system (codesys_obj, codesys_name_sym); \
507 XCODING_SYSTEM_EOL_TYPE (sub_codesys_obj) = Type; \
509 XCODING_SYSTEM_MNEMONIC(sub_codesys_obj) = \
510 build_string (codesys_mnemonic); \
511 CODING_SYSTEM_##Type (codesys) = sub_codesys_obj; \
514 DEFINE_SUB_CODESYS("unix", "", EOL_LF);
515 DEFINE_SUB_CODESYS("dos", ":T", EOL_CRLF);
516 DEFINE_SUB_CODESYS("mac", ":t", EOL_CR);
519 DEFUN("coding-system-p", Fcoding_system_p, 1, 1, 0, /*
520 Return t if OBJECT is a coding system.
521 A coding system is an object that defines how text containing multiple
522 character sets is encoded into a stream of (typically 8-bit) bytes.
523 The coding system is used to decode the stream into a series of
524 characters (which may be from multiple charsets) when the text is read
525 from a file or process, and is used to encode the text back into the
526 same format when it is written out to a file or process.
528 For example, many ISO2022-compliant coding systems (such as Compound
529 Text, which is used for inter-client data under the X Window System)
530 use escape sequences to switch between different charsets -- Japanese
531 Kanji, for example, is invoked with "ESC $ ( B"; ASCII is invoked
532 with "ESC ( B"; and Cyrillic is invoked with "ESC - L". See
533 `make-coding-system' for more information.
535 Coding systems are normally identified using a symbol, and the
536 symbol is accepted in place of the actual coding system object whenever
537 a coding system is called for. (This is similar to how faces work.)
541 return CODING_SYSTEMP(object) ? Qt : Qnil;
544 DEFUN("find-coding-system", Ffind_coding_system, 1, 1, 0, /*
545 Retrieve the coding system of the given name.
547 If CODING-SYSTEM-OR-NAME is a coding-system object, it is simply
548 returned. Otherwise, CODING-SYSTEM-OR-NAME should be a symbol.
549 If there is no such coding system, nil is returned. Otherwise the
550 associated coding system object is returned.
552 (coding_system_or_name))
554 if (NILP(coding_system_or_name))
555 coding_system_or_name = Qbinary;
556 else if (CODING_SYSTEMP(coding_system_or_name))
557 return coding_system_or_name;
559 CHECK_SYMBOL(coding_system_or_name);
562 coding_system_or_name =
563 Fgethash(coding_system_or_name, Vcoding_system_hash_table,
566 if (CODING_SYSTEMP(coding_system_or_name)
567 || NILP(coding_system_or_name))
568 return coding_system_or_name;
572 DEFUN("get-coding-system", Fget_coding_system, 1, 1, 0, /*
573 Retrieve the coding system of the given name.
574 Same as `find-coding-system' except that if there is no such
575 coding system, an error is signaled instead of returning nil.
579 Lisp_Object coding_system = Ffind_coding_system(name);
581 if (NILP(coding_system))
582 signal_simple_error("No such coding system", name);
583 return coding_system;
586 /* We store the coding systems in hash tables with the names as the key and the
587 actual coding system object as the value. Occasionally we need to use them
588 in a list format. These routines provide us with that. */
589 struct coding_system_list_closure {
590 Lisp_Object *coding_system_list;
594 add_coding_system_to_list_mapper(Lisp_Object key, Lisp_Object value,
595 void *coding_system_list_closure)
597 /* This function can GC */
598 struct coding_system_list_closure *cscl =
599 (struct coding_system_list_closure *)coding_system_list_closure;
600 Lisp_Object *coding_system_list = cscl->coding_system_list;
602 *coding_system_list = Fcons(key, *coding_system_list);
606 DEFUN("coding-system-list", Fcoding_system_list, 0, 0, 0, /*
607 Return a list of the names of all defined coding systems.
611 Lisp_Object coding_system_list = Qnil;
613 struct coding_system_list_closure coding_system_list_closure;
615 GCPRO1(coding_system_list);
616 coding_system_list_closure.coding_system_list = &coding_system_list;
617 elisp_maphash(add_coding_system_to_list_mapper,
618 Vcoding_system_hash_table, &coding_system_list_closure);
621 return coding_system_list;
624 DEFUN("coding-system-name", Fcoding_system_name, 1, 1, 0, /*
625 Return the name of the given coding system.
629 coding_system = Fget_coding_system(coding_system);
630 return XCODING_SYSTEM_NAME(coding_system);
633 static Lisp_Coding_System *allocate_coding_system(enum coding_system_type type,
636 Lisp_Coding_System *codesys =
637 alloc_lcrecord_type(Lisp_Coding_System, &lrecord_coding_system);
639 zero_lcrecord(codesys);
640 CODING_SYSTEM_PRE_WRITE_CONVERSION(codesys) = Qnil;
641 CODING_SYSTEM_POST_READ_CONVERSION(codesys) = Qnil;
642 CODING_SYSTEM_EOL_TYPE(codesys) = EOL_AUTODETECT;
643 CODING_SYSTEM_EOL_CRLF(codesys) = Qnil;
644 CODING_SYSTEM_EOL_CR(codesys) = Qnil;
645 CODING_SYSTEM_EOL_LF(codesys) = Qnil;
646 CODING_SYSTEM_TYPE(codesys) = type;
647 CODING_SYSTEM_MNEMONIC(codesys) = Qnil;
649 if (type == CODESYS_ISO2022) {
651 for (i = 0; i < 4; i++)
652 CODING_SYSTEM_ISO2022_INITIAL_CHARSET(codesys, i) =
654 } else if (type == CODESYS_CCL) {
655 CODING_SYSTEM_CCL_DECODE(codesys) = Qnil;
656 CODING_SYSTEM_CCL_ENCODE(codesys) = Qnil;
659 CODING_SYSTEM_NAME(codesys) = name;
665 /* Given a list of charset conversion specs as specified in a Lisp
666 program, parse it into STORE_HERE. */
669 parse_charset_conversion_specs(charset_conversion_spec_dynarr * store_here,
670 Lisp_Object spec_list)
674 EXTERNAL_LIST_LOOP(rest, spec_list) {
675 Lisp_Object car = XCAR(rest);
676 Lisp_Object from, to;
677 struct charset_conversion_spec spec;
679 if (!CONSP(car) || !CONSP(XCDR(car)) || !NILP(XCDR(XCDR(car))))
680 signal_simple_error("Invalid charset conversion spec",
682 from = Fget_charset(XCAR(car));
683 to = Fget_charset(XCAR(XCDR(car)));
684 if (XCHARSET_TYPE(from) != XCHARSET_TYPE(to))
685 signal_simple_error_2
686 ("Attempted conversion between different charset types",
688 spec.from_charset = from;
689 spec.to_charset = to;
691 Dynarr_add(store_here, spec);
695 /* Given a dynarr LOAD_HERE of internally-stored charset conversion
696 specs, return the equivalent as the Lisp programmer would see it.
698 If LOAD_HERE is 0, return Qnil. */
701 unparse_charset_conversion_specs(charset_conversion_spec_dynarr * load_here)
708 for (i = 0, result = Qnil; i < Dynarr_length(load_here); i++) {
709 struct charset_conversion_spec *ccs = Dynarr_atp(load_here, i);
711 Fcons(list2(ccs->from_charset, ccs->to_charset), result);
714 return Fnreverse(result);
719 DEFUN("make-coding-system", Fmake_coding_system, 2, 4, 0, /*
720 Register symbol NAME as a coding system.
722 TYPE describes the conversion method used and should be one of
725 Automatic conversion. SXEmacs attempts to detect the coding system
728 No conversion. Use this for binary files and such. On output,
729 graphic characters that are not in ASCII or Latin-1 will be
730 replaced by a ?. (For a no-conversion-encoded buffer, these
731 characters will only be present if you explicitly insert them.)
733 Shift-JIS (a Japanese encoding commonly used in PC operating systems).
735 ISO 10646 UCS-4 encoding.
737 ISO 10646 UTF-8 encoding.
739 Any ISO2022-compliant encoding. Among other things, this includes
740 JIS (the Japanese encoding commonly used for e-mail), EUC (the
741 standard Unix encoding for Japanese and other languages), and
742 Compound Text (the encoding used in X11). You can specify more
743 specific information about the conversion with the PROPS argument.
745 Big5 (the encoding commonly used for Taiwanese).
747 The conversion is performed using a user-written pseudo-code
748 program. CCL (Code Conversion Language) is the name of this
751 Write out or read in the raw contents of the memory representing
752 the buffer's text. This is primarily useful for debugging
753 purposes, and is only enabled when SXEmacs has been compiled with
754 DEBUG_SXEMACS defined (via the --debug configure option).
755 WARNING: Reading in a file using 'internal conversion can result
756 in an internal inconsistency in the memory representing a
757 buffer's text, which will produce unpredictable results and may
758 cause SXEmacs to crash. Under normal circumstances you should
759 never use 'internal conversion.
761 DOC-STRING is a string describing the coding system.
763 PROPS is a property list, describing the specific nature of the
764 character set. Recognized properties are:
767 String to be displayed in the modeline when this coding system is
771 End-of-line conversion to be used. It should be one of
774 Automatically detect the end-of-line type (LF, CRLF,
775 or CR). Also generate subsidiary coding systems named
776 `NAME-unix', `NAME-dos', and `NAME-mac', that are
777 identical to this coding system but have an EOL-TYPE
778 value of 'lf, 'crlf, and 'cr, respectively.
780 The end of a line is marked externally using ASCII LF.
781 Since this is also the way that SXEmacs represents an
782 end-of-line internally, specifying this option results
783 in no end-of-line conversion. This is the standard
784 format for Unix text files.
786 The end of a line is marked externally using ASCII
787 CRLF. This is the standard format for MS-DOS text
790 The end of a line is marked externally using ASCII CR.
791 This is the standard format for Macintosh text files.
793 Automatically detect the end-of-line type but do not
794 generate subsidiary coding systems. (This value is
795 converted to nil when stored internally, and
796 `coding-system-property' will return nil.)
798 'post-read-conversion
799 Function called after a file has been read in, to perform the
800 decoding. Called with two arguments, START and END, denoting
801 a region of the current buffer to be decoded.
803 'pre-write-conversion
804 Function called before a file is written out, to perform the
805 encoding. Called with two arguments, START and END, denoting
806 a region of the current buffer to be encoded.
808 The following additional properties are recognized if TYPE is 'iso2022:
814 The character set initially designated to the G0 - G3 registers.
815 The value should be one of
817 -- A charset object (designate that character set)
818 -- nil (do not ever use this register)
819 -- t (no character set is initially designated to
820 the register, but may be later on; this automatically
821 sets the corresponding `force-g*-on-output' property)
827 If non-nil, send an explicit designation sequence on output before
828 using the specified register.
831 If non-nil, use the short forms "ESC $ @", "ESC $ A", and
832 "ESC $ B" on output in place of the full designation sequences
833 "ESC $ ( @", "ESC $ ( A", and "ESC $ ( B".
836 If non-nil, don't designate ASCII to G0 at each end of line on output.
837 Setting this to non-nil also suppresses other state-resetting that
838 normally happens at the end of a line.
841 If non-nil, don't designate ASCII to G0 before control chars on output.
844 If non-nil, use 7-bit environment on output. Otherwise, use 8-bit
848 If non-nil, use locking-shift (SO/SI) instead of single-shift
849 or designation by escape sequence.
852 If non-nil, don't use ISO6429's direction specification.
855 If non-nil, literal control characters that are the same as
856 the beginning of a recognized ISO2022 or ISO6429 escape sequence
857 (in particular, ESC (0x1B), SO (0x0E), SI (0x0F), SS2 (0x8E),
858 SS3 (0x8F), and CSI (0x9B)) are "quoted" with an escape character
859 so that they can be properly distinguished from an escape sequence.
860 (Note that doing this results in a non-portable encoding.) This
861 encoding flag is used for byte-compiled files. Note that ESC
862 is a good choice for a quoting character because there are no
863 escape sequences whose second byte is a character from the Control-0
864 or Control-1 character sets; this is explicitly disallowed by the
867 'input-charset-conversion
868 A list of conversion specifications, specifying conversion of
869 characters in one charset to another when decoding is performed.
870 Each specification is a list of two elements: the source charset,
871 and the destination charset.
873 'output-charset-conversion
874 A list of conversion specifications, specifying conversion of
875 characters in one charset to another when encoding is performed.
876 The form of each specification is the same as for
877 'input-charset-conversion.
879 The following additional properties are recognized (and required)
883 CCL program used for decoding (converting to internal format).
886 CCL program used for encoding (converting to external format).
888 (name, type, doc_string, props))
890 Lisp_Coding_System *codesys;
891 enum coding_system_type ty;
892 int need_to_setup_eol_systems = 1;
894 /* Convert type to constant */
895 if (NILP(type) || EQ(type, Qundecided)) {
896 ty = CODESYS_AUTODETECT;
899 else if (EQ(type, Qshift_jis)) {
900 ty = CODESYS_SHIFT_JIS;
901 } else if (EQ(type, Qiso2022)) {
902 ty = CODESYS_ISO2022;
903 } else if (EQ(type, Qbig5)) {
905 } else if (EQ(type, Qucs4)) {
907 } else if (EQ(type, Qutf8)) {
909 } else if (EQ(type, Qccl)) {
913 else if (EQ(type, Qno_conversion)) {
914 ty = CODESYS_NO_CONVERSION;
917 else if (EQ(type, Qinternal)) {
918 ty = CODESYS_INTERNAL;
922 signal_simple_error("Invalid coding system type", type);
926 codesys = allocate_coding_system(ty, name);
928 if (NILP(doc_string))
929 doc_string = build_string("");
931 CHECK_STRING(doc_string);
932 CODING_SYSTEM_DOC_STRING(codesys) = doc_string;
935 EXTERNAL_PROPERTY_LIST_LOOP_3(key, value, props) {
936 if (EQ(key, Qmnemonic)) {
939 CODING_SYSTEM_MNEMONIC(codesys) = value;
942 else if (EQ(key, Qeol_type)) {
943 need_to_setup_eol_systems = NILP(value);
946 CODING_SYSTEM_EOL_TYPE(codesys) =
947 symbol_to_eol_type(value);
950 else if (EQ(key, Qpost_read_conversion))
951 CODING_SYSTEM_POST_READ_CONVERSION(codesys) =
953 else if (EQ(key, Qpre_write_conversion))
954 CODING_SYSTEM_PRE_WRITE_CONVERSION(codesys) =
957 else if (ty == CODESYS_ISO2022) {
958 #define FROB_INITIAL_CHARSET(charset_num) \
959 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, charset_num) = \
960 ((EQ (value, Qt) || EQ (value, Qnil)) ? value : Fget_charset (value))
962 if (EQ(key, Qcharset_g0))
963 FROB_INITIAL_CHARSET(0);
964 else if (EQ(key, Qcharset_g1))
965 FROB_INITIAL_CHARSET(1);
966 else if (EQ(key, Qcharset_g2))
967 FROB_INITIAL_CHARSET(2);
968 else if (EQ(key, Qcharset_g3))
969 FROB_INITIAL_CHARSET(3);
971 #define FROB_FORCE_CHARSET(charset_num) \
972 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (codesys, charset_num) = !NILP (value)
974 else if (EQ(key, Qforce_g0_on_output))
975 FROB_FORCE_CHARSET(0);
976 else if (EQ(key, Qforce_g1_on_output))
977 FROB_FORCE_CHARSET(1);
978 else if (EQ(key, Qforce_g2_on_output))
979 FROB_FORCE_CHARSET(2);
980 else if (EQ(key, Qforce_g3_on_output))
981 FROB_FORCE_CHARSET(3);
983 #define FROB_BOOLEAN_PROPERTY(prop) \
984 CODING_SYSTEM_ISO2022_##prop (codesys) = !NILP (value)
986 else if (EQ(key, Qshort))
987 FROB_BOOLEAN_PROPERTY(SHORT);
988 else if (EQ(key, Qno_ascii_eol))
989 FROB_BOOLEAN_PROPERTY(NO_ASCII_EOL);
990 else if (EQ(key, Qno_ascii_cntl))
991 FROB_BOOLEAN_PROPERTY(NO_ASCII_CNTL);
992 else if (EQ(key, Qseven))
993 FROB_BOOLEAN_PROPERTY(SEVEN);
994 else if (EQ(key, Qlock_shift))
995 FROB_BOOLEAN_PROPERTY(LOCK_SHIFT);
996 else if (EQ(key, Qno_iso6429))
997 FROB_BOOLEAN_PROPERTY(NO_ISO6429);
998 else if (EQ(key, Qescape_quoted))
999 FROB_BOOLEAN_PROPERTY(ESCAPE_QUOTED);
1001 else if (EQ(key, Qinput_charset_conversion)) {
1002 codesys->iso2022.input_conv =
1003 Dynarr_new(charset_conversion_spec);
1004 parse_charset_conversion_specs(codesys->
1008 } else if (EQ(key, Qoutput_charset_conversion)) {
1009 codesys->iso2022.output_conv =
1010 Dynarr_new(charset_conversion_spec);
1011 parse_charset_conversion_specs(codesys->
1017 ("Unrecognized property", key);
1018 } else if (EQ(type, Qccl)) {
1020 struct ccl_program test_ccl;
1023 /* Check key first. */
1024 if (EQ(key, Qdecode))
1025 suffix = "-ccl-decode";
1026 else if (EQ(key, Qencode))
1027 suffix = "-ccl-encode";
1030 ("Unrecognized property", key);
1032 /* If value is vector, register it as a ccl program
1033 associated with an newly created symbol for
1034 backward compatibility. */
1035 if (VECTORP(value)) {
1038 (Fsymbol_name(name),
1039 build_string(suffix)),
1041 Fregister_ccl_program(sym, value);
1043 CHECK_SYMBOL(value);
1046 /* check if the given ccl programs are valid. */
1047 if (setup_ccl_program(&test_ccl, sym) < 0)
1049 ("Invalid CCL program", value);
1051 if (EQ(key, Qdecode))
1052 CODING_SYSTEM_CCL_DECODE(codesys) = sym;
1053 else if (EQ(key, Qencode))
1054 CODING_SYSTEM_CCL_ENCODE(codesys) = sym;
1059 signal_simple_error("Unrecognized property",
1064 if (need_to_setup_eol_systems)
1065 setup_eol_coding_systems(codesys);
1068 Lisp_Object codesys_obj;
1069 XSETCODING_SYSTEM(codesys_obj, codesys);
1070 Fputhash(name, codesys_obj, Vcoding_system_hash_table);
1075 DEFUN("copy-coding-system", Fcopy_coding_system, 2, 2, 0, /*
1076 Copy OLD-CODING-SYSTEM to NEW-NAME.
1077 If NEW-NAME does not name an existing coding system, a new one will
1080 (old_coding_system, new_name))
1082 Lisp_Object new_coding_system;
1083 old_coding_system = Fget_coding_system(old_coding_system);
1084 new_coding_system = Ffind_coding_system(new_name);
1085 if (NILP(new_coding_system)) {
1086 XSETCODING_SYSTEM(new_coding_system,
1087 allocate_coding_system
1088 (XCODING_SYSTEM_TYPE(old_coding_system),
1090 Fputhash(new_name, new_coding_system,
1091 Vcoding_system_hash_table);
1095 Lisp_Coding_System *to = XCODING_SYSTEM(new_coding_system);
1096 Lisp_Coding_System *from = XCODING_SYSTEM(old_coding_system);
1097 memcpy(((char *)to) + sizeof(to->header),
1098 ((char *)from) + sizeof(from->header),
1099 sizeof(*from) - sizeof(from->header));
1100 to->name = new_name;
1102 return new_coding_system;
1105 DEFUN("coding-system-canonical-name-p", Fcoding_system_canonical_name_p, 1, 1, 0, /*
1106 Return t if OBJECT names a coding system, and is not a coding system alias.
1110 Lisp_Object val = Fgethash(object, Vcoding_system_hash_table, Qnil);
1111 return CODING_SYSTEMP(val) ? Qt : Qnil;
1114 DEFUN("coding-system-alias-p", Fcoding_system_alias_p, 1, 1, 0, /*
1115 Return t if OBJECT is a coding system alias.
1116 All coding system aliases are created by `define-coding-system-alias'.
1120 Lisp_Object val = Fgethash(object, Vcoding_system_hash_table, Qzero);
1121 return SYMBOLP(val) ? Qt : Qnil;
1124 DEFUN("coding-system-aliasee", Fcoding_system_aliasee, 1, 1, 0, /*
1125 Return the coding-system symbol for which symbol ALIAS is an alias.
1129 Lisp_Object aliasee = Fgethash(alias, Vcoding_system_hash_table, Qnil);
1130 if (SYMBOLP(aliasee)) {
1133 signal_simple_error(
1134 "Symbol is not a coding system alias", alias);
1136 /* To keep the compiler happy */
1141 append_suffix_to_symbol(Lisp_Object symbol, char *ascii_string)
1143 return Fintern(concat2(Fsymbol_name(symbol),
1144 build_string(ascii_string)), Qnil);
1147 /* A maphash function, for removing dangling coding system aliases. */
1149 dangling_coding_system_alias_p(Lisp_Object alias,
1150 Lisp_Object aliasee, void *dangling_aliases)
1152 if (SYMBOLP(aliasee)
1153 && NILP(Fgethash(aliasee, Vcoding_system_hash_table, Qnil))) {
1154 (*(int *)dangling_aliases)++;
1161 DEFUN("define-coding-system-alias", Fdefine_coding_system_alias, 2, 2, 0, /*
1162 Define symbol ALIAS as an alias for coding system ALIASEE.
1164 You can use this function to redefine an alias that has already been defined,
1165 but you cannot redefine a name which is the canonical name for a coding system.
1166 \(a canonical name of a coding system is what is returned when you call
1167 `coding-system-name' on a coding system).
1169 ALIASEE itself can be an alias, which allows you to define nested aliases.
1171 You are forbidden, however, from creating alias loops or `dangling' aliases.
1172 These will be detected, and an error will be signaled if you attempt to do so.
1174 If ALIASEE is nil, then ALIAS will simply be undefined.
1176 See also `coding-system-alias-p', `coding-system-aliasee',
1177 and `coding-system-canonical-name-p'.
1181 Lisp_Object real_coding_system, probe;
1183 CHECK_SYMBOL(alias);
1185 if (!NILP(Fcoding_system_canonical_name_p(alias)))
1187 ("Symbol is the canonical name of a coding system and cannot be redefined",
1190 if (NILP(aliasee)) {
1191 Lisp_Object subsidiary_unix =
1192 append_suffix_to_symbol(alias, "-unix");
1193 Lisp_Object subsidiary_dos =
1194 append_suffix_to_symbol(alias, "-dos");
1195 Lisp_Object subsidiary_mac =
1196 append_suffix_to_symbol(alias, "-mac");
1198 Fremhash(alias, Vcoding_system_hash_table);
1200 /* Undefine subsidiary aliases,
1201 presumably created by a previous call to this function */
1202 if (!NILP(Fcoding_system_alias_p(subsidiary_unix)) &&
1203 !NILP(Fcoding_system_alias_p(subsidiary_dos)) &&
1204 !NILP(Fcoding_system_alias_p(subsidiary_mac))) {
1205 Fdefine_coding_system_alias(subsidiary_unix, Qnil);
1206 Fdefine_coding_system_alias(subsidiary_dos, Qnil);
1207 Fdefine_coding_system_alias(subsidiary_mac, Qnil);
1210 /* Undefine dangling coding system aliases. */
1212 int dangling_aliases;
1215 dangling_aliases = 0;
1217 (dangling_coding_system_alias_p,
1218 Vcoding_system_hash_table,
1220 } while (dangling_aliases > 0);
1226 if (CODING_SYSTEMP(aliasee))
1227 aliasee = XCODING_SYSTEM_NAME(aliasee);
1229 /* Checks that aliasee names a coding-system */
1230 real_coding_system = Fget_coding_system(aliasee);
1232 /* Check for coding system alias loops */
1233 if (EQ(alias, aliasee))
1234 alias_loop:signal_simple_error_2
1235 ("Attempt to create a coding system alias loop", alias,
1238 for (probe = aliasee;
1240 probe = Fgethash(probe, Vcoding_system_hash_table, Qzero)) {
1241 if (EQ(probe, alias))
1245 Fputhash(alias, aliasee, Vcoding_system_hash_table);
1247 /* Set up aliases for subsidiaries.
1248 #### There must be a better way to handle subsidiary coding
1251 static char *suffixes[] = { "-unix", "-dos", "-mac" };
1253 for (int i = 0; i < countof(suffixes); i++) {
1254 Lisp_Object alias_subsidiary =
1255 append_suffix_to_symbol(alias, suffixes[i]);
1256 Lisp_Object aliasee_subsidiary =
1257 append_suffix_to_symbol(aliasee, suffixes[i]);
1259 if (!NILP(Ffind_coding_system(aliasee_subsidiary))) {
1260 Fdefine_coding_system_alias(alias_subsidiary,
1261 aliasee_subsidiary);
1265 /* FSF return value is a vector of [ALIAS-unix ALIAS-dos ALIAS-mac],
1266 but it doesn't look intentional, so I'd rather return something
1267 meaningful or nothing at all. */
1272 subsidiary_coding_system(Lisp_Object coding_system, eol_type_t type)
1274 Lisp_Coding_System *cs = XCODING_SYSTEM(coding_system);
1275 Lisp_Object new_coding_system;
1277 if (CODING_SYSTEM_EOL_TYPE(cs) != EOL_AUTODETECT)
1278 return coding_system;
1281 case EOL_AUTODETECT:
1282 return coding_system;
1284 new_coding_system = CODING_SYSTEM_EOL_LF(cs);
1287 new_coding_system = CODING_SYSTEM_EOL_CR(cs);
1290 new_coding_system = CODING_SYSTEM_EOL_CRLF(cs);
1297 return NILP(new_coding_system) ? coding_system : new_coding_system;
1300 DEFUN("subsidiary-coding-system", Fsubsidiary_coding_system, 2, 2, 0, /*
1301 Return the subsidiary coding system of CODING-SYSTEM with eol type EOL-TYPE.
1303 (coding_system, eol_type))
1305 coding_system = Fget_coding_system(coding_system);
1307 return subsidiary_coding_system(coding_system,
1308 symbol_to_eol_type(eol_type));
1311 /************************************************************************/
1312 /* Coding system accessors */
1313 /************************************************************************/
1315 DEFUN("coding-system-doc-string", Fcoding_system_doc_string, 1, 1, 0, /*
1316 Return the doc string for CODING-SYSTEM.
1320 coding_system = Fget_coding_system(coding_system);
1321 return XCODING_SYSTEM_DOC_STRING(coding_system);
1324 DEFUN("coding-system-type", Fcoding_system_type, 1, 1, 0, /*
1325 Return the type of CODING-SYSTEM.
1329 Lisp_Object tmp = Fget_coding_system(coding_system);
1331 switch (XCODING_SYSTEM_TYPE(tmp)) {
1335 case CODESYS_AUTODETECT:
1338 case CODESYS_SHIFT_JIS:
1340 case CODESYS_ISO2022:
1351 case CODESYS_NO_CONVERSION:
1352 return Qno_conversion;
1353 #ifdef DEBUG_SXEMACS
1354 case CODESYS_INTERNAL:
1362 Lisp_Object coding_system_charset(Lisp_Object coding_system, int gnum)
1365 = XCODING_SYSTEM_ISO2022_INITIAL_CHARSET(coding_system, gnum);
1367 return CHARSETP(cs) ? XCHARSET_NAME(cs) : Qnil;
1370 DEFUN("coding-system-charset", Fcoding_system_charset, 2, 2, 0, /*
1371 Return initial charset of CODING-SYSTEM designated to GNUM.
1374 (coding_system, gnum))
1376 coding_system = Fget_coding_system(coding_system);
1379 return coding_system_charset(coding_system, XINT(gnum));
1383 DEFUN("coding-system-property", Fcoding_system_property, 2, 2, 0, /*
1384 Return the PROP property of CODING-SYSTEM.
1386 (coding_system, prop))
1389 enum coding_system_type type;
1391 coding_system = Fget_coding_system(coding_system);
1393 type = XCODING_SYSTEM_TYPE(coding_system);
1395 for (i = 0; !ok && i < Dynarr_length(the_codesys_prop_dynarr); i++)
1396 if (EQ(Dynarr_at(the_codesys_prop_dynarr, i).sym, prop)) {
1398 switch (Dynarr_at(the_codesys_prop_dynarr, i).prop_type) {
1399 case CODESYS_PROP_ALL_OK:
1402 case CODESYS_PROP_ISO2022:
1403 if (type != CODESYS_ISO2022)
1405 ("Property only valid in ISO2022 coding systems",
1409 case CODESYS_PROP_CCL:
1410 if (type != CODESYS_CCL)
1412 ("Property only valid in CCL coding systems",
1422 signal_simple_error("Unrecognized property", prop);
1424 if (EQ(prop, Qname))
1425 return XCODING_SYSTEM_NAME(coding_system);
1426 else if (EQ(prop, Qtype))
1427 return Fcoding_system_type(coding_system);
1428 else if (EQ(prop, Qdoc_string))
1429 return XCODING_SYSTEM_DOC_STRING(coding_system);
1430 else if (EQ(prop, Qmnemonic))
1431 return XCODING_SYSTEM_MNEMONIC(coding_system);
1432 else if (EQ(prop, Qeol_type))
1434 eol_type_to_symbol(XCODING_SYSTEM_EOL_TYPE(coding_system));
1435 else if (EQ(prop, Qeol_lf))
1436 return XCODING_SYSTEM_EOL_LF(coding_system);
1437 else if (EQ(prop, Qeol_crlf))
1438 return XCODING_SYSTEM_EOL_CRLF(coding_system);
1439 else if (EQ(prop, Qeol_cr))
1440 return XCODING_SYSTEM_EOL_CR(coding_system);
1441 else if (EQ(prop, Qpost_read_conversion))
1442 return XCODING_SYSTEM_POST_READ_CONVERSION(coding_system);
1443 else if (EQ(prop, Qpre_write_conversion))
1444 return XCODING_SYSTEM_PRE_WRITE_CONVERSION(coding_system);
1446 else if (type == CODESYS_ISO2022) {
1447 if (EQ(prop, Qcharset_g0))
1448 return coding_system_charset(coding_system, 0);
1449 else if (EQ(prop, Qcharset_g1))
1450 return coding_system_charset(coding_system, 1);
1451 else if (EQ(prop, Qcharset_g2))
1452 return coding_system_charset(coding_system, 2);
1453 else if (EQ(prop, Qcharset_g3))
1454 return coding_system_charset(coding_system, 3);
1456 #define FORCE_CHARSET(charset_num) \
1457 (XCODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT \
1458 (coding_system, charset_num) ? Qt : Qnil)
1460 else if (EQ(prop, Qforce_g0_on_output))
1461 return FORCE_CHARSET(0);
1462 else if (EQ(prop, Qforce_g1_on_output))
1463 return FORCE_CHARSET(1);
1464 else if (EQ(prop, Qforce_g2_on_output))
1465 return FORCE_CHARSET(2);
1466 else if (EQ(prop, Qforce_g3_on_output))
1467 return FORCE_CHARSET(3);
1469 #define LISP_BOOLEAN(prop) \
1470 (XCODING_SYSTEM_ISO2022_##prop (coding_system) ? Qt : Qnil)
1472 else if (EQ(prop, Qshort))
1473 return LISP_BOOLEAN(SHORT);
1474 else if (EQ(prop, Qno_ascii_eol))
1475 return LISP_BOOLEAN(NO_ASCII_EOL);
1476 else if (EQ(prop, Qno_ascii_cntl))
1477 return LISP_BOOLEAN(NO_ASCII_CNTL);
1478 else if (EQ(prop, Qseven))
1479 return LISP_BOOLEAN(SEVEN);
1480 else if (EQ(prop, Qlock_shift))
1481 return LISP_BOOLEAN(LOCK_SHIFT);
1482 else if (EQ(prop, Qno_iso6429))
1483 return LISP_BOOLEAN(NO_ISO6429);
1484 else if (EQ(prop, Qescape_quoted))
1485 return LISP_BOOLEAN(ESCAPE_QUOTED);
1487 else if (EQ(prop, Qinput_charset_conversion))
1489 unparse_charset_conversion_specs
1490 (XCODING_SYSTEM(coding_system)->iso2022.input_conv);
1491 else if (EQ(prop, Qoutput_charset_conversion))
1493 unparse_charset_conversion_specs
1494 (XCODING_SYSTEM(coding_system)->iso2022.
1498 } 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 */
1513 /************************************************************************/
1514 /* Coding category functions */
1515 /************************************************************************/
1517 static int decode_coding_category(Lisp_Object symbol)
1521 CHECK_SYMBOL(symbol);
1522 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1523 if (EQ(coding_category_symbol[i], symbol))
1526 signal_simple_error("Unrecognized coding category", symbol);
1527 return 0; /* not reached */
1530 DEFUN("coding-category-list", Fcoding_category_list, 0, 0, 0, /*
1531 Return a list of all recognized coding categories.
1536 Lisp_Object list = Qnil;
1538 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
1539 list = Fcons(coding_category_symbol[i], list);
1543 DEFUN("set-coding-priority-list", Fset_coding_priority_list, 1, 1, 0, /*
1544 Change the priority order of the coding categories.
1545 LIST should be list of coding categories, in descending order of
1546 priority. Unspecified coding categories will be lower in priority
1547 than all specified ones, in the same relative order they were in
1552 int category_to_priority[CODING_CATEGORY_LAST];
1556 /* First generate a list that maps coding categories to priorities. */
1558 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1559 category_to_priority[i] = -1;
1561 /* Highest priority comes from the specified list. */
1563 EXTERNAL_LIST_LOOP(rest, list) {
1564 int cat = decode_coding_category(XCAR(rest));
1566 if (category_to_priority[cat] >= 0)
1567 signal_simple_error("Duplicate coding category in list",
1569 category_to_priority[cat] = i++;
1572 /* Now go through the existing categories by priority to retrieve
1573 the categories not yet specified and preserve their priority
1575 for (j = 0; j < CODING_CATEGORY_LAST; j++) {
1576 int cat = fcd->coding_category_by_priority[j];
1577 if (category_to_priority[cat] < 0)
1578 category_to_priority[cat] = i++;
1581 /* Now we need to construct the inverse of the mapping we just
1584 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1585 fcd->coding_category_by_priority[category_to_priority[i]] = i;
1587 /* Phew! That was confusing. */
1591 DEFUN("coding-priority-list", Fcoding_priority_list, 0, 0, 0, /*
1592 Return a list of coding categories in descending order of priority.
1597 Lisp_Object list = Qnil;
1599 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
1601 Fcons(coding_category_symbol
1602 [fcd->coding_category_by_priority[i]], list);
1606 DEFUN("set-coding-category-system", Fset_coding_category_system, 2, 2, 0, /*
1607 Change the coding system associated with a coding category.
1609 (coding_category, coding_system))
1611 int cat = decode_coding_category(coding_category);
1613 coding_system = Fget_coding_system(coding_system);
1614 fcd->coding_category_system[cat] = coding_system;
1618 DEFUN("coding-category-system", Fcoding_category_system, 1, 1, 0, /*
1619 Return the coding system associated with a coding category.
1623 int cat = decode_coding_category(coding_category);
1624 Lisp_Object sys = fcd->coding_category_system[cat];
1627 return XCODING_SYSTEM_NAME(sys);
1631 /************************************************************************/
1632 /* Detecting the encoding of data */
1633 /************************************************************************/
1635 struct detection_state {
1636 eol_type_t eol_type;
1663 struct iso2022_decoder iso;
1665 int high_byte_count;
1666 unsigned int saw_single_shift:1;
1675 static int acceptable_control_char_p(int c)
1678 /* Allow and ignore control characters that you might
1679 reasonably see in a text file */
1684 case 8: /* backspace */
1685 case 11: /* vertical tab */
1686 case 12: /* form feed */
1687 case 26: /* MS-DOS C-z junk */
1688 case 31: /* '^_' -- for info */
1695 static int mask_has_at_most_one_bit_p(int mask)
1697 /* Perhaps the only thing useful you learn from intensive Microsoft
1698 technical interviews */
1699 return (mask & (mask - 1)) == 0;
1703 detect_eol_type(struct detection_state *st, const Extbyte * src,
1704 Lstream_data_count n)
1707 const unsigned char c = *(const unsigned char*)src++;
1709 if (st->eol.just_saw_cr)
1711 else if (st->eol.seen_anything)
1713 } else if (st->eol.just_saw_cr)
1716 st->eol.just_saw_cr = 1;
1718 st->eol.just_saw_cr = 0;
1719 st->eol.seen_anything = 1;
1722 return EOL_AUTODETECT;
1725 /* Attempt to determine the encoding and EOL type of the given text.
1726 Before calling this function for the first type, you must initialize
1727 st->eol_type as appropriate and initialize st->mask to ~0.
1729 st->eol_type holds the determined EOL type, or EOL_AUTODETECT if
1732 st->mask holds the determined coding category mask, or ~0 if only
1733 ASCII has been seen so far.
1737 0 == st->eol_type is EOL_AUTODETECT and/or more than coding category
1738 is present in st->mask
1739 1 == definitive answers are here for both st->eol_type and st->mask
1743 detect_coding_type(struct detection_state *st, const Extbyte * src,
1744 Lstream_data_count n, int just_do_eol)
1746 if (st->eol_type == EOL_AUTODETECT)
1747 st->eol_type = detect_eol_type(st, src, n);
1750 return st->eol_type != EOL_AUTODETECT;
1752 if (!st->seen_non_ascii) {
1753 for (; n; n--, src++) {
1754 const unsigned char c = *(const unsigned char *)src;
1755 if ((c < 0x20 && !acceptable_control_char_p(c))
1757 st->seen_non_ascii = 1;
1759 st->shift_jis.mask = ~0;
1763 st->iso2022.mask = ~0;
1774 if (!mask_has_at_most_one_bit_p(st->iso2022.mask))
1775 st->iso2022.mask = detect_coding_iso2022(st, src, n);
1776 if (!mask_has_at_most_one_bit_p(st->shift_jis.mask))
1777 st->shift_jis.mask = detect_coding_sjis(st, src, n);
1778 if (!mask_has_at_most_one_bit_p(st->big5.mask))
1779 st->big5.mask = detect_coding_big5(st, src, n);
1780 if (!mask_has_at_most_one_bit_p(st->utf8.mask))
1781 st->utf8.mask = detect_coding_utf8(st, src, n);
1782 if (!mask_has_at_most_one_bit_p(st->ucs4.mask))
1783 st->ucs4.mask = detect_coding_ucs4(st, src, n);
1785 st->mask = st->iso2022.mask | st->shift_jis.mask | st->big5.mask
1786 | st->utf8.mask | st->ucs4.mask;
1789 int retval = mask_has_at_most_one_bit_p(st->mask);
1790 st->mask |= CODING_CATEGORY_NO_CONVERSION_MASK;
1791 return retval && st->eol_type != EOL_AUTODETECT;
1795 static Lisp_Object coding_system_from_mask(int mask)
1798 /* If the file was entirely or basically ASCII, use the
1799 default value of `buffer-file-coding-system'. */
1800 Lisp_Object retval =
1801 XBUFFER(Vbuffer_defaults)->buffer_file_coding_system;
1802 if (!NILP(retval)) {
1803 retval = Ffind_coding_system(retval);
1806 (Qbad_variable, Qwarning,
1807 "Invalid `default-buffer-file-coding-system', set to nil");
1808 XBUFFER(Vbuffer_defaults)->
1809 buffer_file_coding_system = Qnil;
1813 retval = Fget_coding_system(Qraw_text);
1819 mask = postprocess_iso2022_mask(mask);
1821 /* Look through the coding categories by priority and find
1822 the first one that is allowed. */
1823 for (i = 0; i < CODING_CATEGORY_LAST; i++) {
1824 cat = fcd->coding_category_by_priority[i];
1827 if ((mask & (1 << cat)) &&
1828 !NILP(fcd->coding_category_system[cat]))
1832 return fcd->coding_category_system[cat];
1834 return Fget_coding_system(Qraw_text);
1838 /* Given a seekable read stream and potential coding system and EOL type
1839 as specified, do any autodetection that is called for. If the
1840 coding system and/or EOL type are not `autodetect', they will be left
1841 alone; but this function will never return an autodetect coding system
1844 This function does not automatically fetch subsidiary coding systems;
1845 that should be unnecessary with the explicit eol-type argument. */
1847 #define LENGTH(string_constant) (sizeof (string_constant) - 1)
1848 /* number of leading lines to check for a coding cookie */
1849 #define LINES_TO_CHECK 2
1853 autodetect_real_coding_system(lstream_t stream, Lisp_Object * codesys_in_out,
1854 eol_type_t * eol_type_in_out)
1856 static const char mime_name_valid_chars[] =
1857 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1858 "abcdefghijklmnopqrstuvwxyz"
1862 struct detection_state decst;
1865 Lisp_Object coding_system = Qnil;
1868 int lines_checked = 0;
1869 Lstream_data_count nread =
1870 Lstream_read(stream, buf, sizeof(buf));
1873 /* Nothing more can be done here */
1877 decst.eol_type = *eol_type_in_out;
1881 /* Look for initial "-*-"; mode line prefix */
1882 for (p = buf, scan_end = buf + nread - LENGTH("-*-coding:?-*-");
1883 p <= scan_end && lines_checked < LINES_TO_CHECK; p++) {
1884 Extbyte *local_vars_beg = p + 3;
1886 if (*p == '\n' || *p == '\r') {
1887 /* file must use standard EOLs or we miss 2d
1888 line not to mention this is broken for
1891 /* skip past multibyte (DOS) newline */
1892 if (*p == '\r' && *(p + 1) == '\n')
1896 if (*p != '-' || *(p + 1) != '*' || *(p + 2) != '-') {
1900 /* Look for final "-*-"; mode line suffix */
1901 for (p = local_vars_beg, scan_end = buf + nread - LENGTH("-*-");
1902 p <= scan_end && lines_checked < LINES_TO_CHECK; p++) {
1903 Extbyte *suffix = p;
1904 if (*p == '\n' || *p == '\r') {
1905 /* file must use standard EOLs or we
1906 miss 2d line not to mention this is
1907 broken for UTF-16 DOS files */
1909 /* skip past multibyte (DOS) newline */
1911 && *(p + 1) == '\n')
1915 if (*p != '-' || *(p + 1) != '*' || *(p + 2) != '-') {
1919 /* Look for "coding:" */
1920 for (p = local_vars_beg, scan_end = suffix - LENGTH("coding:?");
1921 p <= scan_end; p++) {
1925 if (memcmp("coding:", p, LENGTH("coding:")) != 0) {
1928 if (p != local_vars_beg && strchr(" \t;", *p) == NULL ) {
1931 p += LENGTH("coding:");
1932 while (*p == ' ' || *p == '\t') {
1936 /* Get coding system name */
1939 /* Characters valid in a MIME charset
1940 name (rfc 1521), and in a Lisp
1942 n = strspn((char *)p, mime_name_valid_chars);
1947 coding_system = Ffind_coding_system(
1957 if (NILP(coding_system)) {
1959 if (detect_coding_type(&decst, buf, nread,
1960 XCODING_SYSTEM_TYPE(*codesys_in_out)
1961 != CODESYS_AUTODETECT))
1963 nread = Lstream_read(stream, buf, sizeof(buf));
1968 } else if (XCODING_SYSTEM_TYPE(*codesys_in_out) == CODESYS_AUTODETECT
1969 && XCODING_SYSTEM_EOL_TYPE(coding_system) == EOL_AUTODETECT) {
1971 if (detect_coding_type(&decst, buf, nread, 1))
1973 nread = Lstream_read(stream, buf, sizeof(buf));
1979 *eol_type_in_out = decst.eol_type;
1980 if (XCODING_SYSTEM_TYPE(*codesys_in_out) == CODESYS_AUTODETECT) {
1981 if (NILP(coding_system))
1983 coding_system_from_mask(decst.mask);
1985 *codesys_in_out = coding_system;
1990 determine_real_coding_system(lstream_t stream, Lisp_Object * codesys_in_out,
1991 eol_type_t * eol_type_in_out)
1994 if (*eol_type_in_out == EOL_AUTODETECT)
1995 *eol_type_in_out = XCODING_SYSTEM_EOL_TYPE(*codesys_in_out);
1998 /* If autodetection is called for, do it now. */
1999 if (XCODING_SYSTEM_TYPE(*codesys_in_out) == CODESYS_AUTODETECT
2000 || *eol_type_in_out == EOL_AUTODETECT) {
2001 autodetect_real_coding_system(stream, codesys_in_out,
2005 /* If we absolutely can't determine the EOL type, just assume LF. */
2006 if (*eol_type_in_out == EOL_AUTODETECT)
2007 *eol_type_in_out = EOL_LF;
2009 Lstream_rewind(stream);
2012 DEFUN("detect-coding-region", Fdetect_coding_region, 2, 3, 0, /*
2013 Detect coding system of the text in the region between START and END.
2014 Return a list of possible coding systems ordered by priority.
2015 If only ASCII characters are found, return 'undecided or one of
2016 its subsidiary coding systems according to a detected end-of-line
2017 type. Optional arg BUFFER defaults to the current buffer.
2019 (start, end, buffer))
2021 Lisp_Object val = Qnil;
2022 struct buffer *buf = decode_buffer(buffer, 0);
2024 Lisp_Object instream, lb_instream;
2025 lstream_t istr, lb_istr;
2026 struct detection_state decst;
2027 struct gcpro gcpro1, gcpro2;
2029 get_buffer_range_char(buf, start, end, &b, &e, 0);
2030 lb_instream = make_lisp_buffer_input_stream(buf, b, e, 0);
2031 lb_istr = XLSTREAM(lb_instream);
2033 make_encoding_input_stream(lb_istr, Fget_coding_system(Qbinary));
2034 istr = XLSTREAM(instream);
2035 GCPRO2(instream, lb_instream);
2037 decst.eol_type = EOL_AUTODETECT;
2040 Extbyte random_buffer[4096];
2041 Lstream_data_count nread =
2042 Lstream_read(istr, random_buffer, sizeof(random_buffer));
2046 if (detect_coding_type(&decst, random_buffer, nread, 0))
2050 if (decst.mask == ~0)
2051 val = subsidiary_coding_system(Fget_coding_system(Qundecided),
2058 decst.mask = postprocess_iso2022_mask(decst.mask);
2060 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--) {
2061 int sys = fcd->coding_category_by_priority[i];
2062 if (decst.mask & (1 << sys)) {
2063 Lisp_Object codesys =
2064 fcd->coding_category_system[sys];
2067 subsidiary_coding_system(codesys,
2070 val = Fcons(codesys, val);
2074 Lstream_close(istr);
2076 Lstream_delete(istr);
2077 Lstream_delete(lb_istr);
2081 /************************************************************************/
2082 /* Converting to internal Mule format ("decoding") */
2083 /************************************************************************/
2085 /* A decoding stream is a stream used for decoding text (i.e.
2086 converting from some external format to internal format).
2087 The decoding-stream object keeps track of the actual coding
2088 stream, the stream that is at the other end, and data that
2089 needs to be persistent across the lifetime of the stream. */
2091 /* Handle the EOL stuff related to just-read-in character C.
2092 EOL_TYPE is the EOL type of the coding stream.
2093 FLAGS is the current value of FLAGS in the coding stream, and may
2094 be modified by this macro. (The macro only looks at the
2095 CODING_STATE_CR flag.) DST is the Dynarr to which the decoded
2096 bytes are to be written. You need to also define a local goto
2097 label "label_continue_loop" that is at the end of the main
2098 character-reading loop.
2100 If C is a CR character, then this macro handles it entirely and
2101 jumps to label_continue_loop. Otherwise, this macro does not add
2102 anything to DST, and continues normally. You should continue
2103 processing C normally after this macro. */
2105 #define DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst) \
2109 if (eol_type == EOL_CR) \
2110 Dynarr_add (dst, '\n'); \
2111 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR) \
2112 Dynarr_add (dst, c); \
2114 flags |= CODING_STATE_CR; \
2115 goto label_continue_loop; \
2117 else if (flags & CODING_STATE_CR) \
2118 { /* eol_type == CODING_SYSTEM_EOL_CRLF */ \
2120 Dynarr_add (dst, '\r'); \
2121 flags &= ~CODING_STATE_CR; \
2125 /* C should be a binary character in the range 0 - 255; convert
2126 to internal format and add to Dynarr DST. */
2128 #define DECODE_ADD_BINARY_CHAR(c, dst) \
2130 if (BYTE_ASCII_P (c)) \
2131 Dynarr_add (dst, c); \
2132 else if (BYTE_C1_P (c)) \
2134 Dynarr_add (dst, LEADING_BYTE_CONTROL_1); \
2135 Dynarr_add (dst, c + 0x20); \
2139 Dynarr_add (dst, LEADING_BYTE_LATIN_ISO8859_1); \
2140 Dynarr_add (dst, c); \
2144 #define DECODE_OUTPUT_PARTIAL_CHAR(ch) \
2148 DECODE_ADD_BINARY_CHAR (ch, dst); \
2153 #define DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst) \
2155 if (flags & CODING_STATE_END) \
2157 DECODE_OUTPUT_PARTIAL_CHAR (ch); \
2158 if (flags & CODING_STATE_CR) \
2159 Dynarr_add (dst, '\r'); \
2163 #define DECODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, decoding)
2165 typedef struct decoding_stream_s *decoding_stream_t;
2166 struct decoding_stream_s {
2167 /* Coding system that governs the conversion. */
2168 Lisp_Coding_System *codesys;
2170 /* Stream that we read the encoded data from or
2171 write the decoded data to. */
2172 lstream_t other_end;
2174 /* If we are reading, then we can return only a fixed amount of
2175 data, so if the conversion resulted in too much data, we store it
2176 here for retrieval the next time around. */
2177 unsigned_char_dynarr *runoff;
2179 /* FLAGS holds flags indicating the current state of the decoding.
2180 Some of these flags are dependent on the coding system. */
2183 /* CH holds a partially built-up character. Since we only deal
2184 with one- and two-byte characters at the moment, we only use
2185 this to store the first byte of a two-byte character. */
2188 /* EOL_TYPE specifies the type of end-of-line conversion that
2189 currently applies. We need to keep this separate from the
2190 EOL type stored in CODESYS because the latter might indicate
2191 automatic EOL-type detection while the former will always
2192 indicate a particular EOL type. */
2193 eol_type_t eol_type;
2195 /* Additional ISO2022 information. We define the structure above
2196 because it's also needed by the detection routines. */
2197 struct iso2022_decoder iso2022;
2199 /* Additional information (the state of the running CCL program)
2200 used by the CCL decoder. */
2201 struct ccl_program ccl;
2203 /* counter for UTF-8 or UCS-4 */
2204 unsigned char counter;
2206 struct detection_state decst;
2209 static Lstream_data_count
2210 decoding_reader(lstream_t stream, unsigned char *data, Lstream_data_count size);
2211 static Lstream_data_count
2212 decoding_writer(lstream_t stream,
2213 const unsigned char *data, Lstream_data_count size);
2214 static int decoding_rewinder(lstream_t stream);
2215 static int decoding_seekable_p(lstream_t stream);
2216 static int decoding_flusher(lstream_t stream);
2217 static int decoding_closer(lstream_t stream);
2219 static Lisp_Object decoding_marker(Lisp_Object stream);
2221 DEFINE_LSTREAM_IMPLEMENTATION("decoding", lstream_decoding,
2222 sizeof(struct decoding_stream_s));
2225 decoding_marker(Lisp_Object stream)
2227 lstream_t str = DECODING_STREAM_DATA(XLSTREAM(stream))->other_end;
2228 Lisp_Object str_obj;
2230 /* We do not need to mark the coding systems or charsets stored
2231 within the stream because they are stored in a global list
2232 and automatically marked. */
2234 XSETLSTREAM(str_obj, str);
2235 mark_object(str_obj);
2236 if (str->imp->marker) {
2237 return str->imp->marker(str_obj);
2243 /* Read SIZE bytes of data and store it into DATA. We are a decoding stream
2244 so we read data from the other end, decode it, and store it into DATA. */
2246 static Lstream_data_count
2247 decoding_reader(lstream_t stream, unsigned char *data, Lstream_data_count size)
2249 decoding_stream_t str = DECODING_STREAM_DATA(stream);
2250 unsigned char *orig_data = data;
2251 Lstream_data_count read_size;
2252 int error_occurred = 0;
2254 /* We need to interface to mule_decode(), which expects to take some
2255 amount of data and store the result into a Dynarr. We have
2256 mule_decode() store into str->runoff, and take data from there
2259 /* We loop until we have enough data, reading chunks from the other
2260 end and decoding it. */
2262 /* Take data from the runoff if we can. Make sure to take at
2263 most SIZE bytes, and delete the data from the runoff. */
2264 if (Dynarr_length(str->runoff) > 0) {
2265 Lstream_data_count chunk =
2267 (Lstream_data_count)
2268 Dynarr_length(str->runoff));
2269 memcpy(data, Dynarr_atp(str->runoff, 0), chunk);
2270 Dynarr_delete_many(str->runoff, 0, chunk);
2276 /* No more room for data */
2280 if (str->flags & CODING_STATE_END) {
2281 /* This means that on the previous iteration, we hit the
2282 EOF on the other end. We loop once more so that
2283 mule_decode() can output any final stuff it may be
2284 holding, or any "go back to a sane state" escape
2285 sequences. (This latter makes sense during
2290 /* Exhausted the runoff, so get some more. DATA has at least
2291 SIZE bytes left of storage in it, so it's OK to read directly
2292 into it. (We'll be overwriting above, after we've decoded it
2293 into the runoff.) */
2294 read_size = Lstream_read(str->other_end, data, size);
2295 if (read_size < 0) {
2299 if (read_size == 0) {
2300 /* There might be some more end data produced in the
2301 translation. See the comment above. */
2302 str->flags |= CODING_STATE_END;
2304 mule_decode(stream, (Extbyte *) data, str->runoff, read_size);
2307 if (data - orig_data == 0) {
2308 return error_occurred ? -1 : 0;
2310 return data - orig_data;
2314 static Lstream_data_count
2315 decoding_writer(lstream_t stream, const unsigned char *data,
2316 Lstream_data_count size)
2318 decoding_stream_t str = DECODING_STREAM_DATA(stream);
2319 Lstream_data_count retval;
2321 /* Decode all our data into the runoff, and then attempt to write
2322 it all out to the other end. Remove whatever chunk we succeeded
2324 mule_decode(stream, (const Extbyte *)data, str->runoff, size);
2325 retval = Lstream_write(str->other_end, Dynarr_atp(str->runoff, 0),
2326 Dynarr_length(str->runoff));
2328 Dynarr_delete_many(str->runoff, 0, retval);
2330 /* Do NOT return retval. The return value indicates how much
2331 of the incoming data was written, not how many bytes were
2337 reset_decoding_stream(decoding_stream_t str)
2340 if (CODING_SYSTEM_TYPE(str->codesys) == CODESYS_ISO2022) {
2341 Lisp_Object coding_system;
2342 XSETCODING_SYSTEM(coding_system, str->codesys);
2343 reset_iso2022(coding_system, &str->iso2022);
2344 } else if (CODING_SYSTEM_TYPE(str->codesys) == CODESYS_CCL) {
2345 setup_ccl_program(&str->ccl,
2346 CODING_SYSTEM_CCL_DECODE(str->codesys));
2350 if (CODING_SYSTEM_TYPE(str->codesys) == CODESYS_AUTODETECT
2351 || CODING_SYSTEM_EOL_TYPE(str->codesys) == EOL_AUTODETECT) {
2353 str->decst.eol_type = EOL_AUTODETECT;
2354 str->decst.mask = ~0;
2356 str->flags = str->ch = 0;
2360 decoding_rewinder(lstream_t stream)
2362 decoding_stream_t str = DECODING_STREAM_DATA(stream);
2363 reset_decoding_stream(str);
2364 Dynarr_reset(str->runoff);
2365 return Lstream_rewind(str->other_end);
2369 decoding_seekable_p(lstream_t stream)
2371 decoding_stream_t str = DECODING_STREAM_DATA(stream);
2372 return Lstream_seekable_p(str->other_end);
2376 decoding_flusher(lstream_t stream)
2378 decoding_stream_t str = DECODING_STREAM_DATA(stream);
2379 return Lstream_flush(str->other_end);
2383 decoding_closer(lstream_t stream)
2385 decoding_stream_t str = DECODING_STREAM_DATA(stream);
2386 if (stream->flags & LSTREAM_FL_WRITE) {
2387 str->flags |= CODING_STATE_END;
2388 decoding_writer(stream, 0, 0);
2390 Dynarr_free(str->runoff);
2392 #ifdef ENABLE_COMPOSITE_CHARS
2393 if (str->iso2022.composite_chars) {
2394 Dynarr_free(str->iso2022.composite_chars);
2398 return Lstream_close(str->other_end);
2402 decoding_stream_coding_system(lstream_t stream)
2404 Lisp_Object coding_system;
2405 decoding_stream_t str = DECODING_STREAM_DATA(stream);
2407 XSETCODING_SYSTEM(coding_system, str->codesys);
2408 return subsidiary_coding_system(coding_system, str->eol_type);
2412 set_decoding_stream_coding_system(lstream_t lstr, Lisp_Object codesys)
2414 Lisp_Coding_System *cs = XCODING_SYSTEM(codesys);
2415 decoding_stream_t str = DECODING_STREAM_DATA(lstr);
2417 if (CODING_SYSTEM_EOL_TYPE(cs) != EOL_AUTODETECT) {
2418 str->eol_type = CODING_SYSTEM_EOL_TYPE(cs);
2420 reset_decoding_stream(str);
2424 /* WARNING WARNING WARNING WARNING!!!!! If you open up a decoding
2425 stream for writing, no automatic code detection will be performed.
2426 The reason for this is that automatic code detection requires a
2427 seekable input. Things will also fail if you open a decoding
2428 stream for reading using a non-fully-specified coding system and
2429 a non-seekable input stream. */
2432 make_decoding_stream_1(lstream_t stream, Lisp_Object codesys, const char *mode)
2434 lstream_t lstr = Lstream_new(lstream_decoding, mode);
2435 decoding_stream_t str = DECODING_STREAM_DATA(lstr);
2439 str->other_end = stream;
2440 str->runoff = (unsigned_char_dynarr *) Dynarr_new(unsigned_char);
2441 str->eol_type = EOL_AUTODETECT;
2442 if (!strcmp(mode, "r") && Lstream_seekable_p(stream)) {
2443 /* We can determine the coding system now. */
2444 determine_real_coding_system(stream, &codesys, &str->eol_type);
2446 set_decoding_stream_coding_system(lstr, codesys);
2447 str->decst.eol_type = str->eol_type;
2448 str->decst.mask = ~0;
2449 XSETLSTREAM(obj, lstr);
2454 make_decoding_input_stream(lstream_t stream, Lisp_Object codesys)
2456 return make_decoding_stream_1(stream, codesys, "r");
2460 make_decoding_output_stream(lstream_t stream, Lisp_Object codesys)
2462 return make_decoding_stream_1(stream, codesys, "w");
2465 /* Note: the decode_coding_* functions all take the same
2466 arguments as mule_decode(), which is to say some SRC data of
2467 size N, which is to be stored into dynamic array DST.
2468 DECODING is the stream within which the decoding is
2469 taking place, but no data is actually read from or
2470 written to that stream; that is handled in decoding_reader()
2471 or decoding_writer(). This allows the same functions to
2472 be used for both reading and writing. */
2475 mule_decode(lstream_t decoding, const Extbyte * src,
2476 unsigned_char_dynarr * dst, Lstream_data_count n)
2478 decoding_stream_t str = DECODING_STREAM_DATA(decoding);
2480 /* If necessary, do encoding-detection now. We do this when
2481 we're a writing stream or a non-seekable reading stream,
2482 meaning that we can't just process the whole input,
2483 rewind, and start over. */
2485 if (CODING_SYSTEM_TYPE(str->codesys) == CODESYS_AUTODETECT ||
2486 str->eol_type == EOL_AUTODETECT) {
2487 Lisp_Object codesys;
2489 XSETCODING_SYSTEM(codesys, str->codesys);
2490 detect_coding_type(&str->decst, src, n,
2491 CODING_SYSTEM_TYPE(str->codesys) !=
2492 CODESYS_AUTODETECT);
2493 if (CODING_SYSTEM_TYPE(str->codesys) == CODESYS_AUTODETECT &&
2494 str->decst.mask != ~0)
2495 /* #### This is cheesy. What we really ought to do is
2496 buffer up a certain amount of data so as to get a
2497 less random result. */
2498 codesys = coding_system_from_mask(str->decst.mask);
2499 str->eol_type = str->decst.eol_type;
2500 if (XCODING_SYSTEM(codesys) != str->codesys) {
2501 /* Preserve the CODING_STATE_END flag in case it was set.
2502 If we erase it, bad things might happen. */
2503 int was_end = str->flags & CODING_STATE_END;
2504 set_decoding_stream_coding_system(decoding, codesys);
2506 str->flags |= CODING_STATE_END;
2510 switch (CODING_SYSTEM_TYPE(str->codesys)) {
2511 #ifdef DEBUG_SXEMACS
2512 case CODESYS_INTERNAL:
2513 Dynarr_add_many(dst, src, n);
2516 case CODESYS_AUTODETECT:
2517 /* If we got this far and still haven't decided on the coding
2518 system, then do no conversion. */
2519 case CODESYS_NO_CONVERSION:
2520 decode_coding_no_conversion(decoding, src, dst, n);
2523 case CODESYS_SHIFT_JIS:
2524 decode_coding_sjis(decoding, src, dst, n);
2527 decode_coding_big5(decoding, src, dst, n);
2530 decode_coding_ucs4(decoding, src, dst, n);
2533 decode_coding_utf8(decoding, src, dst, n);
2536 str->ccl.last_block = str->flags & CODING_STATE_END;
2537 /* When applying ccl program to stream, MUST NOT set NULL
2539 ccl_driver(&str->ccl,
2541 ? (const unsigned char *)src
2542 : (const unsigned char *)""),
2543 dst, n, 0, CCL_MODE_DECODING);
2545 case CODESYS_ISO2022:
2546 decode_coding_iso2022(decoding, src, dst, n);
2554 DEFUN("decode-coding-region", Fdecode_coding_region, 3, 4, 0, /*
2555 Decode the text between START and END which is encoded in CODING-SYSTEM.
2556 This is useful if you've read in encoded text from a file without decoding
2557 it (e.g. you read in a JIS-formatted file but used the `binary' or
2558 `no-conversion' coding system, so that it shows up as "^[$B!<!+^[(B").
2559 Return length of decoded text.
2560 BUFFER defaults to the current buffer if unspecified.
2562 (start, end, coding_system, buffer))
2565 struct buffer *buf = decode_buffer(buffer, 0);
2566 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2567 lstream_t istr, ostr;
2568 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2570 get_buffer_range_char(buf, start, end, &b, &e, 0);
2572 barf_if_buffer_read_only(buf, b, e);
2574 coding_system = Fget_coding_system(coding_system);
2575 instream = make_lisp_buffer_input_stream(buf, b, e, 0);
2576 lb_outstream = make_lisp_buffer_output_stream(buf, b, 0);
2577 de_outstream = make_decoding_output_stream(XLSTREAM(lb_outstream),
2579 outstream = make_encoding_output_stream(XLSTREAM(de_outstream),
2580 Fget_coding_system(Qbinary));
2581 istr = XLSTREAM(instream);
2582 ostr = XLSTREAM(outstream);
2583 GCPRO4(instream, lb_outstream, de_outstream, outstream);
2585 /* The chain of streams looks like this:
2587 [BUFFER] <----- send through
2588 ------> [ENCODE AS BINARY]
2589 ------> [DECODE AS SPECIFIED]
2594 char tempbuf[1024]; /* some random amount */
2595 Bufpos newpos, even_newer_pos;
2596 Bufpos oldpos = lisp_buffer_stream_startpos(istr);
2597 Lstream_data_count size_in_bytes =
2598 Lstream_read(istr, tempbuf, sizeof(tempbuf));
2602 newpos = lisp_buffer_stream_startpos(istr);
2603 Lstream_write(ostr, tempbuf, size_in_bytes);
2604 even_newer_pos = lisp_buffer_stream_startpos(istr);
2605 buffer_delete_range(buf, even_newer_pos - (newpos - oldpos),
2608 Lstream_close(istr);
2609 Lstream_close(ostr);
2611 Lstream_delete(istr);
2612 Lstream_delete(ostr);
2613 Lstream_delete(XLSTREAM(de_outstream));
2614 Lstream_delete(XLSTREAM(lb_outstream));
2618 /************************************************************************/
2619 /* Converting to an external encoding ("encoding") */
2620 /************************************************************************/
2622 /* An encoding stream is an output stream. When you create the
2623 stream, you specify the coding system that governs the encoding
2624 and another stream that the resulting encoded data is to be
2625 sent to, and then start sending data to it. */
2627 #define ENCODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, encoding)
2629 typedef struct encoding_stream_s *encoding_stream_t;
2630 struct encoding_stream_s {
2631 /* Coding system that governs the conversion. */
2632 Lisp_Coding_System *codesys;
2634 /* Stream that we read the encoded data from or
2635 write the decoded data to. */
2636 lstream_t other_end;
2638 /* If we are reading, then we can return only a fixed amount of
2639 data, so if the conversion resulted in too much data, we store it
2640 here for retrieval the next time around. */
2641 unsigned_char_dynarr *runoff;
2643 /* FLAGS holds flags indicating the current state of the encoding.
2644 Some of these flags are dependent on the coding system. */
2647 /* CH holds a partially built-up character. Since we only deal
2648 with one- and two-byte characters at the moment, we only use
2649 this to store the first byte of a two-byte character. */
2652 /* Additional information used by the ISO2022 encoder. */
2654 /* CHARSET holds the character sets currently assigned to the G0
2655 through G3 registers. It is initialized from the array
2656 INITIAL_CHARSET in CODESYS. */
2657 Lisp_Object charset[4];
2659 /* Which registers are currently invoked into the left (GL) and
2660 right (GR) halves of the 8-bit encoding space? */
2661 int register_left, register_right;
2663 /* Whether we need to explicitly designate the charset in the
2664 G? register before using it. It is initialized from the
2665 array FORCE_CHARSET_ON_OUTPUT in CODESYS. */
2666 unsigned char force_charset_on_output[4];
2668 /* Other state variables that need to be preserved across
2670 Lisp_Object current_charset;
2672 int current_char_boundary;
2675 /* Additional information (the state of the running CCL program)
2676 used by the CCL encoder. */
2677 struct ccl_program ccl;
2681 static Lstream_data_count
2682 encoding_reader(lstream_t stream, unsigned char *data, Lstream_data_count size);
2683 static Lstream_data_count
2684 encoding_writer(lstream_t stream,
2685 const unsigned char *data, Lstream_data_count size);
2686 static int encoding_rewinder(lstream_t stream);
2687 static int encoding_seekable_p(lstream_t stream);
2688 static int encoding_flusher(lstream_t stream);
2689 static int encoding_closer(lstream_t stream);
2691 static Lisp_Object encoding_marker(Lisp_Object stream);
2693 DEFINE_LSTREAM_IMPLEMENTATION("encoding", lstream_encoding,
2694 sizeof(struct encoding_stream_s));
2697 encoding_marker(Lisp_Object stream)
2699 lstream_t str = ENCODING_STREAM_DATA(XLSTREAM(stream))->other_end;
2700 Lisp_Object str_obj;
2702 /* We do not need to mark the coding systems or charsets stored
2703 within the stream because they are stored in a global list
2704 and automatically marked. */
2706 XSETLSTREAM(str_obj, str);
2707 mark_object(str_obj);
2708 if (str->imp->marker) {
2709 return str->imp->marker(str_obj);
2715 /* Read SIZE bytes of data and store it into DATA. We are a encoding stream
2716 so we read data from the other end, encode it, and store it into DATA. */
2718 static Lstream_data_count
2719 encoding_reader(lstream_t stream, unsigned char *data, Lstream_data_count size)
2721 encoding_stream_t str = ENCODING_STREAM_DATA(stream);
2722 unsigned char *orig_data = data;
2723 Lstream_data_count read_size;
2724 int error_occurred = 0;
2726 /* We need to interface to mule_encode(), which expects to take some
2727 amount of data and store the result into a Dynarr. We have
2728 mule_encode() store into str->runoff, and take data from there
2731 /* We loop until we have enough data, reading chunks from the other
2732 end and encoding it. */
2734 /* Take data from the runoff if we can. Make sure to take at
2735 most SIZE bytes, and delete the data from the runoff. */
2736 if (Dynarr_length(str->runoff) > 0) {
2737 int chunk = min((int)size, Dynarr_length(str->runoff));
2738 memcpy(data, Dynarr_atp(str->runoff, 0), chunk);
2739 Dynarr_delete_many(str->runoff, 0, chunk);
2745 /* No more room for data */
2749 if (str->flags & CODING_STATE_END) {
2750 /* This means that on the previous iteration, we hit the
2751 EOF on the other end. We loop once more so that
2752 mule_encode() can output any final stuff it may be
2753 holding, or any "go back to a sane state" escape
2754 sequences. (This latter makes sense during
2759 /* Exhausted the runoff, so get some more. DATA at least SIZE
2760 bytes left of storage in it, so it's OK to read directly into
2761 it. (We'll be overwriting above, after we've encoded it into
2763 read_size = Lstream_read(str->other_end, data, size);
2764 if (read_size < 0) {
2768 if (read_size == 0) {
2769 /* There might be some more end data produced in the
2770 translation. See the comment above. */
2771 str->flags |= CODING_STATE_END;
2773 mule_encode(stream, data, str->runoff, read_size);
2776 if (data == orig_data) {
2777 return error_occurred ? -1 : 0;
2779 return data - orig_data;
2783 static Lstream_data_count
2784 encoding_writer(lstream_t stream, const unsigned char *data,
2785 Lstream_data_count size)
2787 encoding_stream_t str = ENCODING_STREAM_DATA(stream);
2788 Lstream_data_count retval;
2790 /* Encode all our data into the runoff, and then attempt to write
2791 it all out to the other end. Remove whatever chunk we succeeded
2793 mule_encode(stream, data, str->runoff, size);
2794 retval = Lstream_write(str->other_end, Dynarr_atp(str->runoff, 0),
2795 Dynarr_length(str->runoff));
2797 Dynarr_delete_many(str->runoff, 0, retval);
2799 /* Do NOT return retval. The return value indicates how much
2800 of the incoming data was written, not how many bytes were
2806 reset_encoding_stream(encoding_stream_t str)
2809 switch (CODING_SYSTEM_TYPE(str->codesys)) {
2810 case CODESYS_ISO2022: {
2813 for (i = 0; i < 4; i++) {
2814 str->iso2022.charset[i] =
2815 CODING_SYSTEM_ISO2022_INITIAL_CHARSET(
2817 str->iso2022.force_charset_on_output[i] =
2818 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT(
2821 str->iso2022.register_left = 0;
2822 str->iso2022.register_right = 1;
2823 str->iso2022.current_charset = Qnil;
2824 str->iso2022.current_half = 0;
2825 str->iso2022.current_char_boundary = 1;
2829 setup_ccl_program(&str->ccl,
2830 CODING_SYSTEM_CCL_ENCODE(str->codesys));
2833 /* list the rest of them lot explicitly */
2834 case CODESYS_AUTODETECT:
2835 case CODESYS_SHIFT_JIS:
2839 case CODESYS_NO_CONVERSION:
2840 #ifdef DEBUG_SXEMACS
2841 case CODESYS_INTERNAL:
2848 str->flags = str->ch = 0;
2852 encoding_rewinder(lstream_t stream)
2854 encoding_stream_t str = ENCODING_STREAM_DATA(stream);
2855 reset_encoding_stream(str);
2856 Dynarr_reset(str->runoff);
2857 return Lstream_rewind(str->other_end);
2861 encoding_seekable_p(lstream_t stream)
2863 encoding_stream_t str = ENCODING_STREAM_DATA(stream);
2864 return Lstream_seekable_p(str->other_end);
2868 encoding_flusher(lstream_t stream)
2870 encoding_stream_t str = ENCODING_STREAM_DATA(stream);
2871 return Lstream_flush(str->other_end);
2875 encoding_closer(lstream_t stream)
2877 encoding_stream_t str = ENCODING_STREAM_DATA(stream);
2878 if (stream->flags & LSTREAM_FL_WRITE) {
2879 str->flags |= CODING_STATE_END;
2880 encoding_writer(stream, 0, 0);
2882 Dynarr_free(str->runoff);
2883 return Lstream_close(str->other_end);
2887 encoding_stream_coding_system(lstream_t stream)
2889 Lisp_Object coding_system;
2890 encoding_stream_t str = ENCODING_STREAM_DATA(stream);
2892 XSETCODING_SYSTEM(coding_system, str->codesys);
2893 return coding_system;
2897 set_encoding_stream_coding_system(lstream_t lstr, Lisp_Object codesys)
2899 Lisp_Coding_System *cs = XCODING_SYSTEM(codesys);
2900 encoding_stream_t str = ENCODING_STREAM_DATA(lstr);
2902 reset_encoding_stream(str);
2906 make_encoding_stream_1(lstream_t stream, Lisp_Object codesys, const char *mode)
2908 lstream_t lstr = Lstream_new(lstream_encoding, mode);
2909 encoding_stream_t str = ENCODING_STREAM_DATA(lstr);
2913 str->runoff = Dynarr_new(unsigned_char);
2914 str->other_end = stream;
2915 set_encoding_stream_coding_system(lstr, codesys);
2916 XSETLSTREAM(obj, lstr);
2921 make_encoding_input_stream(lstream_t stream, Lisp_Object codesys)
2923 return make_encoding_stream_1(stream, codesys, "r");
2927 make_encoding_output_stream(lstream_t stream, Lisp_Object codesys)
2929 return make_encoding_stream_1(stream, codesys, "w");
2932 /* Convert N bytes of internally-formatted data stored in SRC to an
2933 external format, according to the encoding stream ENCODING.
2934 Store the encoded data into DST. */
2937 mule_encode(lstream_t encoding, const Bufbyte * src,
2938 unsigned_char_dynarr * dst, Lstream_data_count n)
2940 encoding_stream_t str = ENCODING_STREAM_DATA(encoding);
2942 switch (CODING_SYSTEM_TYPE(str->codesys)) {
2943 #ifdef DEBUG_SXEMACS
2944 case CODESYS_INTERNAL:
2945 Dynarr_add_many(dst, src, n);
2948 case CODESYS_AUTODETECT:
2949 /* If we got this far and still haven't decided on the coding
2950 system, then do no conversion. */
2951 case CODESYS_NO_CONVERSION:
2952 encode_coding_no_conversion(encoding, src, dst, n);
2955 case CODESYS_SHIFT_JIS:
2956 encode_coding_sjis(encoding, src, dst, n);
2959 encode_coding_big5(encoding, src, dst, n);
2962 encode_coding_ucs4(encoding, src, dst, n);
2965 encode_coding_utf8(encoding, src, dst, n);
2968 str->ccl.last_block = str->flags & CODING_STATE_END;
2969 /* When applying ccl program to stream, MUST NOT set NULL
2971 ccl_driver(&str->ccl, ((src) ? src : (unsigned char *)""),
2972 dst, n, 0, CCL_MODE_ENCODING);
2974 case CODESYS_ISO2022:
2975 encode_coding_iso2022(encoding, src, dst, n);
2983 DEFUN("encode-coding-region", Fencode_coding_region, 3, 4, 0, /*
2984 Encode the text between START and END using CODING-SYSTEM.
2985 This will, for example, convert Japanese characters into stuff such as
2986 "^[$B!<!+^[(B" if you use the JIS encoding. Return length of encoded
2987 text. BUFFER defaults to the current buffer if unspecified.
2989 (start, end, coding_system, buffer))
2992 struct buffer *buf = decode_buffer(buffer, 0);
2993 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2994 lstream_t istr, ostr;
2995 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2997 get_buffer_range_char(buf, start, end, &b, &e, 0);
2999 barf_if_buffer_read_only(buf, b, e);
3001 coding_system = Fget_coding_system(coding_system);
3002 instream = make_lisp_buffer_input_stream(buf, b, e, 0);
3003 lb_outstream = make_lisp_buffer_output_stream(buf, b, 0);
3004 de_outstream = make_decoding_output_stream(XLSTREAM(lb_outstream),
3005 Fget_coding_system(Qbinary));
3006 outstream = make_encoding_output_stream(XLSTREAM(de_outstream),
3008 istr = XLSTREAM(instream);
3009 ostr = XLSTREAM(outstream);
3010 GCPRO4(instream, outstream, de_outstream, lb_outstream);
3011 /* The chain of streams looks like this:
3013 [BUFFER] <----- send through
3014 ------> [ENCODE AS SPECIFIED]
3015 ------> [DECODE AS BINARY]
3019 char tempbuf[1024]; /* some random amount */
3020 Bufpos newpos, even_newer_pos;
3021 Bufpos oldpos = lisp_buffer_stream_startpos(istr);
3022 Lstream_data_count size_in_bytes =
3023 Lstream_read(istr, tempbuf, sizeof(tempbuf));
3027 newpos = lisp_buffer_stream_startpos(istr);
3028 Lstream_write(ostr, tempbuf, size_in_bytes);
3029 even_newer_pos = lisp_buffer_stream_startpos(istr);
3030 buffer_delete_range(buf, even_newer_pos - (newpos - oldpos),
3036 lisp_buffer_stream_startpos(XLSTREAM(instream)) - b;
3037 Lstream_close(istr);
3038 Lstream_close(ostr);
3040 Lstream_delete(istr);
3041 Lstream_delete(ostr);
3042 Lstream_delete(XLSTREAM(de_outstream));
3043 Lstream_delete(XLSTREAM(lb_outstream));
3044 return make_int(retlen);
3050 /************************************************************************/
3051 /* Shift-JIS methods */
3052 /************************************************************************/
3054 /* Shift-JIS is a coding system encoding three character sets: ASCII, right
3055 half of JISX0201-Kana, and JISX0208. An ASCII character is encoded
3056 as is. A character of JISX0201-Kana (DIMENSION1_CHARS94 character set) is
3057 encoded by "position-code + 0x80". A character of JISX0208
3058 (DIMENSION2_CHARS94 character set) is encoded in 2-byte but two
3059 position-codes are divided and shifted so that it fit in the range
3062 --- CODE RANGE of Shift-JIS ---
3063 (character set) (range)
3065 JISX0201-Kana 0xA0 .. 0xDF
3066 JISX0208 (1st byte) 0x80 .. 0x9F and 0xE0 .. 0xEF
3067 (2nd byte) 0x40 .. 0x7E and 0x80 .. 0xFC
3068 -------------------------------
3072 /* Is this the first byte of a Shift-JIS two-byte char? */
3074 #define BYTE_SJIS_TWO_BYTE_1_P(c) \
3075 (((c) >= 0x81 && (c) <= 0x9F) || ((c) >= 0xE0 && (c) <= 0xEF))
3077 /* Is this the second byte of a Shift-JIS two-byte char? */
3079 #define BYTE_SJIS_TWO_BYTE_2_P(c) \
3080 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0x80 && (c) <= 0xFC))
3082 #define BYTE_SJIS_KATAKANA_P(c) \
3083 ((c) >= 0xA1 && (c) <= 0xDF)
3086 detect_coding_sjis(struct detection_state *st, const Extbyte * src,
3087 Lstream_data_count n)
3090 const unsigned char c = *(const unsigned char *)src++;
3091 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
3093 if (st->shift_jis.in_second_byte) {
3094 st->shift_jis.in_second_byte = 0;
3097 } else if ((c >= 0x80 && c < 0xA0) || c >= 0xE0)
3098 st->shift_jis.in_second_byte = 1;
3100 return CODING_CATEGORY_SHIFT_JIS_MASK;
3103 /* Convert Shift-JIS data to internal format. */
3106 decode_coding_sjis(lstream_t decoding, const Extbyte * src,
3107 unsigned_char_dynarr * dst, Lstream_data_count n)
3109 decoding_stream_t str = DECODING_STREAM_DATA(decoding);
3110 unsigned int flags = str->flags;
3111 unsigned int ch = str->ch;
3112 eol_type_t eol_type = str->eol_type;
3115 const unsigned char c = *(const unsigned char *)src++;
3118 /* Previous character was first byte of Shift-JIS Kanji
3120 if (BYTE_SJIS_TWO_BYTE_2_P(c)) {
3121 unsigned char e1, e2;
3123 Dynarr_add(dst, LEADING_BYTE_JAPANESE_JISX0208);
3124 DECODE_SJIS(ch, c, e1, e2);
3125 Dynarr_add(dst, e1);
3126 Dynarr_add(dst, e2);
3128 DECODE_ADD_BINARY_CHAR(ch, dst);
3129 DECODE_ADD_BINARY_CHAR(c, dst);
3133 DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst);
3134 if (BYTE_SJIS_TWO_BYTE_1_P(c))
3136 else if (BYTE_SJIS_KATAKANA_P(c)) {
3137 Dynarr_add(dst, LEADING_BYTE_KATAKANA_JISX0201);
3140 DECODE_ADD_BINARY_CHAR(c, dst);
3142 label_continue_loop:;
3145 DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst);
3151 /* Convert internally-formatted data to Shift-JIS. */
3154 encode_coding_sjis(lstream_t encoding, const Bufbyte * src,
3155 unsigned_char_dynarr * dst, Lstream_data_count n)
3157 encoding_stream_t str = ENCODING_STREAM_DATA(encoding);
3158 unsigned int flags = str->flags;
3159 unsigned int ch = str->ch;
3160 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE(str->codesys);
3165 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3166 Dynarr_add(dst, '\r');
3167 if (eol_type != EOL_CR)
3168 Dynarr_add(dst, '\n');
3170 } else if (BYTE_ASCII_P(c)) {
3173 } else if (BUFBYTE_LEADING_BYTE_P(c))
3174 ch = (c == LEADING_BYTE_KATAKANA_JISX0201 ||
3175 c == LEADING_BYTE_JAPANESE_JISX0208_1978 ||
3176 c == LEADING_BYTE_JAPANESE_JISX0208) ? c : 0;
3178 if (ch == LEADING_BYTE_KATAKANA_JISX0201) {
3181 } else if (ch == LEADING_BYTE_JAPANESE_JISX0208_1978 ||
3182 ch == LEADING_BYTE_JAPANESE_JISX0208)
3185 /* j1 is bessel j1 function,
3186 * so we use something else */
3187 /* unsigned char j1, j2; */
3188 unsigned char tt1, tt2;
3190 ENCODE_SJIS(ch, c, tt1, tt2);
3191 Dynarr_add(dst, tt1);
3192 Dynarr_add(dst, tt2);
3202 DEFUN("decode-shift-jis-char", Fdecode_shift_jis_char, 1, 1, 0, /*
3203 Decode a JISX0208 character of Shift-JIS coding-system.
3204 CODE is the character code in Shift-JIS as a cons of type bytes.
3205 Return the corresponding character.
3209 unsigned char c1, c2, s1, s2;
3212 CHECK_INT(XCAR(code));
3213 CHECK_INT(XCDR(code));
3214 s1 = XINT(XCAR(code));
3215 s2 = XINT(XCDR(code));
3216 if (BYTE_SJIS_TWO_BYTE_1_P(s1) && BYTE_SJIS_TWO_BYTE_2_P(s2)) {
3217 DECODE_SJIS(s1, s2, c1, c2);
3218 return make_char(MAKE_CHAR(Vcharset_japanese_jisx0208,
3219 c1 & 0x7F, c2 & 0x7F));
3224 DEFUN("encode-shift-jis-char", Fencode_shift_jis_char, 1, 1, 0, /*
3225 Encode a JISX0208 character CHARACTER to SHIFT-JIS coding-system.
3226 Return the corresponding character code in SHIFT-JIS as a cons of two bytes.
3230 Lisp_Object charset;
3233 CHECK_CHAR_COERCE_INT(character);
3234 BREAKUP_CHAR(XCHAR(character), charset, c1, c2);
3235 if (EQ(charset, Vcharset_japanese_jisx0208)) {
3236 ENCODE_SJIS(c1 | 0x80, c2 | 0x80, s1, s2);
3237 return Fcons(make_int(s1), make_int(s2));
3242 /************************************************************************/
3244 /************************************************************************/
3246 /* BIG5 is a coding system encoding two character sets: ASCII and
3247 Big5. An ASCII character is encoded as is. Big5 is a two-byte
3248 character set and is encoded in two-byte.
3250 --- CODE RANGE of BIG5 ---
3251 (character set) (range)
3253 Big5 (1st byte) 0xA1 .. 0xFE
3254 (2nd byte) 0x40 .. 0x7E and 0xA1 .. 0xFE
3255 --------------------------
3257 Since the number of characters in Big5 is larger than maximum
3258 characters in Emacs' charset (96x96), it can't be handled as one
3259 charset. So, in Emacs, Big5 is divided into two: `charset-big5-1'
3260 and `charset-big5-2'. Both <type>s are DIMENSION2_CHARS94. The former
3261 contains frequently used characters and the latter contains less
3262 frequently used characters. */
3264 #define BYTE_BIG5_TWO_BYTE_1_P(c) \
3265 ((c) >= 0xA1 && (c) <= 0xFE)
3267 /* Is this the second byte of a Shift-JIS two-byte char? */
3269 #define BYTE_BIG5_TWO_BYTE_2_P(c) \
3270 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0xA1 && (c) <= 0xFE))
3272 /* Number of Big5 characters which have the same code in 1st byte. */
3274 #define BIG5_SAME_ROW (0xFF - 0xA1 + 0x7F - 0x40)
3276 /* Code conversion macros. These are macros because they are used in
3277 inner loops during code conversion.
3279 Note that temporary variables in macros introduce the classic
3280 dynamic-scoping problems with variable names. We use capital-
3281 lettered variables in the assumption that SXEmacs does not use
3282 capital letters in variables except in a very formalized way
3285 /* Convert Big5 code (b1, b2) into its internal string representation
3288 /* There is a much simpler way to split the Big5 charset into two.
3289 For the moment I'm going to leave the algorithm as-is because it
3290 claims to separate out the most-used characters into a single
3291 charset, which perhaps will lead to optimizations in various
3294 The way the algorithm works is something like this:
3296 Big5 can be viewed as a 94x157 charset, where the row is
3297 encoded into the bytes 0xA1 .. 0xFE and the column is encoded
3298 into the bytes 0x40 .. 0x7E and 0xA1 .. 0xFE. As for frequency,
3299 the split between low and high column numbers is apparently
3300 meaningless; ascending rows produce less and less frequent chars.
3301 Therefore, we assign the lower half of rows (0xA1 .. 0xC8) to
3302 the first charset, and the upper half (0xC9 .. 0xFE) to the
3303 second. To do the conversion, we convert the character into
3304 a single number where 0 .. 156 is the first row, 157 .. 313
3305 is the second, etc. That way, the characters are ordered by
3306 decreasing frequency. Then we just chop the space in two
3307 and coerce the result into a 94x94 space.
3310 #define DECODE_BIG5(b1, b2, lb, c1, c2) do \
3312 int B1 = b1, B2 = b2; \
3314 = (B1 - 0xA1) * BIG5_SAME_ROW + B2 - (B2 < 0x7F ? 0x40 : 0x62); \
3318 lb = LEADING_BYTE_CHINESE_BIG5_1; \
3322 lb = LEADING_BYTE_CHINESE_BIG5_2; \
3323 I -= (BIG5_SAME_ROW) * (0xC9 - 0xA1); \
3325 c1 = I / (0xFF - 0xA1) + 0xA1; \
3326 c2 = I % (0xFF - 0xA1) + 0xA1; \
3329 /* Convert the internal string representation of a Big5 character
3330 (lb, c1, c2) into Big5 code (b1, b2). */
3332 #define ENCODE_BIG5(lb, c1, c2, b1, b2) do \
3334 unsigned int I = ((c1) - 0xA1) * (0xFF - 0xA1) + ((c2) - 0xA1); \
3336 if (lb == LEADING_BYTE_CHINESE_BIG5_2) \
3338 I += BIG5_SAME_ROW * (0xC9 - 0xA1); \
3340 b1 = I / BIG5_SAME_ROW + 0xA1; \
3341 b2 = I % BIG5_SAME_ROW; \
3342 b2 += b2 < 0x3F ? 0x40 : 0x62; \
3346 detect_coding_big5(struct detection_state *st, const Extbyte * src,
3347 Lstream_data_count n)
3350 const unsigned char c = *(const unsigned char *)src++;
3351 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO ||
3352 (c >= 0x80 && c <= 0xA0))
3354 if (st->big5.in_second_byte) {
3355 st->big5.in_second_byte = 0;
3356 if (c < 0x40 || (c >= 0x80 && c <= 0xA0))
3358 } else if (c >= 0xA1)
3359 st->big5.in_second_byte = 1;
3361 return CODING_CATEGORY_BIG5_MASK;
3364 /* Convert Big5 data to internal format. */
3367 decode_coding_big5(lstream_t decoding, const Extbyte * src,
3368 unsigned_char_dynarr * dst, Lstream_data_count n)
3370 decoding_stream_t str = DECODING_STREAM_DATA(decoding);
3371 unsigned int flags = str->flags;
3372 unsigned int ch = str->ch;
3373 eol_type_t eol_type = str->eol_type;
3376 const unsigned char c = *(const unsigned char *)src++;
3378 /* Previous character was first byte of Big5 char. */
3379 if (BYTE_BIG5_TWO_BYTE_2_P(c)) {
3380 unsigned char b1, b2, b3;
3381 DECODE_BIG5(ch, c, b1, b2, b3);
3382 Dynarr_add(dst, b1);
3383 Dynarr_add(dst, b2);
3384 Dynarr_add(dst, b3);
3386 DECODE_ADD_BINARY_CHAR(ch, dst);
3387 DECODE_ADD_BINARY_CHAR(c, dst);
3391 DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst);
3392 if (BYTE_BIG5_TWO_BYTE_1_P(c))
3395 DECODE_ADD_BINARY_CHAR(c, dst);
3397 label_continue_loop:;
3400 DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst);
3406 /* Convert internally-formatted data to Big5. */
3409 encode_coding_big5(lstream_t encoding, const Bufbyte * src,
3410 unsigned_char_dynarr * dst, Lstream_data_count n)
3413 encoding_stream_t str = ENCODING_STREAM_DATA(encoding);
3414 unsigned int flags = str->flags;
3415 unsigned int ch = str->ch;
3416 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE(str->codesys);
3421 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3422 Dynarr_add(dst, '\r');
3423 if (eol_type != EOL_CR)
3424 Dynarr_add(dst, '\n');
3425 } else if (BYTE_ASCII_P(c)) {
3428 } else if (BUFBYTE_LEADING_BYTE_P(c)) {
3429 if (c == LEADING_BYTE_CHINESE_BIG5_1 ||
3430 c == LEADING_BYTE_CHINESE_BIG5_2) {
3431 /* A recognized leading byte. */
3433 continue; /* not done with this character. */
3435 /* otherwise just ignore this character. */
3436 } else if (ch == LEADING_BYTE_CHINESE_BIG5_1 ||
3437 ch == LEADING_BYTE_CHINESE_BIG5_2) {
3438 /* Previous char was a recognized leading byte. */
3440 continue; /* not done with this character. */
3442 /* Encountering second byte of a Big5 character. */
3443 unsigned char b1, b2;
3445 ENCODE_BIG5(ch >> 8, ch & 0xFF, c, b1, b2);
3446 Dynarr_add(dst, b1);
3447 Dynarr_add(dst, b2);
3457 DEFUN("decode-big5-char", Fdecode_big5_char, 1, 1, 0, /*
3458 Decode a Big5 character CODE of BIG5 coding-system.
3459 CODE is the character code in BIG5, a cons of two integers.
3460 Return the corresponding character.
3464 unsigned char c1, c2, b1, b2;
3467 CHECK_INT(XCAR(code));
3468 CHECK_INT(XCDR(code));
3469 b1 = XINT(XCAR(code));
3470 b2 = XINT(XCDR(code));
3471 if (BYTE_BIG5_TWO_BYTE_1_P(b1) && BYTE_BIG5_TWO_BYTE_2_P(b2)) {
3473 Lisp_Object charset;
3474 DECODE_BIG5(b1, b2, leading_byte, c1, c2);
3475 charset = CHARSET_BY_LEADING_BYTE(leading_byte);
3476 return make_char(MAKE_CHAR(charset, c1 & 0x7F, c2 & 0x7F));
3481 DEFUN("encode-big5-char", Fencode_big5_char, 1, 1, 0, /*
3482 Encode the Big5 character CHARACTER in the BIG5 coding-system.
3483 Return the corresponding character code in Big5.
3487 Lisp_Object charset;
3490 CHECK_CHAR_COERCE_INT(character);
3491 BREAKUP_CHAR(XCHAR(character), charset, c1, c2);
3492 if (EQ(charset, Vcharset_chinese_big5_1) ||
3493 EQ(charset, Vcharset_chinese_big5_2)) {
3494 ENCODE_BIG5(XCHARSET_LEADING_BYTE(charset), c1 | 0x80,
3496 return Fcons(make_int(b1), make_int(b2));
3501 /************************************************************************/
3504 /* UCS-4 character codes are implemented as nonnegative integers. */
3506 /************************************************************************/
3508 DEFUN("set-ucs-char", Fset_ucs_char, 2, 2, 0, /*
3509 Map UCS-4 code CODE to Mule character CHARACTER.
3511 Return T on success, NIL on failure.
3517 CHECK_CHAR(character);
3521 if (c < countof(fcd->ucs_to_mule_table)) {
3522 fcd->ucs_to_mule_table[c] = character;
3528 static Lisp_Object ucs_to_char(unsigned long code)
3530 if (code < countof(fcd->ucs_to_mule_table)) {
3531 return fcd->ucs_to_mule_table[code];
3532 } else if ((0xe00000 <= code) && (code <= 0xe00000 + 94 * 94 * 14)) {
3536 c = code % (94 * 94);
3538 (MAKE_CHAR(CHARSET_BY_ATTRIBUTES
3539 (CHARSET_TYPE_94X94, code / (94 * 94) + '@',
3540 CHARSET_LEFT_TO_RIGHT),
3541 c / 94 + 33, c % 94 + 33));
3546 DEFUN("ucs-char", Fucs_char, 1, 1, 0, /*
3547 Return Mule character corresponding to UCS code CODE (a positive integer).
3552 return ucs_to_char(XINT(code));
3555 DEFUN("set-char-ucs", Fset_char_ucs, 2, 2, 0, /*
3556 Map Mule character CHARACTER to UCS code CODE (a positive integer).
3560 /* #### Isn't this gilding the lily? Fput_char_table checks its args.
3561 Fset_char_ucs is more restrictive on index arg, but should
3562 check code arg in a char_table method. */
3563 CHECK_CHAR(character);
3565 return Fput_char_table(character, code, mule_to_ucs_table);
3568 DEFUN("char-ucs", Fchar_ucs, 1, 1, 0, /*
3569 Return the UCS code (a positive integer) corresponding to CHARACTER.
3573 return Fget_char_table(character, mule_to_ucs_table);
3576 /* Decode a UCS-4 character into a buffer. If the lookup fails, use
3577 <GETA MARK> (U+3013) of JIS X 0208, which means correct character
3578 is not found, instead.
3579 #### do something more appropriate (use blob?)
3580 Danger, Will Robinson! Data loss. Should we signal user? */
3581 static void decode_ucs4(unsigned long ch, unsigned_char_dynarr * dst)
3583 Lisp_Object chr = ucs_to_char(ch);
3586 Bufbyte work[MAX_EMCHAR_LEN];
3591 simple_set_charptr_emchar(work, ch) :
3592 non_ascii_set_charptr_emchar(work, ch);
3593 Dynarr_add_many(dst, work, len);
3595 Dynarr_add(dst, LEADING_BYTE_JAPANESE_JISX0208);
3596 Dynarr_add(dst, 34 + 128);
3597 Dynarr_add(dst, 46 + 128);
3601 static unsigned long
3602 mule_char_to_ucs4(Lisp_Object charset, unsigned char h, unsigned char l)
3605 = Fget_char_table(make_char(MAKE_CHAR(charset, h & 127, l & 127)),
3610 } else if ((XCHARSET_DIMENSION(charset) == 2) &&
3611 (XCHARSET_CHARS(charset) == 94)) {
3612 unsigned char final = XCHARSET_FINAL(charset);
3614 if (('@' <= final) && (final < 0x7f)) {
3615 return 0xe00000 + (final - '@') * 94 * 94
3616 + ((h & 127) - 33) * 94 + (l & 127) - 33;
3626 encode_ucs4(Lisp_Object charset,
3627 unsigned char h, unsigned char l, unsigned_char_dynarr * dst)
3629 unsigned long code = mule_char_to_ucs4(charset, h, l);
3630 Dynarr_add(dst, code >> 24);
3631 Dynarr_add(dst, (code >> 16) & 255);
3632 Dynarr_add(dst, (code >> 8) & 255);
3633 Dynarr_add(dst, code & 255);
3637 detect_coding_ucs4(struct detection_state *st, const Extbyte * src,
3638 Lstream_data_count n)
3641 const unsigned char c = *(const unsigned char *)src++;
3642 switch (st->ucs4.in_byte) {
3650 st->ucs4.in_byte = 0;
3656 return CODING_CATEGORY_UCS4_MASK;
3660 decode_coding_ucs4(lstream_t decoding, const Extbyte * src,
3661 unsigned_char_dynarr * dst, Lstream_data_count n)
3663 decoding_stream_t str = DECODING_STREAM_DATA(decoding);
3664 unsigned int flags = str->flags;
3665 unsigned int ch = str->ch;
3666 unsigned char counter = str->counter;
3669 const unsigned char c = *(const unsigned char *)src++;
3676 decode_ucs4((ch << 8) | c, dst);
3685 if (counter & CODING_STATE_END)
3686 DECODE_OUTPUT_PARTIAL_CHAR(ch);
3690 str->counter = counter;
3694 encode_coding_ucs4(lstream_t encoding, const Bufbyte * src,
3695 unsigned_char_dynarr * dst, Lstream_data_count n)
3697 encoding_stream_t str = ENCODING_STREAM_DATA(encoding);
3698 unsigned int flags = str->flags;
3699 unsigned int ch = str->ch;
3700 unsigned char char_boundary = str->iso2022.current_char_boundary;
3701 Lisp_Object charset = str->iso2022.current_charset;
3703 #ifdef ENABLE_COMPOSITE_CHARS
3704 /* flags for handling composite chars. We do a little switcharoo
3705 on the source while we're outputting the composite char. */
3706 unsigned int saved_n = 0;
3707 const unsigned char *saved_src = NULL;
3708 int in_composite = 0;
3714 unsigned char c = *src++;
3716 if (BYTE_ASCII_P(c)) { /* Processing ASCII character */
3718 encode_ucs4(Vcharset_ascii, c, 0, dst);
3720 } else if (BUFBYTE_LEADING_BYTE_P(c) || BUFBYTE_LEADING_BYTE_P(ch)) { /* Processing Leading Byte */
3722 charset = CHARSET_BY_LEADING_BYTE(c);
3723 if (LEADING_BYTE_PREFIX_P(c))
3726 } else { /* Processing Non-ASCII character */
3728 if (EQ(charset, Vcharset_control_1)) {
3729 encode_ucs4(Vcharset_control_1, c, 0, dst);
3731 switch (XCHARSET_REP_BYTES(charset)) {
3733 encode_ucs4(charset, c, 0, dst);
3736 if (XCHARSET_PRIVATE_P(charset)) {
3737 encode_ucs4(charset, c, 0, dst);
3740 #ifdef ENABLE_COMPOSITE_CHARS
3743 Vcharset_composite)) {
3745 /* #### Bother! We don't know how to
3758 (Vcharset_composite,
3763 composite_char_string
3772 n = XSTRING_LENGTH(lstr);
3775 #endif /* ENABLE_COMPOSITE_CHARS */
3777 encode_ucs4(charset, ch,
3788 encode_ucs4(charset, ch, c,
3803 #ifdef ENABLE_COMPOSITE_CHARS
3808 goto back_to_square_n; /* Wheeeeeeeee ..... */
3810 #endif /* ENABLE_COMPOSITE_CHARS */
3814 str->iso2022.current_char_boundary = char_boundary;
3815 str->iso2022.current_charset = charset;
3817 /* Verbum caro factum est! */
3820 /************************************************************************/
3822 /************************************************************************/
3825 detect_coding_utf8(struct detection_state *st, const Extbyte * src,
3826 Lstream_data_count n)
3829 const unsigned char c = *(const unsigned char *)src++;
3830 switch (st->utf8.in_byte) {
3832 if (c == ISO_CODE_ESC || c == ISO_CODE_SI
3833 || c == ISO_CODE_SO)
3836 st->utf8.in_byte = 5;
3838 st->utf8.in_byte = 4;
3840 st->utf8.in_byte = 3;
3842 st->utf8.in_byte = 2;
3844 st->utf8.in_byte = 1;
3849 if ((c & 0xc0) != 0x80)
3855 return CODING_CATEGORY_UTF8_MASK;
3859 decode_coding_utf8(lstream_t decoding, const Extbyte * src,
3860 unsigned_char_dynarr * dst, Lstream_data_count n)
3862 decoding_stream_t str = DECODING_STREAM_DATA(decoding);
3863 unsigned int flags = str->flags;
3864 unsigned int ch = str->ch;
3865 eol_type_t eol_type = str->eol_type;
3866 unsigned char counter = str->counter;
3869 const unsigned char c = *(const unsigned char *)src++;
3875 } else if (c >= 0xf8) {
3878 } else if (c >= 0xf0) {
3881 } else if (c >= 0xe0) {
3884 } else if (c >= 0xc0) {
3888 DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst);
3889 decode_ucs4(c, dst);
3893 ch = (ch << 6) | (c & 0x3f);
3894 decode_ucs4(ch, dst);
3899 ch = (ch << 6) | (c & 0x3f);
3902 label_continue_loop:;
3905 if (flags & CODING_STATE_END)
3906 DECODE_OUTPUT_PARTIAL_CHAR(ch);
3910 str->counter = counter;
3914 encode_utf8(Lisp_Object charset,
3915 unsigned char h, unsigned char l, unsigned_char_dynarr * dst)
3917 unsigned long code = mule_char_to_ucs4(charset, h, l);
3919 Dynarr_add(dst, code);
3920 } else if (code <= 0x7ff) {
3921 Dynarr_add(dst, (code >> 6) | 0xc0);
3922 Dynarr_add(dst, (code & 0x3f) | 0x80);
3923 } else if (code <= 0xffff) {
3924 Dynarr_add(dst, (code >> 12) | 0xe0);
3925 Dynarr_add(dst, ((code >> 6) & 0x3f) | 0x80);
3926 Dynarr_add(dst, (code & 0x3f) | 0x80);
3927 } else if (code <= 0x1fffff) {
3928 Dynarr_add(dst, (code >> 18) | 0xf0);
3929 Dynarr_add(dst, ((code >> 12) & 0x3f) | 0x80);
3930 Dynarr_add(dst, ((code >> 6) & 0x3f) | 0x80);
3931 Dynarr_add(dst, (code & 0x3f) | 0x80);
3932 } else if (code <= 0x3ffffff) {
3933 Dynarr_add(dst, (code >> 24) | 0xf8);
3934 Dynarr_add(dst, ((code >> 18) & 0x3f) | 0x80);
3935 Dynarr_add(dst, ((code >> 12) & 0x3f) | 0x80);
3936 Dynarr_add(dst, ((code >> 6) & 0x3f) | 0x80);
3937 Dynarr_add(dst, (code & 0x3f) | 0x80);
3939 Dynarr_add(dst, (code >> 30) | 0xfc);
3940 Dynarr_add(dst, ((code >> 24) & 0x3f) | 0x80);
3941 Dynarr_add(dst, ((code >> 18) & 0x3f) | 0x80);
3942 Dynarr_add(dst, ((code >> 12) & 0x3f) | 0x80);
3943 Dynarr_add(dst, ((code >> 6) & 0x3f) | 0x80);
3944 Dynarr_add(dst, (code & 0x3f) | 0x80);
3949 encode_coding_utf8(lstream_t encoding, const Bufbyte * src,
3950 unsigned_char_dynarr * dst, Lstream_data_count n)
3952 encoding_stream_t str = ENCODING_STREAM_DATA(encoding);
3953 unsigned int flags = str->flags;
3954 unsigned int ch = str->ch;
3955 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE(str->codesys);
3956 unsigned char char_boundary = str->iso2022.current_char_boundary;
3957 Lisp_Object charset = str->iso2022.current_charset;
3959 #ifdef ENABLE_COMPOSITE_CHARS
3960 /* flags for handling composite chars. We do a little switcharoo
3961 on the source while we're outputting the composite char. */
3962 unsigned int saved_n = 0;
3963 const unsigned char *saved_src = NULL;
3964 int in_composite = 0;
3967 #endif /* ENABLE_COMPOSITE_CHARS */
3970 unsigned char c = *src++;
3972 if (BYTE_ASCII_P(c)) { /* Processing ASCII character */
3975 if (eol_type != EOL_LF
3976 && eol_type != EOL_AUTODETECT)
3977 Dynarr_add(dst, '\r');
3978 if (eol_type != EOL_CR)
3981 encode_utf8(Vcharset_ascii, c, 0, dst);
3983 } else if (BUFBYTE_LEADING_BYTE_P(c) || BUFBYTE_LEADING_BYTE_P(ch)) { /* Processing Leading Byte */
3985 charset = CHARSET_BY_LEADING_BYTE(c);
3986 if (LEADING_BYTE_PREFIX_P(c))
3989 } else { /* Processing Non-ASCII character */
3991 if (EQ(charset, Vcharset_control_1)) {
3992 encode_utf8(Vcharset_control_1, c, 0, dst);
3994 switch (XCHARSET_REP_BYTES(charset)) {
3996 encode_utf8(charset, c, 0, dst);
3999 if (XCHARSET_PRIVATE_P(charset)) {
4000 encode_utf8(charset, c, 0, dst);
4003 #ifdef ENABLE_COMPOSITE_CHARS
4006 Vcharset_composite)) {
4008 /* #### Bother! We don't know how to
4017 (Vcharset_composite,
4022 composite_char_string
4031 n = XSTRING_LENGTH(lstr);
4034 #endif /* ENABLE_COMPOSITE_CHARS */
4036 encode_utf8(charset, ch,
4047 encode_utf8(charset, ch, c,
4062 #ifdef ENABLE_COMPOSITE_CHARS
4067 goto back_to_square_n; /* Wheeeeeeeee ..... */
4073 str->iso2022.current_char_boundary = char_boundary;
4074 str->iso2022.current_charset = charset;
4076 /* Verbum caro factum est! */
4079 /************************************************************************/
4080 /* ISO2022 methods */
4081 /************************************************************************/
4083 /* The following note describes the coding system ISO2022 briefly.
4084 Since the intention of this note is to help understand the
4085 functions in this file, some parts are NOT ACCURATE or OVERLY
4086 SIMPLIFIED. For thorough understanding, please refer to the
4087 original document of ISO2022.
4089 ISO2022 provides many mechanisms to encode several character sets
4090 in 7-bit and 8-bit environments. For 7-bit environments, all text
4091 is encoded using bytes less than 128. This may make the encoded
4092 text a little bit longer, but the text passes more easily through
4093 several gateways, some of which strip off MSB (Most Signigant Bit).
4095 There are two kinds of character sets: control character set and
4096 graphic character set. The former contains control characters such
4097 as `newline' and `escape' to provide control functions (control
4098 functions are also provided by escape sequences). The latter
4099 contains graphic characters such as 'A' and '-'. Emacs recognizes
4100 two control character sets and many graphic character sets.
4102 Graphic character sets are classified into one of the following
4103 four classes, according to the number of bytes (DIMENSION) and
4104 number of characters in one dimension (CHARS) of the set:
4105 - DIMENSION1_CHARS94
4106 - DIMENSION1_CHARS96
4107 - DIMENSION2_CHARS94
4108 - DIMENSION2_CHARS96
4110 In addition, each character set is assigned an identification tag,
4111 unique for each set, called "final character" (denoted as <F>
4112 hereafter). The <F> of each character set is decided by ECMA(*)
4113 when it is registered in ISO. The code range of <F> is 0x30..0x7F
4114 (0x30..0x3F are for private use only).
4116 Note (*): ECMA = European Computer Manufacturers Association
4118 Here are examples of graphic character set [NAME(<F>)]:
4119 o DIMENSION1_CHARS94 -- ASCII('B'), right-half-of-JISX0201('I'), ...
4120 o DIMENSION1_CHARS96 -- right-half-of-ISO8859-1('A'), ...
4121 o DIMENSION2_CHARS94 -- GB2312('A'), JISX0208('B'), ...
4122 o DIMENSION2_CHARS96 -- none for the moment
4124 A code area (1 byte = 8 bits) is divided into 4 areas, C0, GL, C1, and GR.
4125 C0 [0x00..0x1F] -- control character plane 0
4126 GL [0x20..0x7F] -- graphic character plane 0
4127 C1 [0x80..0x9F] -- control character plane 1
4128 GR [0xA0..0xFF] -- graphic character plane 1
4130 A control character set is directly designated and invoked to C0 or
4131 C1 by an escape sequence. The most common case is that:
4132 - ISO646's control character set is designated/invoked to C0, and
4133 - ISO6429's control character set is designated/invoked to C1,
4134 and usually these designations/invocations are omitted in encoded
4135 text. In a 7-bit environment, only C0 can be used, and a control
4136 character for C1 is encoded by an appropriate escape sequence to
4137 fit into the environment. All control characters for C1 are
4138 defined to have corresponding escape sequences.
4140 A graphic character set is at first designated to one of four
4141 graphic registers (G0 through G3), then these graphic registers are
4142 invoked to GL or GR. These designations and invocations can be
4143 done independently. The most common case is that G0 is invoked to
4144 GL, G1 is invoked to GR, and ASCII is designated to G0. Usually
4145 these invocations and designations are omitted in encoded text.
4146 In a 7-bit environment, only GL can be used.
4148 When a graphic character set of CHARS94 is invoked to GL, codes
4149 0x20 and 0x7F of the GL area work as control characters SPACE and
4150 DEL respectively, and codes 0xA0 and 0xFF of the GR area should not
4153 There are two ways of invocation: locking-shift and single-shift.
4154 With locking-shift, the invocation lasts until the next different
4155 invocation, whereas with single-shift, the invocation affects the
4156 following character only and doesn't affect the locking-shift
4157 state. Invocations are done by the following control characters or
4160 ----------------------------------------------------------------------
4161 abbrev function cntrl escape seq description
4162 ----------------------------------------------------------------------
4163 SI/LS0 (shift-in) 0x0F none invoke G0 into GL
4164 SO/LS1 (shift-out) 0x0E none invoke G1 into GL
4165 LS2 (locking-shift-2) none ESC 'n' invoke G2 into GL
4166 LS3 (locking-shift-3) none ESC 'o' invoke G3 into GL
4167 LS1R (locking-shift-1 right) none ESC '~' invoke G1 into GR (*)
4168 LS2R (locking-shift-2 right) none ESC '}' invoke G2 into GR (*)
4169 LS3R (locking-shift 3 right) none ESC '|' invoke G3 into GR (*)
4170 SS2 (single-shift-2) 0x8E ESC 'N' invoke G2 for one char
4171 SS3 (single-shift-3) 0x8F ESC 'O' invoke G3 for one char
4172 ----------------------------------------------------------------------
4173 (*) These are not used by any known coding system.
4175 Control characters for these functions are defined by macros
4176 ISO_CODE_XXX in `coding.h'.
4178 Designations are done by the following escape sequences:
4179 ----------------------------------------------------------------------
4180 escape sequence description
4181 ----------------------------------------------------------------------
4182 ESC '(' <F> designate DIMENSION1_CHARS94<F> to G0
4183 ESC ')' <F> designate DIMENSION1_CHARS94<F> to G1
4184 ESC '*' <F> designate DIMENSION1_CHARS94<F> to G2
4185 ESC '+' <F> designate DIMENSION1_CHARS94<F> to G3
4186 ESC ',' <F> designate DIMENSION1_CHARS96<F> to G0 (*)
4187 ESC '-' <F> designate DIMENSION1_CHARS96<F> to G1
4188 ESC '.' <F> designate DIMENSION1_CHARS96<F> to G2
4189 ESC '/' <F> designate DIMENSION1_CHARS96<F> to G3
4190 ESC '$' '(' <F> designate DIMENSION2_CHARS94<F> to G0 (**)
4191 ESC '$' ')' <F> designate DIMENSION2_CHARS94<F> to G1
4192 ESC '$' '*' <F> designate DIMENSION2_CHARS94<F> to G2
4193 ESC '$' '+' <F> designate DIMENSION2_CHARS94<F> to G3
4194 ESC '$' ',' <F> designate DIMENSION2_CHARS96<F> to G0 (*)
4195 ESC '$' '-' <F> designate DIMENSION2_CHARS96<F> to G1
4196 ESC '$' '.' <F> designate DIMENSION2_CHARS96<F> to G2
4197 ESC '$' '/' <F> designate DIMENSION2_CHARS96<F> to G3
4198 ----------------------------------------------------------------------
4200 In this list, "DIMENSION1_CHARS94<F>" means a graphic character set
4201 of dimension 1, chars 94, and final character <F>, etc...
4203 Note (*): Although these designations are not allowed in ISO2022,
4204 Emacs accepts them on decoding, and produces them on encoding
4205 CHARS96 character sets in a coding system which is characterized as
4206 7-bit environment, non-locking-shift, and non-single-shift.
4208 Note (**): If <F> is '@', 'A', or 'B', the intermediate character
4209 '(' can be omitted. We refer to this as "short-form" hereafter.
4211 Now you may notice that there are a lot of ways for encoding the
4212 same multilingual text in ISO2022. Actually, there exist many
4213 coding systems such as Compound Text (used in X11's inter client
4214 communication, ISO-2022-JP (used in Japanese internet), ISO-2022-KR
4215 (used in Korean internet), EUC (Extended UNIX Code, used in Asian
4216 localized platforms), and all of these are variants of ISO2022.
4218 In addition to the above, Emacs handles two more kinds of escape
4219 sequences: ISO6429's direction specification and Emacs' private
4220 sequence for specifying character composition.
4222 ISO6429's direction specification takes the following form:
4223 o CSI ']' -- end of the current direction
4224 o CSI '0' ']' -- end of the current direction
4225 o CSI '1' ']' -- start of left-to-right text
4226 o CSI '2' ']' -- start of right-to-left text
4227 The control character CSI (0x9B: control sequence introducer) is
4228 abbreviated to the escape sequence ESC '[' in a 7-bit environment.
4230 Character composition specification takes the following form:
4231 o ESC '0' -- start character composition
4232 o ESC '1' -- end character composition
4233 Since these are not standard escape sequences of any ISO standard,
4234 their use with these meanings is restricted to Emacs only. */
4237 reset_iso2022(Lisp_Object coding_system, struct iso2022_decoder *iso)
4241 for (i = 0; i < 4; i++) {
4242 if (!NILP(coding_system))
4244 XCODING_SYSTEM_ISO2022_INITIAL_CHARSET
4247 iso->charset[i] = Qt;
4248 iso->invalid_designated[i] = 0;
4250 iso->esc = ISO_ESC_NOTHING;
4251 iso->esc_bytes_index = 0;
4252 iso->register_left = 0;
4253 iso->register_right = 1;
4254 iso->switched_dir_and_no_valid_charset_yet = 0;
4255 iso->invalid_switch_dir = 0;
4256 iso->output_direction_sequence = 0;
4257 iso->output_literally = 0;
4258 #ifdef ENABLE_COMPOSITE_CHARS
4259 if (iso->composite_chars)
4260 Dynarr_reset(iso->composite_chars);
4264 static int fit_to_be_escape_quoted(unsigned char c)
4280 /* Parse one byte of an ISO2022 escape sequence.
4281 If the result is an invalid escape sequence, return 0 and
4282 do not change anything in STR. Otherwise, if the result is
4283 an incomplete escape sequence, update ISO2022.ESC and
4284 ISO2022.ESC_BYTES and return -1. Otherwise, update
4285 all the state variables (but not ISO2022.ESC_BYTES) and
4288 If CHECK_INVALID_CHARSETS is non-zero, check for designation
4289 or invocation of an invalid character set and treat that as
4290 an unrecognized escape sequence.
4292 ********************************************************************
4294 #### Strategies for error annotation and coding orthogonalization
4296 We really want to separate out a number of things. Conceptually,
4297 there is a nested syntax.
4299 At the top level is the ISO 2022 extension syntax, including charset
4300 designation and invocation, and certain auxiliary controls such as the
4301 ISO 6429 direction specification. These are octet-oriented, with the
4302 single exception (AFAIK) of the "exit Unicode" sequence which uses the
4303 UTF's natural width (1 byte for UTF-7 and UTF-8, 2 bytes for UCS-2 and
4304 UTF-16, and 4 bytes for UCS-4 and UTF-32). This will be treated as a
4305 (deprecated) special case in Unicode processing.
4307 The middle layer is ISO 2022 character interpretation. This will depend
4308 on the current state of the ISO 2022 registers, and assembles octets
4309 into the character's internal representation.
4311 The lowest level is translating system control conventions. At present
4312 this is restricted to newline translation, but one could imagine doing
4313 tab conversion or line wrapping here. "Escape from Unicode" processing
4314 would be done at this level.
4316 At each level the parser will verify the syntax. In the case of a
4317 syntax error or warning (such as a redundant escape sequence that affects
4318 no characters), the parser will take some action, typically inserting the
4319 erroneous octets directly into the output and creating an annotation
4320 which can be used by higher level I/O to mark the affected region.
4322 This should make it possible to do something sensible about separating
4323 newline convention processing from character construction, and about
4324 preventing ISO 2022 escape sequences from being recognized
4327 The basic strategy will be to have octet classification tables, and
4328 switch processing according to the table entry.
4330 It's possible that, by doing the processing with tables of functions or
4331 the like, the parser can be used for both detection and translation. */
4334 parse_iso2022_esc(Lisp_Object codesys, struct iso2022_decoder *iso,
4335 unsigned char c, unsigned int *flags,
4336 int check_invalid_charsets)
4338 /* (1) If we're at the end of a designation sequence, CS is the
4339 charset being designated and REG is the register to designate
4342 (2) If we're at the end of a locking-shift sequence, REG is
4343 the register to invoke and HALF (0 == left, 1 == right) is
4344 the half to invoke it into.
4346 (3) If we're at the end of a single-shift sequence, REG is
4347 the register to invoke. */
4348 Lisp_Object cs = Qnil;
4351 /* NOTE: This code does goto's all over the fucking place.
4352 The reason for this is that we're basically implementing
4353 a state machine here, and hierarchical languages like C
4354 don't really provide a clean way of doing this. */
4356 if (!(*flags & CODING_STATE_ESCAPE))
4357 /* At beginning of escape sequence; we need to reset our
4358 escape-state variables. */
4359 iso->esc = ISO_ESC_NOTHING;
4361 iso->output_literally = 0;
4362 iso->output_direction_sequence = 0;
4365 case ISO_ESC_NOTHING:
4366 iso->esc_bytes_index = 0;
4368 case ISO_CODE_ESC: /* Start escape sequence */
4369 *flags |= CODING_STATE_ESCAPE;
4373 case ISO_CODE_CSI: /* ISO6429 (specifying directionality) */
4374 *flags |= CODING_STATE_ESCAPE;
4375 iso->esc = ISO_ESC_5_11;
4378 case ISO_CODE_SO: /* locking shift 1 */
4382 case ISO_CODE_SI: /* locking shift 0 */
4387 case ISO_CODE_SS2: /* single shift */
4390 case ISO_CODE_SS3: /* single shift */
4394 default: /* Other control characters */
4400 /**** single shift ****/
4402 case 'N': /* single shift 2 */
4405 case 'O': /* single shift 3 */
4409 /**** locking shift ****/
4411 case '~': /* locking shift 1 right */
4415 case 'n': /* locking shift 2 */
4419 case '}': /* locking shift 2 right */
4423 case 'o': /* locking shift 3 */
4427 case '|': /* locking shift 3 right */
4432 #ifdef ENABLE_COMPOSITE_CHARS
4433 /**** composite ****/
4436 iso->esc = ISO_ESC_START_COMPOSITE;
4437 *flags = (*flags & CODING_STATE_ISO2022_LOCK) |
4438 CODING_STATE_COMPOSITE;
4442 iso->esc = ISO_ESC_END_COMPOSITE;
4443 *flags = (*flags & CODING_STATE_ISO2022_LOCK) &
4444 ~CODING_STATE_COMPOSITE;
4446 #endif /* ENABLE_COMPOSITE_CHARS */
4448 /**** directionality ****/
4451 iso->esc = ISO_ESC_5_11;
4454 /**** designation ****/
4456 case '$': /* multibyte charset prefix */
4457 iso->esc = ISO_ESC_2_4;
4461 if (0x28 <= c && c <= 0x2F) {
4463 (enum iso_esc_flag)(c - 0x28 + ISO_ESC_2_8);
4467 /* This function is called with CODESYS equal to nil when
4468 doing coding-system detection. */
4470 && XCODING_SYSTEM_ISO2022_ESCAPE_QUOTED(codesys)
4471 && fit_to_be_escape_quoted(c)) {
4472 iso->esc = ISO_ESC_LITERAL;
4473 *flags &= CODING_STATE_ISO2022_LOCK;
4481 /**** directionality ****/
4483 case ISO_ESC_5_11: /* ISO6429 direction control */
4486 (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4487 goto directionality;
4490 iso->esc = ISO_ESC_5_11_0;
4492 iso->esc = ISO_ESC_5_11_1;
4494 iso->esc = ISO_ESC_5_11_2;
4499 case ISO_ESC_5_11_0:
4502 (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4503 goto directionality;
4507 case ISO_ESC_5_11_1:
4510 (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4511 goto directionality;
4515 case ISO_ESC_5_11_2:
4518 (*flags & CODING_STATE_ISO2022_LOCK) |
4520 goto directionality;
4525 iso->esc = ISO_ESC_DIRECTIONALITY;
4526 /* Various junk here to attempt to preserve the direction
4527 sequences literally in the text if they would otherwise be
4528 swallowed due to invalid designations that don't show up as
4529 actual charset changes in the text. */
4530 if (iso->invalid_switch_dir) {
4531 /* We already inserted a direction switch literally into
4532 the text. We assume (#### this may not be right)
4533 that the next direction switch is the one going the
4534 other way, and we need to output that literally as
4536 iso->output_literally = 1;
4537 iso->invalid_switch_dir = 0;
4541 /* If we are in the thrall of an invalid designation,
4542 then stick the directionality sequence literally into
4543 the output stream so it ends up in the original text
4545 for (jj = 0; jj < 4; jj++)
4546 if (iso->invalid_designated[jj])
4549 iso->output_literally = 1;
4550 iso->invalid_switch_dir = 1;
4552 /* Indicate that we haven't yet seen a valid
4553 designation, so that if a switch-dir is
4554 directly followed by an invalid designation,
4555 both get inserted literally. */
4556 iso->switched_dir_and_no_valid_charset_yet = 1;
4560 /**** designation ****/
4563 if (0x28 <= c && c <= 0x2F) {
4565 (enum iso_esc_flag)(c - 0x28 + ISO_ESC_2_4_8);
4568 if (0x40 <= c && c <= 0x42) {
4569 cs = CHARSET_BY_ATTRIBUTES(CHARSET_TYPE_94X94, c,
4570 *flags & CODING_STATE_R2L ?
4571 CHARSET_RIGHT_TO_LEFT :
4572 CHARSET_LEFT_TO_RIGHT);
4589 case ISO_ESC_2_4_10:
4590 case ISO_ESC_2_4_11:
4591 case ISO_ESC_2_4_12:
4592 case ISO_ESC_2_4_13:
4593 case ISO_ESC_2_4_14:
4594 case ISO_ESC_2_4_15:
4595 case ISO_ESC_SINGLE_SHIFT:
4596 case ISO_ESC_LOCKING_SHIFT:
4597 case ISO_ESC_DESIGNATE:
4598 case ISO_ESC_DIRECTIONALITY:
4599 case ISO_ESC_LITERAL:
4604 if (c < '0' || c > '~')
4605 return 0; /* bad final byte */
4607 if (iso->esc >= ISO_ESC_2_8 && iso->esc <= ISO_ESC_2_15) {
4608 type = ((iso->esc >= ISO_ESC_2_12) ?
4609 CHARSET_TYPE_96 : CHARSET_TYPE_94);
4610 reg = (iso->esc - ISO_ESC_2_8) & 3;
4611 } else if (iso->esc >= ISO_ESC_2_4_8 &&
4612 iso->esc <= ISO_ESC_2_4_15) {
4613 type = ((iso->esc >= ISO_ESC_2_4_12) ?
4614 CHARSET_TYPE_96X96 :
4615 CHARSET_TYPE_94X94);
4616 reg = (iso->esc - ISO_ESC_2_4_8) & 3;
4618 /* Can this ever be reached? -slb */
4623 cs = CHARSET_BY_ATTRIBUTES(type, c,
4624 *flags & CODING_STATE_R2L ?
4625 CHARSET_RIGHT_TO_LEFT :
4626 CHARSET_LEFT_TO_RIGHT);
4632 iso->esc_bytes[iso->esc_bytes_index++] = (unsigned char)c;
4636 if (check_invalid_charsets && !CHARSETP(iso->charset[reg]))
4637 /* can't invoke something that ain't there. */
4639 iso->esc = ISO_ESC_SINGLE_SHIFT;
4640 *flags &= CODING_STATE_ISO2022_LOCK;
4642 *flags |= CODING_STATE_SS2;
4644 *flags |= CODING_STATE_SS3;
4648 if (check_invalid_charsets && !CHARSETP(iso->charset[reg]))
4649 /* can't invoke something that ain't there. */
4652 iso->register_right = reg;
4654 iso->register_left = reg;
4655 *flags &= CODING_STATE_ISO2022_LOCK;
4656 iso->esc = ISO_ESC_LOCKING_SHIFT;
4660 if (NILP(cs) && check_invalid_charsets) {
4661 iso->invalid_designated[reg] = 1;
4662 iso->charset[reg] = Vcharset_ascii;
4663 iso->esc = ISO_ESC_DESIGNATE;
4664 *flags &= CODING_STATE_ISO2022_LOCK;
4665 iso->output_literally = 1;
4666 if (iso->switched_dir_and_no_valid_charset_yet) {
4667 /* We encountered a switch-direction followed by an
4668 invalid designation. Ensure that the switch-direction
4669 gets outputted; otherwise it will probably get eaten
4670 when the text is written out again. */
4671 iso->switched_dir_and_no_valid_charset_yet = 0;
4672 iso->output_direction_sequence = 1;
4673 /* And make sure that the switch-dir going the other
4674 way gets outputted, as well. */
4675 iso->invalid_switch_dir = 1;
4679 /* This function is called with CODESYS equal to nil when
4680 doing coding-system detection. */
4681 if (!NILP(codesys)) {
4682 charset_conversion_spec_dynarr *dyn =
4683 XCODING_SYSTEM(codesys)->iso2022.input_conv;
4688 for (i = 0; i < Dynarr_length(dyn); i++) {
4689 struct charset_conversion_spec *spec =
4691 if (EQ(cs, spec->from_charset))
4692 cs = spec->to_charset;
4697 iso->charset[reg] = cs;
4698 iso->esc = ISO_ESC_DESIGNATE;
4699 *flags &= CODING_STATE_ISO2022_LOCK;
4700 if (iso->invalid_designated[reg]) {
4701 iso->invalid_designated[reg] = 0;
4702 iso->output_literally = 1;
4704 if (iso->switched_dir_and_no_valid_charset_yet)
4705 iso->switched_dir_and_no_valid_charset_yet = 0;
4710 detect_coding_iso2022(struct detection_state *st, const Extbyte * src,
4711 Lstream_data_count n)
4715 /* #### There are serious deficiencies in the recognition mechanism
4716 here. This needs to be much smarter if it's going to cut it.
4717 The sequence "\xff\x0f" is currently detected as LOCK_SHIFT while
4718 it should be detected as Latin-1.
4719 All the ISO2022 stuff in this file should be synced up with the
4720 code from FSF Emacs-20.4, in which Mule should be more or less stable.
4721 Perhaps we should wait till R2L works in FSF Emacs? */
4723 if (!st->iso2022.initted) {
4724 reset_iso2022(Qnil, &st->iso2022.iso);
4725 st->iso2022.mask = (CODING_CATEGORY_ISO_7_MASK |
4726 CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
4727 CODING_CATEGORY_ISO_8_1_MASK |
4728 CODING_CATEGORY_ISO_8_2_MASK |
4729 CODING_CATEGORY_ISO_LOCK_SHIFT_MASK);
4730 st->iso2022.flags = 0;
4731 st->iso2022.high_byte_count = 0;
4732 st->iso2022.saw_single_shift = 0;
4733 st->iso2022.initted = 1;
4736 mask = st->iso2022.mask;
4739 const unsigned char c = *(const unsigned char *)src++;
4741 mask &= ~CODING_CATEGORY_ISO_7_MASK;
4742 st->iso2022.high_byte_count++;
4744 if (st->iso2022.high_byte_count
4745 && !st->iso2022.saw_single_shift) {
4746 if (st->iso2022.high_byte_count & 1)
4747 /* odd number of high bytes; assume not iso-8-2 */
4748 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
4750 st->iso2022.high_byte_count = 0;
4751 st->iso2022.saw_single_shift = 0;
4753 mask &= ~CODING_CATEGORY_ISO_7_MASK;
4755 if (!(st->iso2022.flags & CODING_STATE_ESCAPE)
4756 && (BYTE_C0_P(c) || BYTE_C1_P(c))) { /* control chars */
4758 /* Allow and ignore control characters that you might
4759 reasonably see in a text file */
4764 case 8: /* backspace */
4765 case 11: /* vertical tab */
4766 case 12: /* form feed */
4767 case 26: /* MS-DOS C-z junk */
4768 case 31: /* '^_' -- for info */
4769 goto label_continue_loop;
4776 if ((st->iso2022.flags & CODING_STATE_ESCAPE) || BYTE_C0_P(c)
4778 if (parse_iso2022_esc(Qnil, &st->iso2022.iso, c,
4779 &st->iso2022.flags, 0)) {
4780 switch (st->iso2022.iso.esc) {
4781 case ISO_ESC_DESIGNATE:
4782 mask &= ~CODING_CATEGORY_ISO_8_1_MASK;
4783 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
4785 case ISO_ESC_LOCKING_SHIFT:
4786 mask = CODING_CATEGORY_ISO_LOCK_SHIFT_MASK;
4787 goto ran_out_of_chars;
4788 case ISO_ESC_SINGLE_SHIFT:
4789 mask &= ~CODING_CATEGORY_ISO_8_DESIGNATE_MASK;
4790 st->iso2022.saw_single_shift = 1;
4794 case ISO_ESC_NOTHING:
4807 case ISO_ESC_2_4_10:
4808 case ISO_ESC_2_4_11:
4809 case ISO_ESC_2_4_12:
4810 case ISO_ESC_2_4_13:
4811 case ISO_ESC_2_4_14:
4812 case ISO_ESC_2_4_15:
4814 case ISO_ESC_5_11_0:
4815 case ISO_ESC_5_11_1:
4816 case ISO_ESC_5_11_2:
4817 case ISO_ESC_DIRECTIONALITY:
4818 case ISO_ESC_LITERAL:
4824 goto ran_out_of_chars;
4827 label_continue_loop:;
4834 static int postprocess_iso2022_mask(int mask)
4836 /* #### kind of cheesy */
4837 /* If seven-bit ISO is allowed, then assume that the encoding is
4838 entirely seven-bit and turn off the eight-bit ones. */
4839 if (mask & CODING_CATEGORY_ISO_7_MASK)
4840 mask &= ~(CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
4841 CODING_CATEGORY_ISO_8_1_MASK |
4842 CODING_CATEGORY_ISO_8_2_MASK);
4846 /* If FLAGS is a null pointer or specifies right-to-left motion,
4847 output a switch-dir-to-left-to-right sequence to DST.
4848 Also update FLAGS if it is not a null pointer.
4849 If INTERNAL_P is set, we are outputting in internal format and
4850 need to handle the CSI differently. */
4853 restore_left_to_right_direction(Lisp_Coding_System * codesys,
4854 unsigned_char_dynarr * dst,
4855 unsigned int *flags, int internal_p)
4857 if (!flags || (*flags & CODING_STATE_R2L)) {
4858 if (CODING_SYSTEM_ISO2022_SEVEN(codesys)) {
4859 Dynarr_add(dst, ISO_CODE_ESC);
4860 Dynarr_add(dst, '[');
4861 } else if (internal_p)
4862 DECODE_ADD_BINARY_CHAR(ISO_CODE_CSI, dst);
4864 Dynarr_add(dst, ISO_CODE_CSI);
4865 Dynarr_add(dst, '0');
4866 Dynarr_add(dst, ']');
4868 *flags &= ~CODING_STATE_R2L;
4872 /* If FLAGS is a null pointer or specifies a direction different from
4873 DIRECTION (which should be either CHARSET_RIGHT_TO_LEFT or
4874 CHARSET_LEFT_TO_RIGHT), output the appropriate switch-dir escape
4875 sequence to DST. Also update FLAGS if it is not a null pointer.
4876 If INTERNAL_P is set, we are outputting in internal format and
4877 need to handle the CSI differently. */
4880 ensure_correct_direction(int direction, Lisp_Coding_System * codesys,
4881 unsigned_char_dynarr * dst, unsigned int *flags,
4884 if ((!flags || (*flags & CODING_STATE_R2L)) &&
4885 direction == CHARSET_LEFT_TO_RIGHT)
4886 restore_left_to_right_direction(codesys, dst, flags,
4888 else if (!CODING_SYSTEM_ISO2022_NO_ISO6429(codesys)
4889 && (!flags || !(*flags & CODING_STATE_R2L)) &&
4890 direction == CHARSET_RIGHT_TO_LEFT) {
4891 if (CODING_SYSTEM_ISO2022_SEVEN(codesys)) {
4892 Dynarr_add(dst, ISO_CODE_ESC);
4893 Dynarr_add(dst, '[');
4894 } else if (internal_p)
4895 DECODE_ADD_BINARY_CHAR(ISO_CODE_CSI, dst);
4897 Dynarr_add(dst, ISO_CODE_CSI);
4898 Dynarr_add(dst, '2');
4899 Dynarr_add(dst, ']');
4901 *flags |= CODING_STATE_R2L;
4905 /* Convert ISO2022-format data to internal format. */
4908 decode_coding_iso2022(lstream_t decoding, const Extbyte * src,
4909 unsigned_char_dynarr * dst, Lstream_data_count n)
4911 decoding_stream_t str = DECODING_STREAM_DATA(decoding);
4912 unsigned int flags = str->flags;
4913 unsigned int ch = str->ch;
4914 eol_type_t eol_type = str->eol_type;
4915 #ifdef ENABLE_COMPOSITE_CHARS
4916 unsigned_char_dynarr *real_dst = dst;
4918 Lisp_Object coding_system;
4920 XSETCODING_SYSTEM(coding_system, str->codesys);
4922 #ifdef ENABLE_COMPOSITE_CHARS
4923 if (flags & CODING_STATE_COMPOSITE)
4924 dst = str->iso2022.composite_chars;
4925 #endif /* ENABLE_COMPOSITE_CHARS */
4928 const unsigned char c = *(const unsigned char *)src++;
4929 if (flags & CODING_STATE_ESCAPE) {
4930 /* Within ESC sequence */
4931 int retval = parse_iso2022_esc(
4932 coding_system, &str->iso2022, c, &flags, 1);
4935 switch (str->iso2022.esc) {
4936 #ifdef ENABLE_COMPOSITE_CHARS
4937 case ISO_ESC_START_COMPOSITE:
4938 if (str->iso2022.composite_chars)
4939 Dynarr_reset(str->iso2022.
4942 str->iso2022.composite_chars =
4943 Dynarr_new(unsigned_char);
4944 dst = str->iso2022.composite_chars;
4946 case ISO_ESC_END_COMPOSITE:
4948 Bufbyte comstr[MAX_EMCHAR_LEN];
4951 lookup_composite_char
4952 (Dynarr_atp(dst, 0),
4953 Dynarr_length(dst));
4956 set_charptr_emchar(comstr,
4958 Dynarr_add_many(dst, comstr,
4962 #endif /* ENABLE_COMPOSITE_CHARS */
4964 case ISO_ESC_LITERAL:
4965 DECODE_ADD_BINARY_CHAR(c, dst);
4968 case ISO_ESC_NOTHING:
4981 case ISO_ESC_2_4_10:
4982 case ISO_ESC_2_4_11:
4983 case ISO_ESC_2_4_12:
4984 case ISO_ESC_2_4_13:
4985 case ISO_ESC_2_4_14:
4986 case ISO_ESC_2_4_15:
4988 case ISO_ESC_5_11_0:
4989 case ISO_ESC_5_11_1:
4990 case ISO_ESC_5_11_2:
4991 case ISO_ESC_SINGLE_SHIFT:
4992 case ISO_ESC_LOCKING_SHIFT:
4993 case ISO_ESC_DESIGNATE:
4994 case ISO_ESC_DIRECTIONALITY:
4997 /* Everything else handled already */
5002 /* Attempted error recovery. */
5003 if (str->iso2022.output_direction_sequence)
5004 ensure_correct_direction(flags &
5006 CHARSET_RIGHT_TO_LEFT :
5007 CHARSET_LEFT_TO_RIGHT,
5008 str->codesys, dst, 0,
5010 /* More error recovery. */
5011 if (!retval || str->iso2022.output_literally) {
5012 /* Output the (possibly invalid) sequence */
5014 for (i = 0; i < str->iso2022.esc_bytes_index;
5016 DECODE_ADD_BINARY_CHAR(str->iso2022.
5019 flags &= CODING_STATE_ISO2022_LOCK;
5021 n++, src--; /* Repeat the loop with the same character. */
5023 /* No sense in reprocessing the final byte of the
5024 escape sequence; it could mess things up anyway.
5026 DECODE_ADD_BINARY_CHAR(c, dst);
5030 } else if (BYTE_C0_P(c) || BYTE_C1_P(c)) { /* Control characters */
5032 /***** Error-handling *****/
5034 /* If we were in the middle of a character, dump out the
5035 partial character. */
5036 DECODE_OUTPUT_PARTIAL_CHAR(ch);
5038 /* If we just saw a single-shift character, dump it out.
5039 This may dump out the wrong sort of single-shift character,
5040 but least it will give an indication that something went
5042 if (flags & CODING_STATE_SS2) {
5043 DECODE_ADD_BINARY_CHAR(ISO_CODE_SS2, dst);
5044 flags &= ~CODING_STATE_SS2;
5046 if (flags & CODING_STATE_SS3) {
5047 DECODE_ADD_BINARY_CHAR(ISO_CODE_SS3, dst);
5048 flags &= ~CODING_STATE_SS3;
5051 /***** Now handle the control characters. *****/
5054 DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst);
5056 flags &= CODING_STATE_ISO2022_LOCK;
5058 if (!parse_iso2022_esc
5059 (coding_system, &str->iso2022, c, &flags, 1))
5060 DECODE_ADD_BINARY_CHAR(c, dst);
5061 } else { /* Graphic characters */
5062 Lisp_Object charset;
5066 DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst);
5068 /* Now determine the charset. */
5069 reg = ((flags & CODING_STATE_SS2) ? 2
5070 : (flags & CODING_STATE_SS3) ? 3
5071 : !BYTE_ASCII_P(c) ? str->iso2022.register_right
5072 : str->iso2022.register_left);
5073 charset = str->iso2022.charset[reg];
5075 /* Error checking: */
5076 if (!CHARSETP(charset)
5077 || str->iso2022.invalid_designated[reg]
5079 (((c & 0x7F) == ' ' || (c & 0x7F) == ISO_CODE_DEL)
5080 && XCHARSET_CHARS(charset) == 94))
5081 /* Mrmph. We are trying to invoke a register that has no
5082 or an invalid charset in it, or trying to add a character
5083 outside the range of the charset. Insert that char literally
5084 to preserve it for the output. */
5086 DECODE_OUTPUT_PARTIAL_CHAR(ch);
5087 DECODE_ADD_BINARY_CHAR(c, dst);
5091 /* Things are probably hunky-dorey. */
5093 /* Fetch reverse charset, maybe. */
5094 if (((flags & CODING_STATE_R2L) &&
5095 XCHARSET_DIRECTION(charset) ==
5096 CHARSET_LEFT_TO_RIGHT)
5097 || (!(flags & CODING_STATE_R2L)
5098 && XCHARSET_DIRECTION(charset) ==
5099 CHARSET_RIGHT_TO_LEFT)) {
5100 Lisp_Object new_charset =
5101 XCHARSET_REVERSE_DIRECTION_CHARSET
5103 if (!NILP(new_charset))
5104 charset = new_charset;
5107 lb = XCHARSET_LEADING_BYTE(charset);
5108 switch (XCHARSET_REP_BYTES(charset)) {
5110 DECODE_OUTPUT_PARTIAL_CHAR(ch);
5111 Dynarr_add(dst, c & 0x7F);
5114 case 2: /* one-byte official */
5115 DECODE_OUTPUT_PARTIAL_CHAR(ch);
5116 Dynarr_add(dst, lb);
5117 Dynarr_add(dst, c | 0x80);
5120 case 3: /* one-byte private or two-byte official */
5121 if (XCHARSET_PRIVATE_P(charset)) {
5122 DECODE_OUTPUT_PARTIAL_CHAR(ch);
5124 PRE_LEADING_BYTE_PRIVATE_1);
5125 Dynarr_add(dst, lb);
5126 Dynarr_add(dst, c | 0x80);
5129 Dynarr_add(dst, lb);
5140 default: /* two-byte private */
5143 PRE_LEADING_BYTE_PRIVATE_2);
5144 Dynarr_add(dst, lb);
5145 Dynarr_add(dst, ch | 0x80);
5146 Dynarr_add(dst, c | 0x80);
5154 flags &= CODING_STATE_ISO2022_LOCK;
5157 label_continue_loop:;
5160 if (flags & CODING_STATE_END)
5161 DECODE_OUTPUT_PARTIAL_CHAR(ch);
5167 /***** ISO2022 encoder *****/
5169 /* Designate CHARSET into register REG. */
5172 iso2022_designate(Lisp_Object charset, unsigned char reg,
5173 encoding_stream_t str, unsigned_char_dynarr * dst)
5175 static const char inter94[] = "()*+";
5176 static const char inter96[] = ",-./";
5178 unsigned char final;
5179 Lisp_Object old_charset = str->iso2022.charset[reg];
5181 str->iso2022.charset[reg] = charset;
5182 if (!CHARSETP(charset))
5183 /* charset might be an initial nil or t. */
5185 type = XCHARSET_TYPE(charset);
5186 final = XCHARSET_FINAL(charset);
5187 if (!str->iso2022.force_charset_on_output[reg] &&
5188 CHARSETP(old_charset) &&
5189 XCHARSET_TYPE(old_charset) == type &&
5190 XCHARSET_FINAL(old_charset) == final)
5193 str->iso2022.force_charset_on_output[reg] = 0;
5196 charset_conversion_spec_dynarr *dyn =
5197 str->codesys->iso2022.output_conv;
5202 for (i = 0; i < Dynarr_length(dyn); i++) {
5203 struct charset_conversion_spec *spec =
5205 if (EQ(charset, spec->from_charset))
5206 charset = spec->to_charset;
5211 Dynarr_add(dst, ISO_CODE_ESC);
5213 case CHARSET_TYPE_94:
5214 Dynarr_add(dst, inter94[reg]);
5216 case CHARSET_TYPE_96:
5217 Dynarr_add(dst, inter96[reg]);
5219 case CHARSET_TYPE_94X94:
5220 Dynarr_add(dst, '$');
5221 if (reg != 0 || !(CODING_SYSTEM_ISO2022_SHORT(str->codesys))
5222 || final < '@' || final > 'B')
5223 Dynarr_add(dst, inter94[reg]);
5225 case CHARSET_TYPE_96X96:
5226 Dynarr_add(dst, '$');
5227 Dynarr_add(dst, inter96[reg]);
5232 Dynarr_add(dst, final);
5236 ensure_normal_shift(encoding_stream_t str, unsigned_char_dynarr * dst)
5238 if (str->iso2022.register_left != 0) {
5239 Dynarr_add(dst, ISO_CODE_SI);
5240 str->iso2022.register_left = 0;
5245 ensure_shift_out(encoding_stream_t str, unsigned_char_dynarr * dst)
5247 if (str->iso2022.register_left != 1) {
5248 Dynarr_add(dst, ISO_CODE_SO);
5249 str->iso2022.register_left = 1;
5253 /* Convert internally-formatted data to ISO2022 format. */
5256 encode_coding_iso2022(lstream_t encoding, const Bufbyte * src,
5257 unsigned_char_dynarr * dst, Lstream_data_count n)
5259 unsigned char charmask, c;
5260 unsigned char char_boundary;
5261 encoding_stream_t str = ENCODING_STREAM_DATA(encoding);
5262 unsigned int flags = str->flags;
5263 unsigned int ch = str->ch;
5264 Lisp_Coding_System *codesys = str->codesys;
5265 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE(str->codesys);
5267 Lisp_Object charset;
5270 #ifdef ENABLE_COMPOSITE_CHARS
5271 /* flags for handling composite chars. We do a little switcharoo
5272 on the source while we're outputting the composite char. */
5273 unsigned int saved_n = 0;
5274 const unsigned char *saved_src = NULL;
5275 int in_composite = 0;
5276 #endif /* ENABLE_COMPOSITE_CHARS */
5278 char_boundary = str->iso2022.current_char_boundary;
5279 charset = str->iso2022.current_charset;
5280 half = str->iso2022.current_half;
5282 #ifdef ENABLE_COMPOSITE_CHARS
5288 if (BYTE_ASCII_P(c)) { /* Processing ASCII character */
5291 restore_left_to_right_direction(codesys, dst, &flags,
5294 /* Make sure G0 contains ASCII */
5295 if ((c > ' ' && c < ISO_CODE_DEL) ||
5296 !CODING_SYSTEM_ISO2022_NO_ASCII_CNTL(codesys)) {
5297 ensure_normal_shift(str, dst);
5298 iso2022_designate(Vcharset_ascii, 0, str, dst);
5301 /* If necessary, restore everything to the default state
5304 !(CODING_SYSTEM_ISO2022_NO_ASCII_EOL(codesys))) {
5305 restore_left_to_right_direction(codesys, dst,
5308 ensure_normal_shift(str, dst);
5310 for (i = 0; i < 4; i++) {
5311 Lisp_Object initial_charset =
5312 CODING_SYSTEM_ISO2022_INITIAL_CHARSET
5314 iso2022_designate(initial_charset, i,
5319 if (eol_type != EOL_LF
5320 && eol_type != EOL_AUTODETECT)
5321 Dynarr_add(dst, '\r');
5322 if (eol_type != EOL_CR)
5325 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED(codesys)
5326 && fit_to_be_escape_quoted(c))
5327 Dynarr_add(dst, ISO_CODE_ESC);
5333 else if (BUFBYTE_LEADING_BYTE_P(c) || BUFBYTE_LEADING_BYTE_P(ch)) { /* Processing Leading Byte */
5335 charset = CHARSET_BY_LEADING_BYTE(c);
5336 if (LEADING_BYTE_PREFIX_P(c))
5338 else if (!EQ(charset, Vcharset_control_1)
5339 #ifdef ENABLE_COMPOSITE_CHARS
5340 && !EQ(charset, Vcharset_composite)
5345 ensure_correct_direction(XCHARSET_DIRECTION
5349 /* Now determine which register to use. */
5351 for (i = 0; i < 4; i++) {
5352 if (EQ(charset, str->iso2022.charset[i])
5354 CODING_SYSTEM_ISO2022_INITIAL_CHARSET
5362 if (XCHARSET_GRAPHIC(charset) != 0) {
5364 (str->iso2022.charset[1])
5366 (!CODING_SYSTEM_ISO2022_SEVEN
5369 CODING_SYSTEM_ISO2022_LOCK_SHIFT
5386 iso2022_designate(charset, reg, str, dst);
5388 /* Now invoke that register. */
5391 ensure_normal_shift(str, dst);
5396 if (CODING_SYSTEM_ISO2022_SEVEN
5398 ensure_shift_out(str, dst);
5405 if (CODING_SYSTEM_ISO2022_SEVEN
5407 Dynarr_add(dst, ISO_CODE_ESC);
5408 Dynarr_add(dst, 'N');
5411 Dynarr_add(dst, ISO_CODE_SS2);
5417 if (CODING_SYSTEM_ISO2022_SEVEN
5419 Dynarr_add(dst, ISO_CODE_ESC);
5420 Dynarr_add(dst, 'O');
5423 Dynarr_add(dst, ISO_CODE_SS3);
5433 } else { /* Processing Non-ASCII character */
5434 charmask = (half == 0 ? 0x7F : 0xFF);
5436 if (EQ(charset, Vcharset_control_1)) {
5437 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED(codesys)
5438 && fit_to_be_escape_quoted(c))
5439 Dynarr_add(dst, ISO_CODE_ESC);
5440 /* you asked for it ... */
5441 Dynarr_add(dst, c - 0x20);
5443 switch (XCHARSET_REP_BYTES(charset)) {
5445 Dynarr_add(dst, c & charmask);
5448 if (XCHARSET_PRIVATE_P(charset)) {
5449 Dynarr_add(dst, c & charmask);
5452 #ifdef ENABLE_COMPOSITE_CHARS
5455 Vcharset_composite)) {
5457 /* #### Bother! We don't know how to
5464 (Vcharset_composite,
5469 composite_char_string
5478 n = XSTRING_LENGTH(lstr);
5481 Dynarr_add(dst, '0'); /* start composing */
5484 #endif /* ENABLE_COMPOSITE_CHARS */
5501 Dynarr_add(dst, ch & charmask);
5502 Dynarr_add(dst, c & charmask);
5516 #ifdef ENABLE_COMPOSITE_CHARS
5521 Dynarr_add(dst, ISO_CODE_ESC);
5522 Dynarr_add(dst, '1'); /* end composing */
5523 goto back_to_square_n; /* Wheeeeeeeee ..... */
5525 #endif /* ENABLE_COMPOSITE_CHARS */
5527 if (char_boundary && flags & CODING_STATE_END) {
5528 restore_left_to_right_direction(codesys, dst, &flags, 0);
5529 ensure_normal_shift(str, dst);
5530 for (i = 0; i < 4; i++) {
5531 Lisp_Object initial_charset =
5532 CODING_SYSTEM_ISO2022_INITIAL_CHARSET(codesys, i);
5533 iso2022_designate(initial_charset, i, str, dst);
5539 str->iso2022.current_char_boundary = char_boundary;
5540 str->iso2022.current_charset = charset;
5541 str->iso2022.current_half = half;
5543 /* Verbum caro factum est! */
5547 /************************************************************************/
5548 /* No-conversion methods */
5549 /************************************************************************/
5551 /* This is used when reading in "binary" files -- i.e. files that may
5552 contain all 256 possible byte values and that are not to be
5553 interpreted as being in any particular decoding. */
5555 decode_coding_no_conversion(lstream_t decoding, const Extbyte * src,
5556 unsigned_char_dynarr * dst, Lstream_data_count n)
5558 decoding_stream_t str = DECODING_STREAM_DATA(decoding);
5559 unsigned int flags = str->flags;
5560 unsigned int ch = str->ch;
5561 eol_type_t eol_type = str->eol_type;
5564 const unsigned char c = *(const unsigned char *)src++;
5566 DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst);
5567 DECODE_ADD_BINARY_CHAR(c, dst);
5568 label_continue_loop:;
5571 DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst);
5578 encode_coding_no_conversion(lstream_t encoding, const Bufbyte * src,
5579 unsigned_char_dynarr * dst, Lstream_data_count n)
5582 encoding_stream_t str = ENCODING_STREAM_DATA(encoding);
5583 unsigned int flags = str->flags;
5584 unsigned int ch = str->ch;
5585 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE(str->codesys);
5590 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5591 Dynarr_add(dst, '\r');
5592 if (eol_type != EOL_CR)
5593 Dynarr_add(dst, '\n');
5595 } else if (BYTE_ASCII_P(c)) {
5598 } else if (BUFBYTE_LEADING_BYTE_P(c)) {
5600 if (c == LEADING_BYTE_LATIN_ISO8859_1 ||
5601 c == LEADING_BYTE_CONTROL_1)
5604 Dynarr_add(dst, '~'); /* untranslatable character */
5606 if (ch == LEADING_BYTE_LATIN_ISO8859_1)
5608 else if (ch == LEADING_BYTE_CONTROL_1) {
5610 Dynarr_add(dst, c - 0x20);
5612 /* else it should be the second or third byte of an
5613 untranslatable character, so ignore it */
5622 /************************************************************************/
5623 /* Initialization */
5624 /************************************************************************/
5626 void syms_of_file_coding(void)
5628 INIT_LRECORD_IMPLEMENTATION(coding_system);
5630 DEFERROR_STANDARD(Qcoding_system_error, Qio_error);
5632 DEFSUBR(Fcoding_system_p);
5633 DEFSUBR(Ffind_coding_system);
5634 DEFSUBR(Fget_coding_system);
5635 DEFSUBR(Fcoding_system_list);
5636 DEFSUBR(Fcoding_system_name);
5637 DEFSUBR(Fmake_coding_system);
5638 DEFSUBR(Fcopy_coding_system);
5639 DEFSUBR(Fcoding_system_canonical_name_p);
5640 DEFSUBR(Fcoding_system_alias_p);
5641 DEFSUBR(Fcoding_system_aliasee);
5642 DEFSUBR(Fdefine_coding_system_alias);
5643 DEFSUBR(Fsubsidiary_coding_system);
5645 DEFSUBR(Fcoding_system_type);
5646 DEFSUBR(Fcoding_system_doc_string);
5648 DEFSUBR(Fcoding_system_charset);
5650 DEFSUBR(Fcoding_system_property);
5652 DEFSUBR(Fcoding_category_list);
5653 DEFSUBR(Fset_coding_priority_list);
5654 DEFSUBR(Fcoding_priority_list);
5655 DEFSUBR(Fset_coding_category_system);
5656 DEFSUBR(Fcoding_category_system);
5658 DEFSUBR(Fdetect_coding_region);
5659 DEFSUBR(Fdecode_coding_region);
5660 DEFSUBR(Fencode_coding_region);
5662 DEFSUBR(Fdecode_shift_jis_char);
5663 DEFSUBR(Fencode_shift_jis_char);
5664 DEFSUBR(Fdecode_big5_char);
5665 DEFSUBR(Fencode_big5_char);
5666 DEFSUBR(Fset_ucs_char);
5668 DEFSUBR(Fset_char_ucs);
5671 defsymbol(&Qcoding_systemp, "coding-system-p");
5672 defsymbol(&Qno_conversion, "no-conversion");
5673 defsymbol(&Qraw_text, "raw-text");
5675 defsymbol(&Qbig5, "big5");
5676 defsymbol(&Qshift_jis, "shift-jis");
5677 defsymbol(&Qucs4, "ucs-4");
5678 defsymbol(&Qutf8, "utf-8");
5679 defsymbol(&Qccl, "ccl");
5680 defsymbol(&Qiso2022, "iso2022");
5682 defsymbol(&Qmnemonic, "mnemonic");
5683 defsymbol(&Qeol_type, "eol-type");
5684 defsymbol(&Qpost_read_conversion, "post-read-conversion");
5685 defsymbol(&Qpre_write_conversion, "pre-write-conversion");
5687 defsymbol(&Qcr, "cr");
5688 defsymbol(&Qlf, "lf");
5689 defsymbol(&Qcrlf, "crlf");
5690 defsymbol(&Qeol_cr, "eol-cr");
5691 defsymbol(&Qeol_lf, "eol-lf");
5692 defsymbol(&Qeol_crlf, "eol-crlf");
5694 defsymbol(&Qcharset_g0, "charset-g0");
5695 defsymbol(&Qcharset_g1, "charset-g1");
5696 defsymbol(&Qcharset_g2, "charset-g2");
5697 defsymbol(&Qcharset_g3, "charset-g3");
5698 defsymbol(&Qforce_g0_on_output, "force-g0-on-output");
5699 defsymbol(&Qforce_g1_on_output, "force-g1-on-output");
5700 defsymbol(&Qforce_g2_on_output, "force-g2-on-output");
5701 defsymbol(&Qforce_g3_on_output, "force-g3-on-output");
5702 defsymbol(&Qno_iso6429, "no-iso6429");
5703 defsymbol(&Qinput_charset_conversion, "input-charset-conversion");
5704 defsymbol(&Qoutput_charset_conversion, "output-charset-conversion");
5706 defsymbol(&Qno_ascii_eol, "no-ascii-eol");
5707 defsymbol(&Qno_ascii_cntl, "no-ascii-cntl");
5708 defsymbol(&Qseven, "seven");
5709 defsymbol(&Qlock_shift, "lock-shift");
5710 defsymbol(&Qescape_quoted, "escape-quoted");
5712 defsymbol(&Qencode, "encode");
5713 defsymbol(&Qdecode, "decode");
5716 defsymbol(&coding_category_symbol[CODING_CATEGORY_SHIFT_JIS],
5718 defsymbol(&coding_category_symbol[CODING_CATEGORY_BIG5], "big5");
5719 defsymbol(&coding_category_symbol[CODING_CATEGORY_UCS4], "ucs-4");
5720 defsymbol(&coding_category_symbol[CODING_CATEGORY_UTF8], "utf-8");
5721 defsymbol(&coding_category_symbol[CODING_CATEGORY_ISO_7], "iso-7");
5722 defsymbol(&coding_category_symbol[CODING_CATEGORY_ISO_8_DESIGNATE],
5724 defsymbol(&coding_category_symbol[CODING_CATEGORY_ISO_8_1], "iso-8-1");
5725 defsymbol(&coding_category_symbol[CODING_CATEGORY_ISO_8_2], "iso-8-2");
5726 defsymbol(&coding_category_symbol[CODING_CATEGORY_ISO_LOCK_SHIFT],
5729 defsymbol(&coding_category_symbol[CODING_CATEGORY_NO_CONVERSION],
5733 void lstream_type_create_file_coding(void)
5735 LSTREAM_HAS_METHOD(decoding, reader);
5736 LSTREAM_HAS_METHOD(decoding, writer);
5737 LSTREAM_HAS_METHOD(decoding, rewinder);
5738 LSTREAM_HAS_METHOD(decoding, seekable_p);
5739 LSTREAM_HAS_METHOD(decoding, flusher);
5740 LSTREAM_HAS_METHOD(decoding, closer);
5741 LSTREAM_HAS_METHOD(decoding, marker);
5743 LSTREAM_HAS_METHOD(encoding, reader);
5744 LSTREAM_HAS_METHOD(encoding, writer);
5745 LSTREAM_HAS_METHOD(encoding, rewinder);
5746 LSTREAM_HAS_METHOD(encoding, seekable_p);
5747 LSTREAM_HAS_METHOD(encoding, flusher);
5748 LSTREAM_HAS_METHOD(encoding, closer);
5749 LSTREAM_HAS_METHOD(encoding, marker);
5752 void vars_of_file_coding(void)
5756 fcd = xnew(struct file_coding_dump);
5757 dump_add_root_struct_ptr(&fcd, &fcd_description);
5759 /* Initialize to something reasonable ... */
5760 for (i = 0; i < CODING_CATEGORY_LAST; i++) {
5761 fcd->coding_category_system[i] = Qnil;
5762 fcd->coding_category_by_priority[i] = i;
5765 Fprovide(intern("file-coding"));
5767 DEFVAR_LISP("keyboard-coding-system", &Vkeyboard_coding_system /*
5768 Coding system used for TTY keyboard input.
5769 Not used under a windowing system.
5771 Vkeyboard_coding_system = Qnil;
5773 DEFVAR_LISP("terminal-coding-system", &Vterminal_coding_system /*
5774 Coding system used for TTY display output.
5775 Not used under a windowing system.
5777 Vterminal_coding_system = Qnil;
5779 DEFVAR_LISP("coding-system-for-read", &Vcoding_system_for_read /*
5780 Overriding coding system used when reading from a file or process.
5781 You should bind this variable with `let', but do not set it globally.
5782 If this is non-nil, it specifies the coding system that will be used
5783 to decode input on read operations, such as from a file or process.
5784 It overrides `buffer-file-coding-system-for-read',
5785 `insert-file-contents-pre-hook', etc. Use those variables instead of
5786 this one for permanent changes to the environment. */ );
5787 Vcoding_system_for_read = Qnil;
5789 DEFVAR_LISP("coding-system-for-write", &Vcoding_system_for_write /*
5790 Overriding coding system used when writing to a file or process.
5791 You should bind this variable with `let', but do not set it globally.
5792 If this is non-nil, it specifies the coding system that will be used
5793 to encode output for write operations, such as to a file or process.
5794 It overrides `buffer-file-coding-system', `write-region-pre-hook', etc.
5795 Use those variables instead of this one for permanent changes to the
5797 Vcoding_system_for_write = Qnil;
5799 DEFVAR_LISP("file-name-coding-system", &Vfile_name_coding_system /*
5800 Coding system used to convert pathnames when accessing files.
5802 Vfile_name_coding_system = Qnil;
5804 DEFVAR_BOOL("enable-multibyte-characters", &enable_multibyte_characters /*
5805 Non-nil means the buffer contents are regarded as multi-byte form
5806 of characters, not a binary code. This affects the display, file I/O,
5807 and behaviors of various editing commands.
5809 Setting this to nil does not do anything.
5811 enable_multibyte_characters = 1;
5814 void complex_vars_of_file_coding(void)
5816 staticpro(&Vcoding_system_hash_table);
5817 Vcoding_system_hash_table =
5818 make_lisp_hash_table(50, HASH_TABLE_NON_WEAK, HASH_TABLE_EQ);
5820 the_codesys_prop_dynarr = Dynarr_new(codesys_prop);
5821 dump_add_root_struct_ptr(&the_codesys_prop_dynarr,
5822 &codesys_prop_dynarr_description);
5824 #define DEFINE_CODESYS_PROP(Prop_Type, Sym) do \
5826 struct codesys_prop csp; \
5828 csp.prop_type = (Prop_Type); \
5829 Dynarr_add (the_codesys_prop_dynarr, csp); \
5832 DEFINE_CODESYS_PROP(CODESYS_PROP_ALL_OK, Qmnemonic);
5833 DEFINE_CODESYS_PROP(CODESYS_PROP_ALL_OK, Qeol_type);
5834 DEFINE_CODESYS_PROP(CODESYS_PROP_ALL_OK, Qeol_cr);
5835 DEFINE_CODESYS_PROP(CODESYS_PROP_ALL_OK, Qeol_crlf);
5836 DEFINE_CODESYS_PROP(CODESYS_PROP_ALL_OK, Qeol_lf);
5837 DEFINE_CODESYS_PROP(CODESYS_PROP_ALL_OK, Qpost_read_conversion);
5838 DEFINE_CODESYS_PROP(CODESYS_PROP_ALL_OK, Qpre_write_conversion);
5840 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qcharset_g0);
5841 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qcharset_g1);
5842 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qcharset_g2);
5843 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qcharset_g3);
5844 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qforce_g0_on_output);
5845 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qforce_g1_on_output);
5846 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qforce_g2_on_output);
5847 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qforce_g3_on_output);
5848 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qshort);
5849 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qno_ascii_eol);
5850 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qno_ascii_cntl);
5851 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qseven);
5852 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qlock_shift);
5853 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qno_iso6429);
5854 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qescape_quoted);
5855 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qinput_charset_conversion);
5856 DEFINE_CODESYS_PROP(CODESYS_PROP_ISO2022, Qoutput_charset_conversion);
5858 DEFINE_CODESYS_PROP(CODESYS_PROP_CCL, Qencode);
5859 DEFINE_CODESYS_PROP(CODESYS_PROP_CCL, Qdecode);
5861 /* Need to create this here or we're really screwed. */
5863 (Qraw_text, Qno_conversion,
5865 ("Raw text, which means it converts only line-break-codes."),
5866 list2(Qmnemonic, build_string("Raw")));
5869 (Qbinary, Qno_conversion,
5870 build_string("Binary, which means it does not convert anything."),
5871 list4(Qeol_type, Qlf, Qmnemonic, build_string("Binary")));
5873 Fdefine_coding_system_alias(Qno_conversion, Qraw_text);
5875 Fdefine_coding_system_alias(Qfile_name, Qbinary);
5877 Fdefine_coding_system_alias(Qterminal, Qbinary);
5878 Fdefine_coding_system_alias(Qkeyboard, Qbinary);
5880 /* Need this for bootstrapping */
5881 fcd->coding_category_system[CODING_CATEGORY_NO_CONVERSION] =
5882 Fget_coding_system(Qraw_text);
5888 for (i = 0; i < countof(fcd->ucs_to_mule_table); i++)
5889 fcd->ucs_to_mule_table[i] = Qnil;
5891 staticpro(&mule_to_ucs_table);
5892 mule_to_ucs_table = Fmake_char_table(Qgeneric);