1 /* Code conversion functions.
2 Copyright (C) 1991, 1995 Free Software Foundation, Inc.
3 Copyright (C) 1995 Sun Microsystems, Inc.
4 Copyright (C) 1999,2000,2001,2002 MORIOKA Tomohiko
6 This file is part of XEmacs.
8 XEmacs is free software; you can redistribute it and/or modify it
9 under the terms of the GNU General Public License as published by the
10 Free Software Foundation; either version 2, or (at your option) any
13 XEmacs is distributed in the hope that it will be useful, but WITHOUT
14 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
18 You should have received a copy of the GNU General Public License
19 along with XEmacs; see the file COPYING. If not, write to
20 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
21 Boston, MA 02111-1307, USA. */
23 /* Synched up with: Mule 2.3. Not in FSF. */
25 /* Rewritten by Ben Wing <ben@xemacs.org>. */
26 /* Rewritten by MORIOKA Tomohiko <tomo@m17n.org> for XEmacs UTF-2000. */
40 #include "file-coding.h"
42 Lisp_Object Qcoding_system_error;
44 Lisp_Object Vkeyboard_coding_system;
45 Lisp_Object Vterminal_coding_system;
46 Lisp_Object Vcoding_system_for_read;
47 Lisp_Object Vcoding_system_for_write;
48 Lisp_Object Vfile_name_coding_system;
50 Lisp_Object Vcoded_charset_entity_reference_alist;
52 /* Table of symbols identifying each coding category. */
53 Lisp_Object coding_category_symbol[CODING_CATEGORY_LAST];
57 struct file_coding_dump {
58 /* Coding system currently associated with each coding category. */
59 Lisp_Object coding_category_system[CODING_CATEGORY_LAST];
61 /* Table of all coding categories in decreasing order of priority.
62 This describes a permutation of the possible coding categories. */
63 int coding_category_by_priority[CODING_CATEGORY_LAST];
65 #if defined(MULE) && !defined(UTF2000)
66 Lisp_Object ucs_to_mule_table[65536];
70 static const struct lrecord_description fcd_description_1[] = {
71 { XD_LISP_OBJECT_ARRAY, offsetof (struct file_coding_dump, coding_category_system), CODING_CATEGORY_LAST },
72 #if defined(MULE) && !defined(UTF2000)
73 { XD_LISP_OBJECT_ARRAY, offsetof (struct file_coding_dump, ucs_to_mule_table), countof (fcd->ucs_to_mule_table) },
78 static const struct struct_description fcd_description = {
79 sizeof (struct file_coding_dump),
83 Lisp_Object mule_to_ucs_table;
85 Lisp_Object Qcoding_systemp;
87 Lisp_Object Qraw_text, Qno_conversion, Qccl, Qiso2022;
88 /* Qinternal in general.c */
90 Lisp_Object Qmnemonic, Qeol_type;
91 Lisp_Object Qcr, Qcrlf, Qlf;
92 Lisp_Object Qeol_cr, Qeol_crlf, Qeol_lf;
93 Lisp_Object Qpost_read_conversion;
94 Lisp_Object Qpre_write_conversion;
97 Lisp_Object Qucs4, Qutf16, Qutf8;
98 Lisp_Object Qbig5, Qshift_jis;
99 Lisp_Object Qcharset_g0, Qcharset_g1, Qcharset_g2, Qcharset_g3;
100 Lisp_Object Qforce_g0_on_output, Qforce_g1_on_output;
101 Lisp_Object Qforce_g2_on_output, Qforce_g3_on_output;
102 Lisp_Object Qno_iso6429;
103 Lisp_Object Qinput_charset_conversion, Qoutput_charset_conversion;
104 Lisp_Object Qescape_quoted;
105 Lisp_Object Qshort, Qno_ascii_eol, Qno_ascii_cntl, Qseven, Qlock_shift;
108 Lisp_Object Qutf_8_mcs;
109 Lisp_Object Qdisable_composition;
110 Lisp_Object Quse_entity_reference;
111 Lisp_Object Qd, Qx, QX;
113 Lisp_Object Qencode, Qdecode;
115 Lisp_Object Vcoding_system_hash_table;
117 int enable_multibyte_characters;
120 /* Additional information used by the ISO2022 decoder and detector. */
121 struct iso2022_decoder
123 /* CHARSET holds the character sets currently assigned to the G0
124 through G3 variables. It is initialized from the array
125 INITIAL_CHARSET in CODESYS. */
126 Lisp_Object charset[4];
128 /* Which registers are currently invoked into the left (GL) and
129 right (GR) halves of the 8-bit encoding space? */
130 int register_left, register_right;
132 /* ISO_ESC holds a value indicating part of an escape sequence
133 that has already been seen. */
134 enum iso_esc_flag esc;
136 /* This records the bytes we've seen so far in an escape sequence,
137 in case the sequence is invalid (we spit out the bytes unchanged). */
138 unsigned char esc_bytes[8];
140 /* Index for next byte to store in ISO escape sequence. */
143 #ifdef ENABLE_COMPOSITE_CHARS
144 /* Stuff seen so far when composing a string. */
145 unsigned_char_dynarr *composite_chars;
148 /* If we saw an invalid designation sequence for a particular
149 register, we flag it here and switch to ASCII. The next time we
150 see a valid designation for this register, we turn off the flag
151 and do the designation normally, but pretend the sequence was
152 invalid. The effect of all this is that (most of the time) the
153 escape sequences for both the switch to the unknown charset, and
154 the switch back to the known charset, get inserted literally into
155 the buffer and saved out as such. The hope is that we can
156 preserve the escape sequences so that the resulting written out
157 file makes sense. If we don't do any of this, the designation
158 to the invalid charset will be preserved but that switch back
159 to the known charset will probably get eaten because it was
160 the same charset that was already present in the register. */
161 unsigned char invalid_designated[4];
163 /* We try to do similar things as above for direction-switching
164 sequences. If we encountered a direction switch while an
165 invalid designation was present, or an invalid designation
166 just after a direction switch (i.e. no valid designation
167 encountered yet), we insert the direction-switch escape
168 sequence literally into the output stream, and later on
169 insert the corresponding direction-restoring escape sequence
171 unsigned int switched_dir_and_no_valid_charset_yet :1;
172 unsigned int invalid_switch_dir :1;
174 /* Tells the decoder to output the escape sequence literally
175 even though it was valid. Used in the games we play to
176 avoid lossage when we encounter invalid designations. */
177 unsigned int output_literally :1;
178 /* We encountered a direction switch followed by an invalid
179 designation. We didn't output the direction switch
180 literally because we didn't know about the invalid designation;
181 but we have to do so now. */
182 unsigned int output_direction_sequence :1;
185 EXFUN (Fcopy_coding_system, 2);
187 struct detection_state;
190 text_encode_generic (Lstream *encoding, const Bufbyte *src,
191 unsigned_char_dynarr *dst, Lstream_data_count n);
193 static int detect_coding_sjis (struct detection_state *st,
194 const Extbyte *src, Lstream_data_count n);
195 static void decode_coding_sjis (Lstream *decoding, const Extbyte *src,
196 unsigned_char_dynarr *dst, Lstream_data_count n);
197 void char_encode_shift_jis (struct encoding_stream *str, Emchar c,
198 unsigned_char_dynarr *dst, unsigned int *flags);
199 void char_finish_shift_jis (struct encoding_stream *str,
200 unsigned_char_dynarr *dst, unsigned int *flags);
202 static int detect_coding_big5 (struct detection_state *st,
203 const Extbyte *src, Lstream_data_count n);
204 static void decode_coding_big5 (Lstream *decoding, const Extbyte *src,
205 unsigned_char_dynarr *dst, Lstream_data_count n);
206 void char_encode_big5 (struct encoding_stream *str, Emchar c,
207 unsigned_char_dynarr *dst, unsigned int *flags);
208 void char_finish_big5 (struct encoding_stream *str,
209 unsigned_char_dynarr *dst, unsigned int *flags);
211 static int detect_coding_ucs4 (struct detection_state *st,
212 const Extbyte *src, Lstream_data_count n);
213 static void decode_coding_ucs4 (Lstream *decoding, const Extbyte *src,
214 unsigned_char_dynarr *dst, Lstream_data_count n);
215 void char_encode_ucs4 (struct encoding_stream *str, Emchar c,
216 unsigned_char_dynarr *dst, unsigned int *flags);
217 void char_finish_ucs4 (struct encoding_stream *str,
218 unsigned_char_dynarr *dst, unsigned int *flags);
220 static int detect_coding_utf16 (struct detection_state *st,
221 const Extbyte *src, Lstream_data_count n);
222 static void decode_coding_utf16 (Lstream *decoding, const Extbyte *src,
223 unsigned_char_dynarr *dst, Lstream_data_count n);
224 void char_encode_utf16 (struct encoding_stream *str, Emchar c,
225 unsigned_char_dynarr *dst, unsigned int *flags);
226 void char_finish_utf16 (struct encoding_stream *str,
227 unsigned_char_dynarr *dst, unsigned int *flags);
229 static int detect_coding_utf8 (struct detection_state *st,
230 const Extbyte *src, Lstream_data_count n);
231 static void decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
232 unsigned_char_dynarr *dst, Lstream_data_count n);
233 void char_encode_utf8 (struct encoding_stream *str, Emchar c,
234 unsigned_char_dynarr *dst, unsigned int *flags);
235 void char_finish_utf8 (struct encoding_stream *str,
236 unsigned_char_dynarr *dst, unsigned int *flags);
238 static int postprocess_iso2022_mask (int mask);
239 static void reset_iso2022 (Lisp_Object coding_system,
240 struct iso2022_decoder *iso);
241 static int detect_coding_iso2022 (struct detection_state *st,
242 const Extbyte *src, Lstream_data_count n);
243 static void decode_coding_iso2022 (Lstream *decoding, const Extbyte *src,
244 unsigned_char_dynarr *dst, Lstream_data_count n);
245 void char_encode_iso2022 (struct encoding_stream *str, Emchar c,
246 unsigned_char_dynarr *dst, unsigned int *flags);
247 void char_finish_iso2022 (struct encoding_stream *str,
248 unsigned_char_dynarr *dst, unsigned int *flags);
250 static void decode_coding_no_conversion (Lstream *decoding, const Extbyte *src,
251 unsigned_char_dynarr *dst, Lstream_data_count n);
252 static void encode_coding_no_conversion (Lstream *encoding, const Bufbyte *src,
253 unsigned_char_dynarr *dst, Lstream_data_count n);
254 static void mule_decode (Lstream *decoding, const Extbyte *src,
255 unsigned_char_dynarr *dst, Lstream_data_count n);
256 static void mule_encode (Lstream *encoding, const Bufbyte *src,
257 unsigned_char_dynarr *dst, Lstream_data_count n);
259 typedef struct codesys_prop codesys_prop;
268 Dynarr_declare (codesys_prop);
269 } codesys_prop_dynarr;
271 static const struct lrecord_description codesys_prop_description_1[] = {
272 { XD_LISP_OBJECT, offsetof (codesys_prop, sym) },
276 static const struct struct_description codesys_prop_description = {
277 sizeof (codesys_prop),
278 codesys_prop_description_1
281 static const struct lrecord_description codesys_prop_dynarr_description_1[] = {
282 XD_DYNARR_DESC (codesys_prop_dynarr, &codesys_prop_description),
286 static const struct struct_description codesys_prop_dynarr_description = {
287 sizeof (codesys_prop_dynarr),
288 codesys_prop_dynarr_description_1
291 codesys_prop_dynarr *the_codesys_prop_dynarr;
293 enum codesys_prop_enum
296 CODESYS_PROP_ISO2022,
301 /************************************************************************/
302 /* Coding system functions */
303 /************************************************************************/
305 static Lisp_Object mark_coding_system (Lisp_Object);
306 static void print_coding_system (Lisp_Object, Lisp_Object, int);
307 static void finalize_coding_system (void *header, int for_disksave);
310 static const struct lrecord_description ccs_description_1[] = {
311 { XD_LISP_OBJECT, offsetof (charset_conversion_spec, from_charset) },
312 { XD_LISP_OBJECT, offsetof (charset_conversion_spec, to_charset) },
316 static const struct struct_description ccs_description = {
317 sizeof (charset_conversion_spec),
321 static const struct lrecord_description ccsd_description_1[] = {
322 XD_DYNARR_DESC (charset_conversion_spec_dynarr, &ccs_description),
326 static const struct struct_description ccsd_description = {
327 sizeof (charset_conversion_spec_dynarr),
332 static const struct lrecord_description coding_system_description[] = {
333 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, name) },
334 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, doc_string) },
335 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, mnemonic) },
336 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, post_read_conversion) },
337 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, pre_write_conversion) },
338 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_lf) },
339 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_crlf) },
340 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_cr) },
342 { XD_LISP_OBJECT_ARRAY, offsetof (Lisp_Coding_System, iso2022.initial_charset), 4 },
343 { XD_STRUCT_PTR, offsetof (Lisp_Coding_System, iso2022.input_conv), 1, &ccsd_description },
344 { XD_STRUCT_PTR, offsetof (Lisp_Coding_System, iso2022.output_conv), 1, &ccsd_description },
345 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccl.decode) },
346 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccl.encode) },
348 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccs_priority_list) },
354 DEFINE_LRECORD_IMPLEMENTATION ("coding-system", coding_system,
355 mark_coding_system, print_coding_system,
356 finalize_coding_system,
357 0, 0, coding_system_description,
361 mark_coding_system (Lisp_Object obj)
363 Lisp_Coding_System *codesys = XCODING_SYSTEM (obj);
365 mark_object (CODING_SYSTEM_NAME (codesys));
366 mark_object (CODING_SYSTEM_DOC_STRING (codesys));
367 mark_object (CODING_SYSTEM_MNEMONIC (codesys));
368 mark_object (CODING_SYSTEM_EOL_LF (codesys));
369 mark_object (CODING_SYSTEM_EOL_CRLF (codesys));
370 mark_object (CODING_SYSTEM_EOL_CR (codesys));
372 switch (CODING_SYSTEM_TYPE (codesys))
376 case CODESYS_ISO2022:
377 for (i = 0; i < 4; i++)
378 mark_object (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i));
379 if (codesys->iso2022.input_conv)
381 for (i = 0; i < Dynarr_length (codesys->iso2022.input_conv); i++)
383 struct charset_conversion_spec *ccs =
384 Dynarr_atp (codesys->iso2022.input_conv, i);
385 mark_object (ccs->from_charset);
386 mark_object (ccs->to_charset);
389 if (codesys->iso2022.output_conv)
391 for (i = 0; i < Dynarr_length (codesys->iso2022.output_conv); i++)
393 struct charset_conversion_spec *ccs =
394 Dynarr_atp (codesys->iso2022.output_conv, i);
395 mark_object (ccs->from_charset);
396 mark_object (ccs->to_charset);
403 mark_object (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 0));
404 mark_object (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 1));
409 mark_object (CODING_SYSTEM_CCL_DECODE (codesys));
410 mark_object (CODING_SYSTEM_CCL_ENCODE (codesys));
417 mark_object (CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys));
419 mark_object (CODING_SYSTEM_CCS_PRIORITY_LIST (codesys));
421 return CODING_SYSTEM_POST_READ_CONVERSION (codesys);
425 print_coding_system (Lisp_Object obj, Lisp_Object printcharfun,
428 Lisp_Coding_System *c = XCODING_SYSTEM (obj);
430 error ("printing unreadable object #<coding_system 0x%x>",
433 write_c_string ("#<coding_system ", printcharfun);
434 print_internal (c->name, printcharfun, 1);
435 write_c_string (">", printcharfun);
439 finalize_coding_system (void *header, int for_disksave)
441 Lisp_Coding_System *c = (Lisp_Coding_System *) header;
442 /* Since coding systems never go away, this function is not
443 necessary. But it would be necessary if we changed things
444 so that coding systems could go away. */
445 if (!for_disksave) /* see comment in lstream.c */
447 switch (CODING_SYSTEM_TYPE (c))
450 case CODESYS_ISO2022:
451 if (c->iso2022.input_conv)
453 Dynarr_free (c->iso2022.input_conv);
454 c->iso2022.input_conv = 0;
456 if (c->iso2022.output_conv)
458 Dynarr_free (c->iso2022.output_conv);
459 c->iso2022.output_conv = 0;
470 symbol_to_eol_type (Lisp_Object symbol)
472 CHECK_SYMBOL (symbol);
473 if (NILP (symbol)) return EOL_AUTODETECT;
474 if (EQ (symbol, Qlf)) return EOL_LF;
475 if (EQ (symbol, Qcrlf)) return EOL_CRLF;
476 if (EQ (symbol, Qcr)) return EOL_CR;
478 signal_simple_error ("Unrecognized eol type", symbol);
479 return EOL_AUTODETECT; /* not reached */
483 eol_type_to_symbol (eol_type_t type)
488 case EOL_LF: return Qlf;
489 case EOL_CRLF: return Qcrlf;
490 case EOL_CR: return Qcr;
491 case EOL_AUTODETECT: return Qnil;
496 setup_eol_coding_systems (Lisp_Coding_System *codesys)
498 Lisp_Object codesys_obj;
499 int len = string_length (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name);
500 char *codesys_name = (char *) alloca (len + 7);
502 char *codesys_mnemonic=0;
504 Lisp_Object codesys_name_sym, sub_codesys_obj;
508 XSETCODING_SYSTEM (codesys_obj, codesys);
510 memcpy (codesys_name,
511 string_data (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name), len);
513 if (STRINGP (CODING_SYSTEM_MNEMONIC (codesys)))
515 mlen = XSTRING_LENGTH (CODING_SYSTEM_MNEMONIC (codesys));
516 codesys_mnemonic = (char *) alloca (mlen + 7);
517 memcpy (codesys_mnemonic,
518 XSTRING_DATA (CODING_SYSTEM_MNEMONIC (codesys)), mlen);
521 #define DEFINE_SUB_CODESYS(op_sys, op_sys_abbr, Type) do { \
522 strcpy (codesys_name + len, "-" op_sys); \
524 strcpy (codesys_mnemonic + mlen, op_sys_abbr); \
525 codesys_name_sym = intern (codesys_name); \
526 sub_codesys_obj = Fcopy_coding_system (codesys_obj, codesys_name_sym); \
527 XCODING_SYSTEM_EOL_TYPE (sub_codesys_obj) = Type; \
529 XCODING_SYSTEM_MNEMONIC(sub_codesys_obj) = \
530 build_string (codesys_mnemonic); \
531 CODING_SYSTEM_##Type (codesys) = sub_codesys_obj; \
534 DEFINE_SUB_CODESYS("unix", "", EOL_LF);
535 DEFINE_SUB_CODESYS("dos", ":T", EOL_CRLF);
536 DEFINE_SUB_CODESYS("mac", ":t", EOL_CR);
539 DEFUN ("coding-system-p", Fcoding_system_p, 1, 1, 0, /*
540 Return t if OBJECT is a coding system.
541 A coding system is an object that defines how text containing multiple
542 character sets is encoded into a stream of (typically 8-bit) bytes.
543 The coding system is used to decode the stream into a series of
544 characters (which may be from multiple charsets) when the text is read
545 from a file or process, and is used to encode the text back into the
546 same format when it is written out to a file or process.
548 For example, many ISO2022-compliant coding systems (such as Compound
549 Text, which is used for inter-client data under the X Window System)
550 use escape sequences to switch between different charsets -- Japanese
551 Kanji, for example, is invoked with "ESC $ ( B"; ASCII is invoked
552 with "ESC ( B"; and Cyrillic is invoked with "ESC - L". See
553 `make-coding-system' for more information.
555 Coding systems are normally identified using a symbol, and the
556 symbol is accepted in place of the actual coding system object whenever
557 a coding system is called for. (This is similar to how faces work.)
561 return CODING_SYSTEMP (object) ? Qt : Qnil;
564 DEFUN ("find-coding-system", Ffind_coding_system, 1, 1, 0, /*
565 Retrieve the coding system of the given name.
567 If CODING-SYSTEM-OR-NAME is a coding-system object, it is simply
568 returned. Otherwise, CODING-SYSTEM-OR-NAME should be a symbol.
569 If there is no such coding system, nil is returned. Otherwise the
570 associated coding system object is returned.
572 (coding_system_or_name))
574 if (NILP (coding_system_or_name))
575 coding_system_or_name = Qbinary;
576 else if (CODING_SYSTEMP (coding_system_or_name))
577 return coding_system_or_name;
579 CHECK_SYMBOL (coding_system_or_name);
583 coding_system_or_name =
584 Fgethash (coding_system_or_name, Vcoding_system_hash_table, Qnil);
586 if (CODING_SYSTEMP (coding_system_or_name) || NILP (coding_system_or_name))
587 return coding_system_or_name;
591 DEFUN ("get-coding-system", Fget_coding_system, 1, 1, 0, /*
592 Retrieve the coding system of the given name.
593 Same as `find-coding-system' except that if there is no such
594 coding system, an error is signaled instead of returning nil.
598 Lisp_Object coding_system = Ffind_coding_system (name);
600 if (NILP (coding_system))
601 signal_simple_error ("No such coding system", name);
602 return coding_system;
605 /* We store the coding systems in hash tables with the names as the key and the
606 actual coding system object as the value. Occasionally we need to use them
607 in a list format. These routines provide us with that. */
608 struct coding_system_list_closure
610 Lisp_Object *coding_system_list;
614 add_coding_system_to_list_mapper (Lisp_Object key, Lisp_Object value,
615 void *coding_system_list_closure)
617 /* This function can GC */
618 struct coding_system_list_closure *cscl =
619 (struct coding_system_list_closure *) coding_system_list_closure;
620 Lisp_Object *coding_system_list = cscl->coding_system_list;
622 *coding_system_list = Fcons (key, *coding_system_list);
626 DEFUN ("coding-system-list", Fcoding_system_list, 0, 0, 0, /*
627 Return a list of the names of all defined coding systems.
631 Lisp_Object coding_system_list = Qnil;
633 struct coding_system_list_closure coding_system_list_closure;
635 GCPRO1 (coding_system_list);
636 coding_system_list_closure.coding_system_list = &coding_system_list;
637 elisp_maphash (add_coding_system_to_list_mapper, Vcoding_system_hash_table,
638 &coding_system_list_closure);
641 return coding_system_list;
644 DEFUN ("coding-system-name", Fcoding_system_name, 1, 1, 0, /*
645 Return the name of the given coding system.
649 coding_system = Fget_coding_system (coding_system);
650 return XCODING_SYSTEM_NAME (coding_system);
653 static Lisp_Coding_System *
654 allocate_coding_system (enum coding_system_type type, Lisp_Object name)
656 Lisp_Coding_System *codesys =
657 alloc_lcrecord_type (Lisp_Coding_System, &lrecord_coding_system);
659 zero_lcrecord (codesys);
660 CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = Qnil;
661 CODING_SYSTEM_POST_READ_CONVERSION (codesys) = Qnil;
662 CODING_SYSTEM_EOL_TYPE (codesys) = EOL_AUTODETECT;
663 CODING_SYSTEM_EOL_CRLF (codesys) = Qnil;
664 CODING_SYSTEM_EOL_CR (codesys) = Qnil;
665 CODING_SYSTEM_EOL_LF (codesys) = Qnil;
666 CODING_SYSTEM_TYPE (codesys) = type;
667 CODING_SYSTEM_MNEMONIC (codesys) = Qnil;
670 CODING_SYSTEM_CCS_PRIORITY_LIST (codesys) = Qnil;
672 if (type == CODESYS_ISO2022)
675 for (i = 0; i < 4; i++)
676 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i) = Qnil;
679 if (type == CODESYS_UTF8)
681 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 0)
683 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 1)
685 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 2)
687 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 3)
690 else if (type == CODESYS_BIG5)
692 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 0)
694 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 1)
695 = Vcharset_chinese_big5;
696 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 2)
698 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 3)
702 else if (type == CODESYS_CCL)
704 CODING_SYSTEM_CCL_DECODE (codesys) = Qnil;
705 CODING_SYSTEM_CCL_ENCODE (codesys) = Qnil;
708 CODING_SYSTEM_NAME (codesys) = name;
714 /* Given a list of charset conversion specs as specified in a Lisp
715 program, parse it into STORE_HERE. */
718 parse_charset_conversion_specs (charset_conversion_spec_dynarr *store_here,
719 Lisp_Object spec_list)
723 EXTERNAL_LIST_LOOP (rest, spec_list)
725 Lisp_Object car = XCAR (rest);
726 Lisp_Object from, to;
727 struct charset_conversion_spec spec;
729 if (!CONSP (car) || !CONSP (XCDR (car)) || !NILP (XCDR (XCDR (car))))
730 signal_simple_error ("Invalid charset conversion spec", car);
731 from = Fget_charset (XCAR (car));
732 to = Fget_charset (XCAR (XCDR (car)));
733 if ( (XCHARSET_CHARS (from) != XCHARSET_CHARS (to)) ||
734 (XCHARSET_DIMENSION (from) != XCHARSET_DIMENSION (to)) )
735 signal_simple_error_2
736 ("Attempted conversion between different charset types",
738 spec.from_charset = from;
739 spec.to_charset = to;
741 Dynarr_add (store_here, spec);
745 /* Given a dynarr LOAD_HERE of internally-stored charset conversion
746 specs, return the equivalent as the Lisp programmer would see it.
748 If LOAD_HERE is 0, return Qnil. */
751 unparse_charset_conversion_specs (charset_conversion_spec_dynarr *load_here)
758 for (i = 0, result = Qnil; i < Dynarr_length (load_here); i++)
760 struct charset_conversion_spec *ccs = Dynarr_atp (load_here, i);
761 result = Fcons (list2 (ccs->from_charset, ccs->to_charset), result);
764 return Fnreverse (result);
769 DEFUN ("make-coding-system", Fmake_coding_system, 2, 4, 0, /*
770 Register symbol NAME as a coding system.
772 TYPE describes the conversion method used and should be one of
775 Automatic conversion. XEmacs attempts to detect the coding system
778 No conversion. Use this for binary files and such. On output,
779 graphic characters that are not in ASCII or Latin-1 will be
780 replaced by a ?. (For a no-conversion-encoded buffer, these
781 characters will only be present if you explicitly insert them.)
783 Shift-JIS (a Japanese encoding commonly used in PC operating systems).
785 ISO 10646 UCS-4 encoding.
787 ISO 10646 UTF-8 encoding.
789 Any ISO2022-compliant encoding. Among other things, this includes
790 JIS (the Japanese encoding commonly used for e-mail), EUC (the
791 standard Unix encoding for Japanese and other languages), and
792 Compound Text (the encoding used in X11). You can specify more
793 specific information about the conversion with the PROPS argument.
795 Big5 (the encoding commonly used for Taiwanese).
797 The conversion is performed using a user-written pseudo-code
798 program. CCL (Code Conversion Language) is the name of this
801 Write out or read in the raw contents of the memory representing
802 the buffer's text. This is primarily useful for debugging
803 purposes, and is only enabled when XEmacs has been compiled with
804 DEBUG_XEMACS defined (via the --debug configure option).
805 WARNING: Reading in a file using 'internal conversion can result
806 in an internal inconsistency in the memory representing a
807 buffer's text, which will produce unpredictable results and may
808 cause XEmacs to crash. Under normal circumstances you should
809 never use 'internal conversion.
811 DOC-STRING is a string describing the coding system.
813 PROPS is a property list, describing the specific nature of the
814 character set. Recognized properties are:
817 String to be displayed in the modeline when this coding system is
821 End-of-line conversion to be used. It should be one of
824 Automatically detect the end-of-line type (LF, CRLF,
825 or CR). Also generate subsidiary coding systems named
826 `NAME-unix', `NAME-dos', and `NAME-mac', that are
827 identical to this coding system but have an EOL-TYPE
828 value of 'lf, 'crlf, and 'cr, respectively.
830 The end of a line is marked externally using ASCII LF.
831 Since this is also the way that XEmacs represents an
832 end-of-line internally, specifying this option results
833 in no end-of-line conversion. This is the standard
834 format for Unix text files.
836 The end of a line is marked externally using ASCII
837 CRLF. This is the standard format for MS-DOS text
840 The end of a line is marked externally using ASCII CR.
841 This is the standard format for Macintosh text files.
843 Automatically detect the end-of-line type but do not
844 generate subsidiary coding systems. (This value is
845 converted to nil when stored internally, and
846 `coding-system-property' will return nil.)
849 If non-nil, composition/decomposition for combining characters
852 'use-entity-reference
853 If non-nil, SGML style entity-reference is used for non-system-characters.
855 'post-read-conversion
856 Function called after a file has been read in, to perform the
857 decoding. Called with two arguments, START and END, denoting
858 a region of the current buffer to be decoded.
860 'pre-write-conversion
861 Function called before a file is written out, to perform the
862 encoding. Called with two arguments, START and END, denoting
863 a region of the current buffer to be encoded.
866 The following additional properties are recognized if TYPE is 'iso2022:
872 The character set initially designated to the G0 - G3 registers.
873 The value should be one of
875 -- A charset object (designate that character set)
876 -- nil (do not ever use this register)
877 -- t (no character set is initially designated to
878 the register, but may be later on; this automatically
879 sets the corresponding `force-g*-on-output' property)
885 If non-nil, send an explicit designation sequence on output before
886 using the specified register.
889 If non-nil, use the short forms "ESC $ @", "ESC $ A", and
890 "ESC $ B" on output in place of the full designation sequences
891 "ESC $ ( @", "ESC $ ( A", and "ESC $ ( B".
894 If non-nil, don't designate ASCII to G0 at each end of line on output.
895 Setting this to non-nil also suppresses other state-resetting that
896 normally happens at the end of a line.
899 If non-nil, don't designate ASCII to G0 before control chars on output.
902 If non-nil, use 7-bit environment on output. Otherwise, use 8-bit
906 If non-nil, use locking-shift (SO/SI) instead of single-shift
907 or designation by escape sequence.
910 If non-nil, don't use ISO6429's direction specification.
913 If non-nil, literal control characters that are the same as
914 the beginning of a recognized ISO2022 or ISO6429 escape sequence
915 (in particular, ESC (0x1B), SO (0x0E), SI (0x0F), SS2 (0x8E),
916 SS3 (0x8F), and CSI (0x9B)) are "quoted" with an escape character
917 so that they can be properly distinguished from an escape sequence.
918 (Note that doing this results in a non-portable encoding.) This
919 encoding flag is used for byte-compiled files. Note that ESC
920 is a good choice for a quoting character because there are no
921 escape sequences whose second byte is a character from the Control-0
922 or Control-1 character sets; this is explicitly disallowed by the
925 'input-charset-conversion
926 A list of conversion specifications, specifying conversion of
927 characters in one charset to another when decoding is performed.
928 Each specification is a list of two elements: the source charset,
929 and the destination charset.
931 'output-charset-conversion
932 A list of conversion specifications, specifying conversion of
933 characters in one charset to another when encoding is performed.
934 The form of each specification is the same as for
935 'input-charset-conversion.
938 The following additional properties are recognized (and required)
942 CCL program used for decoding (converting to internal format).
945 CCL program used for encoding (converting to external format).
947 (name, type, doc_string, props))
949 Lisp_Coding_System *codesys;
950 enum coding_system_type ty;
951 int need_to_setup_eol_systems = 1;
953 /* Convert type to constant */
954 if (NILP (type) || EQ (type, Qundecided))
955 { ty = CODESYS_AUTODETECT; }
957 else if (EQ (type, Qshift_jis)) { ty = CODESYS_SHIFT_JIS; }
958 else if (EQ (type, Qiso2022)) { ty = CODESYS_ISO2022; }
959 else if (EQ (type, Qbig5)) { ty = CODESYS_BIG5; }
960 else if (EQ (type, Qucs4)) { ty = CODESYS_UCS4; }
961 else if (EQ (type, Qutf16)) { ty = CODESYS_UTF16; }
962 else if (EQ (type, Qutf8)) { ty = CODESYS_UTF8; }
963 else if (EQ (type, Qccl)) { ty = CODESYS_CCL; }
965 else if (EQ (type, Qno_conversion)) { ty = CODESYS_NO_CONVERSION; }
967 else if (EQ (type, Qinternal)) { ty = CODESYS_INTERNAL; }
970 signal_simple_error ("Invalid coding system type", type);
974 codesys = allocate_coding_system (ty, name);
976 if (NILP (doc_string))
977 doc_string = build_string ("");
979 CHECK_STRING (doc_string);
980 CODING_SYSTEM_DOC_STRING (codesys) = doc_string;
983 EXTERNAL_PROPERTY_LIST_LOOP_3 (key, value, props)
985 if (EQ (key, Qmnemonic))
988 CHECK_STRING (value);
989 CODING_SYSTEM_MNEMONIC (codesys) = value;
992 else if (EQ (key, Qeol_type))
994 need_to_setup_eol_systems = NILP (value);
997 CODING_SYSTEM_EOL_TYPE (codesys) = symbol_to_eol_type (value);
1000 else if (EQ (key, Qpost_read_conversion))
1001 CODING_SYSTEM_POST_READ_CONVERSION (codesys) = value;
1002 else if (EQ (key, Qpre_write_conversion))
1003 CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = value;
1005 else if (EQ (key, Qdisable_composition))
1006 CODING_SYSTEM_DISABLE_COMPOSITION (codesys) = !NILP (value);
1007 else if (EQ (key, Quse_entity_reference))
1008 CODING_SYSTEM_USE_ENTITY_REFERENCE (codesys) = !NILP (value);
1011 else if (ty == CODESYS_ISO2022)
1013 #define FROB_INITIAL_CHARSET(charset_num) \
1014 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, charset_num) = \
1015 ((EQ (value, Qt) || EQ (value, Qnil)) ? value : Fget_charset (value))
1017 if (EQ (key, Qcharset_g0)) FROB_INITIAL_CHARSET (0);
1018 else if (EQ (key, Qcharset_g1)) FROB_INITIAL_CHARSET (1);
1019 else if (EQ (key, Qcharset_g2)) FROB_INITIAL_CHARSET (2);
1020 else if (EQ (key, Qcharset_g3)) FROB_INITIAL_CHARSET (3);
1022 #define FROB_FORCE_CHARSET(charset_num) \
1023 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (codesys, charset_num) = !NILP (value)
1025 else if (EQ (key, Qforce_g0_on_output)) FROB_FORCE_CHARSET (0);
1026 else if (EQ (key, Qforce_g1_on_output)) FROB_FORCE_CHARSET (1);
1027 else if (EQ (key, Qforce_g2_on_output)) FROB_FORCE_CHARSET (2);
1028 else if (EQ (key, Qforce_g3_on_output)) FROB_FORCE_CHARSET (3);
1030 #define FROB_BOOLEAN_PROPERTY(prop) \
1031 CODING_SYSTEM_ISO2022_##prop (codesys) = !NILP (value)
1033 else if (EQ (key, Qshort)) FROB_BOOLEAN_PROPERTY (SHORT);
1034 else if (EQ (key, Qno_ascii_eol)) FROB_BOOLEAN_PROPERTY (NO_ASCII_EOL);
1035 else if (EQ (key, Qno_ascii_cntl)) FROB_BOOLEAN_PROPERTY (NO_ASCII_CNTL);
1036 else if (EQ (key, Qseven)) FROB_BOOLEAN_PROPERTY (SEVEN);
1037 else if (EQ (key, Qlock_shift)) FROB_BOOLEAN_PROPERTY (LOCK_SHIFT);
1038 else if (EQ (key, Qno_iso6429)) FROB_BOOLEAN_PROPERTY (NO_ISO6429);
1039 else if (EQ (key, Qescape_quoted)) FROB_BOOLEAN_PROPERTY (ESCAPE_QUOTED);
1041 else if (EQ (key, Qinput_charset_conversion))
1043 codesys->iso2022.input_conv =
1044 Dynarr_new (charset_conversion_spec);
1045 parse_charset_conversion_specs (codesys->iso2022.input_conv,
1048 else if (EQ (key, Qoutput_charset_conversion))
1050 codesys->iso2022.output_conv =
1051 Dynarr_new (charset_conversion_spec);
1052 parse_charset_conversion_specs (codesys->iso2022.output_conv,
1056 signal_simple_error ("Unrecognized property", key);
1059 else if (ty == CODESYS_UTF8)
1061 if (EQ (key, Qcharset_g0)) FROB_INITIAL_CHARSET (0);
1062 else if (EQ (key, Qcharset_g1))
1063 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 1) = value;
1064 else if (EQ (key, Qcharset_g2))
1065 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 2) = value;
1067 signal_simple_error ("Unrecognized property", key);
1069 else if (ty == CODESYS_BIG5)
1071 if (EQ (key, Qcharset_g0)) FROB_INITIAL_CHARSET (0);
1072 else if (EQ (key, Qcharset_g1)) FROB_INITIAL_CHARSET (1);
1074 signal_simple_error ("Unrecognized property", key);
1077 else if (EQ (type, Qccl))
1080 struct ccl_program test_ccl;
1083 /* Check key first. */
1084 if (EQ (key, Qdecode))
1085 suffix = "-ccl-decode";
1086 else if (EQ (key, Qencode))
1087 suffix = "-ccl-encode";
1089 signal_simple_error ("Unrecognized property", key);
1091 /* If value is vector, register it as a ccl program
1092 associated with an newly created symbol for
1093 backward compatibility. */
1094 if (VECTORP (value))
1096 sym = Fintern (concat2 (Fsymbol_name (name),
1097 build_string (suffix)),
1099 Fregister_ccl_program (sym, value);
1103 CHECK_SYMBOL (value);
1106 /* check if the given ccl programs are valid. */
1107 if (setup_ccl_program (&test_ccl, sym) < 0)
1108 signal_simple_error ("Invalid CCL program", value);
1110 if (EQ (key, Qdecode))
1111 CODING_SYSTEM_CCL_DECODE (codesys) = sym;
1112 else if (EQ (key, Qencode))
1113 CODING_SYSTEM_CCL_ENCODE (codesys) = sym;
1118 signal_simple_error ("Unrecognized property", key);
1122 if (need_to_setup_eol_systems)
1123 setup_eol_coding_systems (codesys);
1126 Lisp_Object codesys_obj;
1127 XSETCODING_SYSTEM (codesys_obj, codesys);
1128 Fputhash (name, codesys_obj, Vcoding_system_hash_table);
1133 DEFUN ("copy-coding-system", Fcopy_coding_system, 2, 2, 0, /*
1134 Copy OLD-CODING-SYSTEM to NEW-NAME.
1135 If NEW-NAME does not name an existing coding system, a new one will
1138 (old_coding_system, new_name))
1140 Lisp_Object new_coding_system;
1141 old_coding_system = Fget_coding_system (old_coding_system);
1142 new_coding_system = Ffind_coding_system (new_name);
1143 if (NILP (new_coding_system))
1145 XSETCODING_SYSTEM (new_coding_system,
1146 allocate_coding_system
1147 (XCODING_SYSTEM_TYPE (old_coding_system),
1149 Fputhash (new_name, new_coding_system, Vcoding_system_hash_table);
1153 Lisp_Coding_System *to = XCODING_SYSTEM (new_coding_system);
1154 Lisp_Coding_System *from = XCODING_SYSTEM (old_coding_system);
1155 memcpy (((char *) to ) + sizeof (to->header),
1156 ((char *) from) + sizeof (from->header),
1157 sizeof (*from) - sizeof (from->header));
1158 to->name = new_name;
1160 return new_coding_system;
1163 DEFUN ("coding-system-canonical-name-p", Fcoding_system_canonical_name_p, 1, 1, 0, /*
1164 Return t if OBJECT names a coding system, and is not a coding system alias.
1168 return CODING_SYSTEMP (Fgethash (object, Vcoding_system_hash_table, Qnil))
1172 DEFUN ("coding-system-alias-p", Fcoding_system_alias_p, 1, 1, 0, /*
1173 Return t if OBJECT is a coding system alias.
1174 All coding system aliases are created by `define-coding-system-alias'.
1178 return SYMBOLP (Fgethash (object, Vcoding_system_hash_table, Qzero))
1182 DEFUN ("coding-system-aliasee", Fcoding_system_aliasee, 1, 1, 0, /*
1183 Return the coding-system symbol for which symbol ALIAS is an alias.
1187 Lisp_Object aliasee = Fgethash (alias, Vcoding_system_hash_table, Qnil);
1188 if (SYMBOLP (aliasee))
1191 signal_simple_error ("Symbol is not a coding system alias", alias);
1192 return Qnil; /* To keep the compiler happy */
1196 append_suffix_to_symbol (Lisp_Object symbol, const char *ascii_string)
1198 return Fintern (concat2 (Fsymbol_name (symbol), build_string (ascii_string)),
1202 /* A maphash function, for removing dangling coding system aliases. */
1204 dangling_coding_system_alias_p (Lisp_Object alias,
1205 Lisp_Object aliasee,
1206 void *dangling_aliases)
1208 if (SYMBOLP (aliasee)
1209 && NILP (Fgethash (aliasee, Vcoding_system_hash_table, Qnil)))
1211 (*(int *) dangling_aliases)++;
1218 DEFUN ("define-coding-system-alias", Fdefine_coding_system_alias, 2, 2, 0, /*
1219 Define symbol ALIAS as an alias for coding system ALIASEE.
1221 You can use this function to redefine an alias that has already been defined,
1222 but you cannot redefine a name which is the canonical name for a coding system.
1223 \(a canonical name of a coding system is what is returned when you call
1224 `coding-system-name' on a coding system).
1226 ALIASEE itself can be an alias, which allows you to define nested aliases.
1228 You are forbidden, however, from creating alias loops or `dangling' aliases.
1229 These will be detected, and an error will be signaled if you attempt to do so.
1231 If ALIASEE is nil, then ALIAS will simply be undefined.
1233 See also `coding-system-alias-p', `coding-system-aliasee',
1234 and `coding-system-canonical-name-p'.
1238 Lisp_Object real_coding_system, probe;
1240 CHECK_SYMBOL (alias);
1242 if (!NILP (Fcoding_system_canonical_name_p (alias)))
1244 ("Symbol is the canonical name of a coding system and cannot be redefined",
1249 Lisp_Object subsidiary_unix = append_suffix_to_symbol (alias, "-unix");
1250 Lisp_Object subsidiary_dos = append_suffix_to_symbol (alias, "-dos");
1251 Lisp_Object subsidiary_mac = append_suffix_to_symbol (alias, "-mac");
1253 Fremhash (alias, Vcoding_system_hash_table);
1255 /* Undefine subsidiary aliases,
1256 presumably created by a previous call to this function */
1257 if (! NILP (Fcoding_system_alias_p (subsidiary_unix)) &&
1258 ! NILP (Fcoding_system_alias_p (subsidiary_dos)) &&
1259 ! NILP (Fcoding_system_alias_p (subsidiary_mac)))
1261 Fdefine_coding_system_alias (subsidiary_unix, Qnil);
1262 Fdefine_coding_system_alias (subsidiary_dos, Qnil);
1263 Fdefine_coding_system_alias (subsidiary_mac, Qnil);
1266 /* Undefine dangling coding system aliases. */
1268 int dangling_aliases;
1271 dangling_aliases = 0;
1272 elisp_map_remhash (dangling_coding_system_alias_p,
1273 Vcoding_system_hash_table,
1275 } while (dangling_aliases > 0);
1281 if (CODING_SYSTEMP (aliasee))
1282 aliasee = XCODING_SYSTEM_NAME (aliasee);
1284 /* Checks that aliasee names a coding-system */
1285 real_coding_system = Fget_coding_system (aliasee);
1287 /* Check for coding system alias loops */
1288 if (EQ (alias, aliasee))
1289 alias_loop: signal_simple_error_2
1290 ("Attempt to create a coding system alias loop", alias, aliasee);
1292 for (probe = aliasee;
1294 probe = Fgethash (probe, Vcoding_system_hash_table, Qzero))
1296 if (EQ (probe, alias))
1300 Fputhash (alias, aliasee, Vcoding_system_hash_table);
1302 /* Set up aliases for subsidiaries.
1303 #### There must be a better way to handle subsidiary coding systems. */
1305 static const char *suffixes[] = { "-unix", "-dos", "-mac" };
1307 for (i = 0; i < countof (suffixes); i++)
1309 Lisp_Object alias_subsidiary =
1310 append_suffix_to_symbol (alias, suffixes[i]);
1311 Lisp_Object aliasee_subsidiary =
1312 append_suffix_to_symbol (aliasee, suffixes[i]);
1314 if (! NILP (Ffind_coding_system (aliasee_subsidiary)))
1315 Fdefine_coding_system_alias (alias_subsidiary, aliasee_subsidiary);
1318 /* FSF return value is a vector of [ALIAS-unix ALIAS-dos ALIAS-mac],
1319 but it doesn't look intentional, so I'd rather return something
1320 meaningful or nothing at all. */
1325 subsidiary_coding_system (Lisp_Object coding_system, eol_type_t type)
1327 Lisp_Coding_System *cs = XCODING_SYSTEM (coding_system);
1328 Lisp_Object new_coding_system;
1330 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
1331 return coding_system;
1335 case EOL_AUTODETECT: return coding_system;
1336 case EOL_LF: new_coding_system = CODING_SYSTEM_EOL_LF (cs); break;
1337 case EOL_CR: new_coding_system = CODING_SYSTEM_EOL_CR (cs); break;
1338 case EOL_CRLF: new_coding_system = CODING_SYSTEM_EOL_CRLF (cs); break;
1339 default: abort (); return Qnil;
1342 return NILP (new_coding_system) ? coding_system : new_coding_system;
1345 DEFUN ("subsidiary-coding-system", Fsubsidiary_coding_system, 2, 2, 0, /*
1346 Return the subsidiary coding system of CODING-SYSTEM with eol type EOL-TYPE.
1348 (coding_system, eol_type))
1350 coding_system = Fget_coding_system (coding_system);
1352 return subsidiary_coding_system (coding_system,
1353 symbol_to_eol_type (eol_type));
1357 /************************************************************************/
1358 /* Coding system accessors */
1359 /************************************************************************/
1361 DEFUN ("coding-system-doc-string", Fcoding_system_doc_string, 1, 1, 0, /*
1362 Return the doc string for CODING-SYSTEM.
1366 coding_system = Fget_coding_system (coding_system);
1367 return XCODING_SYSTEM_DOC_STRING (coding_system);
1370 DEFUN ("coding-system-type", Fcoding_system_type, 1, 1, 0, /*
1371 Return the type of CODING-SYSTEM.
1375 switch (XCODING_SYSTEM_TYPE (Fget_coding_system (coding_system)))
1378 case CODESYS_AUTODETECT: return Qundecided;
1380 case CODESYS_SHIFT_JIS: return Qshift_jis;
1381 case CODESYS_ISO2022: return Qiso2022;
1382 case CODESYS_BIG5: return Qbig5;
1383 case CODESYS_UCS4: return Qucs4;
1384 case CODESYS_UTF16: return Qutf16;
1385 case CODESYS_UTF8: return Qutf8;
1386 case CODESYS_CCL: return Qccl;
1388 case CODESYS_NO_CONVERSION: return Qno_conversion;
1390 case CODESYS_INTERNAL: return Qinternal;
1397 Lisp_Object coding_system_charset (Lisp_Object coding_system, int gnum)
1400 = XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, gnum);
1402 return CHARSETP (cs) ? XCHARSET_NAME (cs) : Qnil;
1405 DEFUN ("coding-system-charset", Fcoding_system_charset, 2, 2, 0, /*
1406 Return initial charset of CODING-SYSTEM designated to GNUM.
1409 (coding_system, gnum))
1411 coding_system = Fget_coding_system (coding_system);
1414 return coding_system_charset (coding_system, XINT (gnum));
1418 DEFUN ("coding-system-property", Fcoding_system_property, 2, 2, 0, /*
1419 Return the PROP property of CODING-SYSTEM.
1421 (coding_system, prop))
1424 enum coding_system_type type;
1426 coding_system = Fget_coding_system (coding_system);
1427 CHECK_SYMBOL (prop);
1428 type = XCODING_SYSTEM_TYPE (coding_system);
1430 for (i = 0; !ok && i < Dynarr_length (the_codesys_prop_dynarr); i++)
1431 if (EQ (Dynarr_at (the_codesys_prop_dynarr, i).sym, prop))
1434 switch (Dynarr_at (the_codesys_prop_dynarr, i).prop_type)
1436 case CODESYS_PROP_ALL_OK:
1439 case CODESYS_PROP_ISO2022:
1440 if (type != CODESYS_ISO2022)
1442 ("Property only valid in ISO2022 coding systems",
1446 case CODESYS_PROP_CCL:
1447 if (type != CODESYS_CCL)
1449 ("Property only valid in CCL coding systems",
1459 signal_simple_error ("Unrecognized property", prop);
1461 if (EQ (prop, Qname))
1462 return XCODING_SYSTEM_NAME (coding_system);
1463 else if (EQ (prop, Qtype))
1464 return Fcoding_system_type (coding_system);
1465 else if (EQ (prop, Qdoc_string))
1466 return XCODING_SYSTEM_DOC_STRING (coding_system);
1467 else if (EQ (prop, Qmnemonic))
1468 return XCODING_SYSTEM_MNEMONIC (coding_system);
1469 else if (EQ (prop, Qeol_type))
1470 return eol_type_to_symbol (XCODING_SYSTEM_EOL_TYPE (coding_system));
1471 else if (EQ (prop, Qeol_lf))
1472 return XCODING_SYSTEM_EOL_LF (coding_system);
1473 else if (EQ (prop, Qeol_crlf))
1474 return XCODING_SYSTEM_EOL_CRLF (coding_system);
1475 else if (EQ (prop, Qeol_cr))
1476 return XCODING_SYSTEM_EOL_CR (coding_system);
1477 else if (EQ (prop, Qpost_read_conversion))
1478 return XCODING_SYSTEM_POST_READ_CONVERSION (coding_system);
1479 else if (EQ (prop, Qpre_write_conversion))
1480 return XCODING_SYSTEM_PRE_WRITE_CONVERSION (coding_system);
1483 else if (EQ (prop, Qdisable_composition))
1484 return XCODING_SYSTEM_DISABLE_COMPOSITION (coding_system) ? Qt : Qnil;
1485 else if (EQ (prop, Quse_entity_reference))
1486 return XCODING_SYSTEM_USE_ENTITY_REFERENCE (coding_system) ? Qt : Qnil;
1488 else if (type == CODESYS_ISO2022)
1490 if (EQ (prop, Qcharset_g0))
1491 return coding_system_charset (coding_system, 0);
1492 else if (EQ (prop, Qcharset_g1))
1493 return coding_system_charset (coding_system, 1);
1494 else if (EQ (prop, Qcharset_g2))
1495 return coding_system_charset (coding_system, 2);
1496 else if (EQ (prop, Qcharset_g3))
1497 return coding_system_charset (coding_system, 3);
1499 #define FORCE_CHARSET(charset_num) \
1500 (XCODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT \
1501 (coding_system, charset_num) ? Qt : Qnil)
1503 else if (EQ (prop, Qforce_g0_on_output)) return FORCE_CHARSET (0);
1504 else if (EQ (prop, Qforce_g1_on_output)) return FORCE_CHARSET (1);
1505 else if (EQ (prop, Qforce_g2_on_output)) return FORCE_CHARSET (2);
1506 else if (EQ (prop, Qforce_g3_on_output)) return FORCE_CHARSET (3);
1508 #define LISP_BOOLEAN(prop) \
1509 (XCODING_SYSTEM_ISO2022_##prop (coding_system) ? Qt : Qnil)
1511 else if (EQ (prop, Qshort)) return LISP_BOOLEAN (SHORT);
1512 else if (EQ (prop, Qno_ascii_eol)) return LISP_BOOLEAN (NO_ASCII_EOL);
1513 else if (EQ (prop, Qno_ascii_cntl)) return LISP_BOOLEAN (NO_ASCII_CNTL);
1514 else if (EQ (prop, Qseven)) return LISP_BOOLEAN (SEVEN);
1515 else if (EQ (prop, Qlock_shift)) return LISP_BOOLEAN (LOCK_SHIFT);
1516 else if (EQ (prop, Qno_iso6429)) return LISP_BOOLEAN (NO_ISO6429);
1517 else if (EQ (prop, Qescape_quoted)) return LISP_BOOLEAN (ESCAPE_QUOTED);
1519 else if (EQ (prop, Qinput_charset_conversion))
1521 unparse_charset_conversion_specs
1522 (XCODING_SYSTEM (coding_system)->iso2022.input_conv);
1523 else if (EQ (prop, Qoutput_charset_conversion))
1525 unparse_charset_conversion_specs
1526 (XCODING_SYSTEM (coding_system)->iso2022.output_conv);
1530 else if (type == CODESYS_CCL)
1532 if (EQ (prop, Qdecode))
1533 return XCODING_SYSTEM_CCL_DECODE (coding_system);
1534 else if (EQ (prop, Qencode))
1535 return XCODING_SYSTEM_CCL_ENCODE (coding_system);
1543 return Qnil; /* not reached */
1547 /************************************************************************/
1548 /* Coding category functions */
1549 /************************************************************************/
1552 decode_coding_category (Lisp_Object symbol)
1556 CHECK_SYMBOL (symbol);
1557 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1558 if (EQ (coding_category_symbol[i], symbol))
1561 signal_simple_error ("Unrecognized coding category", symbol);
1562 return 0; /* not reached */
1565 DEFUN ("coding-category-list", Fcoding_category_list, 0, 0, 0, /*
1566 Return a list of all recognized coding categories.
1571 Lisp_Object list = Qnil;
1573 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
1574 list = Fcons (coding_category_symbol[i], list);
1578 DEFUN ("set-coding-priority-list", Fset_coding_priority_list, 1, 1, 0, /*
1579 Change the priority order of the coding categories.
1580 LIST should be list of coding categories, in descending order of
1581 priority. Unspecified coding categories will be lower in priority
1582 than all specified ones, in the same relative order they were in
1587 int category_to_priority[CODING_CATEGORY_LAST];
1591 /* First generate a list that maps coding categories to priorities. */
1593 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1594 category_to_priority[i] = -1;
1596 /* Highest priority comes from the specified list. */
1598 EXTERNAL_LIST_LOOP (rest, list)
1600 int cat = decode_coding_category (XCAR (rest));
1602 if (category_to_priority[cat] >= 0)
1603 signal_simple_error ("Duplicate coding category in list", XCAR (rest));
1604 category_to_priority[cat] = i++;
1607 /* Now go through the existing categories by priority to retrieve
1608 the categories not yet specified and preserve their priority
1610 for (j = 0; j < CODING_CATEGORY_LAST; j++)
1612 int cat = fcd->coding_category_by_priority[j];
1613 if (category_to_priority[cat] < 0)
1614 category_to_priority[cat] = i++;
1617 /* Now we need to construct the inverse of the mapping we just
1620 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1621 fcd->coding_category_by_priority[category_to_priority[i]] = i;
1623 /* Phew! That was confusing. */
1627 DEFUN ("coding-priority-list", Fcoding_priority_list, 0, 0, 0, /*
1628 Return a list of coding categories in descending order of priority.
1633 Lisp_Object list = Qnil;
1635 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
1636 list = Fcons (coding_category_symbol[fcd->coding_category_by_priority[i]],
1641 DEFUN ("set-coding-category-system", Fset_coding_category_system, 2, 2, 0, /*
1642 Change the coding system associated with a coding category.
1644 (coding_category, coding_system))
1646 int cat = decode_coding_category (coding_category);
1648 coding_system = Fget_coding_system (coding_system);
1649 fcd->coding_category_system[cat] = coding_system;
1653 DEFUN ("coding-category-system", Fcoding_category_system, 1, 1, 0, /*
1654 Return the coding system associated with a coding category.
1658 int cat = decode_coding_category (coding_category);
1659 Lisp_Object sys = fcd->coding_category_system[cat];
1662 return XCODING_SYSTEM_NAME (sys);
1667 /************************************************************************/
1668 /* Detecting the encoding of data */
1669 /************************************************************************/
1671 struct detection_state
1673 eol_type_t eol_type;
1716 struct iso2022_decoder iso;
1718 int high_byte_count;
1719 unsigned int saw_single_shift:1;
1732 acceptable_control_char_p (int c)
1736 /* Allow and ignore control characters that you might
1737 reasonably see in a text file */
1742 case 8: /* backspace */
1743 case 11: /* vertical tab */
1744 case 12: /* form feed */
1745 case 26: /* MS-DOS C-z junk */
1746 case 31: /* '^_' -- for info */
1754 mask_has_at_most_one_bit_p (int mask)
1756 /* Perhaps the only thing useful you learn from intensive Microsoft
1757 technical interviews */
1758 return (mask & (mask - 1)) == 0;
1762 detect_eol_type (struct detection_state *st, const Extbyte *src,
1763 Lstream_data_count n)
1767 unsigned char c = *(unsigned char *)src++;
1770 if (st->eol.just_saw_cr)
1772 else if (st->eol.seen_anything)
1775 else if (st->eol.just_saw_cr)
1778 st->eol.just_saw_cr = 1;
1780 st->eol.just_saw_cr = 0;
1781 st->eol.seen_anything = 1;
1784 return EOL_AUTODETECT;
1787 /* Attempt to determine the encoding and EOL type of the given text.
1788 Before calling this function for the first type, you must initialize
1789 st->eol_type as appropriate and initialize st->mask to ~0.
1791 st->eol_type holds the determined EOL type, or EOL_AUTODETECT if
1794 st->mask holds the determined coding category mask, or ~0 if only
1795 ASCII has been seen so far.
1799 0 == st->eol_type is EOL_AUTODETECT and/or more than coding category
1800 is present in st->mask
1801 1 == definitive answers are here for both st->eol_type and st->mask
1805 detect_coding_type (struct detection_state *st, const Extbyte *src,
1806 Lstream_data_count n, int just_do_eol)
1808 if (st->eol_type == EOL_AUTODETECT)
1809 st->eol_type = detect_eol_type (st, src, n);
1812 return st->eol_type != EOL_AUTODETECT;
1814 if (!st->seen_non_ascii)
1816 for (; n; n--, src++)
1818 unsigned char c = *(unsigned char *) src;
1819 if ((c < 0x20 && !acceptable_control_char_p (c)) || c >= 0x80)
1821 st->seen_non_ascii = 1;
1823 st->shift_jis.mask = ~0;
1826 st->utf16.mask = ~0;
1828 st->iso2022.mask = ~0;
1838 if (!mask_has_at_most_one_bit_p (st->iso2022.mask))
1839 st->iso2022.mask = detect_coding_iso2022 (st, src, n);
1840 if (!mask_has_at_most_one_bit_p (st->shift_jis.mask))
1841 st->shift_jis.mask = detect_coding_sjis (st, src, n);
1842 if (!mask_has_at_most_one_bit_p (st->big5.mask))
1843 st->big5.mask = detect_coding_big5 (st, src, n);
1844 if (!mask_has_at_most_one_bit_p (st->utf8.mask))
1845 st->utf8.mask = detect_coding_utf8 (st, src, n);
1846 if (!mask_has_at_most_one_bit_p (st->utf16.mask))
1847 st->utf16.mask = detect_coding_utf16 (st, src, n);
1848 if (!mask_has_at_most_one_bit_p (st->ucs4.mask))
1849 st->ucs4.mask = detect_coding_ucs4 (st, src, n);
1852 = st->iso2022.mask | st->shift_jis.mask | st->big5.mask
1853 | st->utf8.mask | st->ucs4.mask;
1856 int retval = mask_has_at_most_one_bit_p (st->mask);
1857 st->mask |= CODING_CATEGORY_NO_CONVERSION_MASK;
1858 return retval && st->eol_type != EOL_AUTODETECT;
1863 coding_system_from_mask (int mask)
1867 /* If the file was entirely or basically ASCII, use the
1868 default value of `buffer-file-coding-system'. */
1869 Lisp_Object retval =
1870 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system;
1873 retval = Ffind_coding_system (retval);
1877 (Qbad_variable, Qwarning,
1878 "Invalid `default-buffer-file-coding-system', set to nil");
1879 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system = Qnil;
1883 retval = Fget_coding_system (Qraw_text);
1891 mask = postprocess_iso2022_mask (mask);
1893 /* Look through the coding categories by priority and find
1894 the first one that is allowed. */
1895 for (i = 0; i < CODING_CATEGORY_LAST; i++)
1897 cat = fcd->coding_category_by_priority[i];
1898 if ((mask & (1 << cat)) &&
1899 !NILP (fcd->coding_category_system[cat]))
1903 return fcd->coding_category_system[cat];
1905 return Fget_coding_system (Qraw_text);
1909 /* Given a seekable read stream and potential coding system and EOL type
1910 as specified, do any autodetection that is called for. If the
1911 coding system and/or EOL type are not `autodetect', they will be left
1912 alone; but this function will never return an autodetect coding system
1915 This function does not automatically fetch subsidiary coding systems;
1916 that should be unnecessary with the explicit eol-type argument. */
1918 #define LENGTH(string_constant) (sizeof (string_constant) - 1)
1919 /* number of leading lines to check for a coding cookie */
1920 #define LINES_TO_CHECK 2
1923 determine_real_coding_system (Lstream *stream, Lisp_Object *codesys_in_out,
1924 eol_type_t *eol_type_in_out)
1926 struct detection_state decst;
1928 if (*eol_type_in_out == EOL_AUTODETECT)
1929 *eol_type_in_out = XCODING_SYSTEM_EOL_TYPE (*codesys_in_out);
1932 decst.eol_type = *eol_type_in_out;
1935 /* If autodetection is called for, do it now. */
1936 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT
1937 || *eol_type_in_out == EOL_AUTODETECT)
1940 Lisp_Object coding_system = Qnil;
1942 Lstream_data_count nread = Lstream_read (stream, buf, sizeof (buf));
1944 int lines_checked = 0;
1946 /* Look for initial "-*-"; mode line prefix */
1948 scan_end = buf + nread - LENGTH ("-*-coding:?-*-");
1950 && lines_checked < LINES_TO_CHECK;
1952 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-')
1954 Extbyte *local_vars_beg = p + 3;
1955 /* Look for final "-*-"; mode line suffix */
1956 for (p = local_vars_beg,
1957 scan_end = buf + nread - LENGTH ("-*-");
1959 && lines_checked < LINES_TO_CHECK;
1961 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-')
1963 Extbyte *suffix = p;
1964 /* Look for "coding:" */
1965 for (p = local_vars_beg,
1966 scan_end = suffix - LENGTH ("coding:?");
1969 if (memcmp ("coding:", p, LENGTH ("coding:")) == 0
1970 && (p == local_vars_beg
1971 || (*(p-1) == ' ' ||
1977 p += LENGTH ("coding:");
1978 while (*p == ' ' || *p == '\t') p++;
1980 /* Get coding system name */
1981 save = *suffix; *suffix = '\0';
1982 /* Characters valid in a MIME charset name (rfc 1521),
1983 and in a Lisp symbol name. */
1984 n = strspn ( (char *) p,
1985 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1986 "abcdefghijklmnopqrstuvwxyz"
1992 save = p[n]; p[n] = '\0';
1994 Ffind_coding_system (intern ((char *) p));
2001 /* #### file must use standard EOLs or we miss 2d line */
2002 /* #### not to mention this is broken for UTF-16 DOS files */
2003 else if (*p == '\n' || *p == '\r')
2006 /* skip past multibyte (DOS) newline */
2007 if (*p == '\r' && *(p+1) == '\n') p++;
2011 /* #### file must use standard EOLs or we miss 2d line */
2012 /* #### not to mention this is broken for UTF-16 DOS files */
2013 else if (*p == '\n' || *p == '\r')
2016 /* skip past multibyte (DOS) newline */
2017 if (*p == '\r' && *(p+1) == '\n') p++;
2020 if (NILP (coding_system))
2023 if (detect_coding_type (&decst, buf, nread,
2024 XCODING_SYSTEM_TYPE (*codesys_in_out)
2025 != CODESYS_AUTODETECT))
2027 nread = Lstream_read (stream, buf, sizeof (buf));
2033 else if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT
2034 && XCODING_SYSTEM_EOL_TYPE (coding_system) == EOL_AUTODETECT)
2037 if (detect_coding_type (&decst, buf, nread, 1))
2039 nread = Lstream_read (stream, buf, sizeof (buf));
2045 *eol_type_in_out = decst.eol_type;
2046 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT)
2048 if (NILP (coding_system))
2049 *codesys_in_out = coding_system_from_mask (decst.mask);
2051 *codesys_in_out = coding_system;
2055 /* If we absolutely can't determine the EOL type, just assume LF. */
2056 if (*eol_type_in_out == EOL_AUTODETECT)
2057 *eol_type_in_out = EOL_LF;
2059 Lstream_rewind (stream);
2062 DEFUN ("detect-coding-region", Fdetect_coding_region, 2, 3, 0, /*
2063 Detect coding system of the text in the region between START and END.
2064 Return a list of possible coding systems ordered by priority.
2065 If only ASCII characters are found, return 'undecided or one of
2066 its subsidiary coding systems according to a detected end-of-line
2067 type. Optional arg BUFFER defaults to the current buffer.
2069 (start, end, buffer))
2071 Lisp_Object val = Qnil;
2072 struct buffer *buf = decode_buffer (buffer, 0);
2074 Lisp_Object instream, lb_instream;
2075 Lstream *istr, *lb_istr;
2076 struct detection_state decst;
2077 struct gcpro gcpro1, gcpro2;
2079 get_buffer_range_char (buf, start, end, &b, &e, 0);
2080 lb_instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2081 lb_istr = XLSTREAM (lb_instream);
2082 instream = make_encoding_input_stream (lb_istr, Fget_coding_system (Qbinary));
2083 istr = XLSTREAM (instream);
2084 GCPRO2 (instream, lb_instream);
2086 decst.eol_type = EOL_AUTODETECT;
2090 Extbyte random_buffer[4096];
2091 Lstream_data_count nread = Lstream_read (istr, random_buffer, sizeof (random_buffer));
2095 if (detect_coding_type (&decst, random_buffer, nread, 0))
2099 if (decst.mask == ~0)
2100 val = subsidiary_coding_system (Fget_coding_system (Qundecided),
2108 decst.mask = postprocess_iso2022_mask (decst.mask);
2110 for (i = CODING_CATEGORY_LAST - 1; i >= 0; i--)
2112 int sys = fcd->coding_category_by_priority[i];
2113 if (decst.mask & (1 << sys))
2115 Lisp_Object codesys = fcd->coding_category_system[sys];
2116 if (!NILP (codesys))
2117 codesys = subsidiary_coding_system (codesys, decst.eol_type);
2118 val = Fcons (codesys, val);
2122 Lstream_close (istr);
2124 Lstream_delete (istr);
2125 Lstream_delete (lb_istr);
2130 /************************************************************************/
2131 /* Converting to internal Mule format ("decoding") */
2132 /************************************************************************/
2134 /* A decoding stream is a stream used for decoding text (i.e.
2135 converting from some external format to internal format).
2136 The decoding-stream object keeps track of the actual coding
2137 stream, the stream that is at the other end, and data that
2138 needs to be persistent across the lifetime of the stream. */
2140 /* Handle the EOL stuff related to just-read-in character C.
2141 EOL_TYPE is the EOL type of the coding stream.
2142 FLAGS is the current value of FLAGS in the coding stream, and may
2143 be modified by this macro. (The macro only looks at the
2144 CODING_STATE_CR flag.) DST is the Dynarr to which the decoded
2145 bytes are to be written. You need to also define a local goto
2146 label "label_continue_loop" that is at the end of the main
2147 character-reading loop.
2149 If C is a CR character, then this macro handles it entirely and
2150 jumps to label_continue_loop. Otherwise, this macro does not add
2151 anything to DST, and continues normally. You should continue
2152 processing C normally after this macro. */
2154 #define DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst) \
2158 if (eol_type == EOL_CR) \
2159 Dynarr_add (dst, '\n'); \
2160 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR) \
2161 Dynarr_add (dst, c); \
2163 flags |= CODING_STATE_CR; \
2164 goto label_continue_loop; \
2166 else if (flags & CODING_STATE_CR) \
2167 { /* eol_type == CODING_SYSTEM_EOL_CRLF */ \
2169 Dynarr_add (dst, '\r'); \
2170 flags &= ~CODING_STATE_CR; \
2174 /* C should be a binary character in the range 0 - 255; convert
2175 to internal format and add to Dynarr DST. */
2178 #define DECODE_ADD_BINARY_CHAR(c, dst) \
2180 if (BYTE_ASCII_P (c)) \
2181 Dynarr_add (dst, c); \
2184 Dynarr_add (dst, (c >> 6) | 0xc0); \
2185 Dynarr_add (dst, (c & 0x3f) | 0x80); \
2189 INLINE_HEADER void DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst);
2191 DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst)
2195 Dynarr_add (dst, c);
2197 else if ( c <= 0x7ff )
2199 Dynarr_add (dst, (c >> 6) | 0xc0);
2200 Dynarr_add (dst, (c & 0x3f) | 0x80);
2202 else if ( c <= 0xffff )
2204 Dynarr_add (dst, (c >> 12) | 0xe0);
2205 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2206 Dynarr_add (dst, (c & 0x3f) | 0x80);
2208 else if ( c <= 0x1fffff )
2210 Dynarr_add (dst, (c >> 18) | 0xf0);
2211 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
2212 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2213 Dynarr_add (dst, (c & 0x3f) | 0x80);
2215 else if ( c <= 0x3ffffff )
2217 Dynarr_add (dst, (c >> 24) | 0xf8);
2218 Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
2219 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
2220 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2221 Dynarr_add (dst, (c & 0x3f) | 0x80);
2225 Dynarr_add (dst, (c >> 30) | 0xfc);
2226 Dynarr_add (dst, ((c >> 24) & 0x3f) | 0x80);
2227 Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
2228 Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
2229 Dynarr_add (dst, ((c >> 6) & 0x3f) | 0x80);
2230 Dynarr_add (dst, (c & 0x3f) | 0x80);
2234 #define DECODE_ADD_BINARY_CHAR(c, dst) \
2236 if (BYTE_ASCII_P (c)) \
2237 Dynarr_add (dst, c); \
2238 else if (BYTE_C1_P (c)) \
2240 Dynarr_add (dst, LEADING_BYTE_CONTROL_1); \
2241 Dynarr_add (dst, c + 0x20); \
2245 Dynarr_add (dst, LEADING_BYTE_LATIN_ISO8859_1); \
2246 Dynarr_add (dst, c); \
2251 #define DECODE_OUTPUT_PARTIAL_CHAR(ch) \
2255 DECODE_ADD_BINARY_CHAR (ch, dst); \
2260 #define DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst) \
2262 if (flags & CODING_STATE_END) \
2264 DECODE_OUTPUT_PARTIAL_CHAR (ch); \
2265 if (flags & CODING_STATE_CR) \
2266 Dynarr_add (dst, '\r'); \
2270 #define DECODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, decoding)
2272 struct decoding_stream
2274 /* Coding system that governs the conversion. */
2275 Lisp_Coding_System *codesys;
2277 /* Stream that we read the encoded data from or
2278 write the decoded data to. */
2281 /* If we are reading, then we can return only a fixed amount of
2282 data, so if the conversion resulted in too much data, we store it
2283 here for retrieval the next time around. */
2284 unsigned_char_dynarr *runoff;
2286 /* FLAGS holds flags indicating the current state of the decoding.
2287 Some of these flags are dependent on the coding system. */
2290 /* CPOS holds a partially built-up code-point of character. */
2293 /* EOL_TYPE specifies the type of end-of-line conversion that
2294 currently applies. We need to keep this separate from the
2295 EOL type stored in CODESYS because the latter might indicate
2296 automatic EOL-type detection while the former will always
2297 indicate a particular EOL type. */
2298 eol_type_t eol_type;
2300 /* Additional ISO2022 information. We define the structure above
2301 because it's also needed by the detection routines. */
2302 struct iso2022_decoder iso2022;
2304 /* Additional information (the state of the running CCL program)
2305 used by the CCL decoder. */
2306 struct ccl_program ccl;
2308 /* counter for UTF-8 or UCS-4 */
2309 unsigned char counter;
2312 unsigned char er_counter;
2313 unsigned char er_buf[16];
2315 unsigned combined_char_count;
2316 Emchar combined_chars[16];
2317 Lisp_Object combining_table;
2319 struct detection_state decst;
2322 static Lstream_data_count decoding_reader (Lstream *stream,
2323 unsigned char *data, Lstream_data_count size);
2324 static Lstream_data_count decoding_writer (Lstream *stream,
2325 const unsigned char *data, Lstream_data_count size);
2326 static int decoding_rewinder (Lstream *stream);
2327 static int decoding_seekable_p (Lstream *stream);
2328 static int decoding_flusher (Lstream *stream);
2329 static int decoding_closer (Lstream *stream);
2331 static Lisp_Object decoding_marker (Lisp_Object stream);
2333 DEFINE_LSTREAM_IMPLEMENTATION ("decoding", lstream_decoding,
2334 sizeof (struct decoding_stream));
2337 decoding_marker (Lisp_Object stream)
2339 Lstream *str = DECODING_STREAM_DATA (XLSTREAM (stream))->other_end;
2340 Lisp_Object str_obj;
2342 /* We do not need to mark the coding systems or charsets stored
2343 within the stream because they are stored in a global list
2344 and automatically marked. */
2346 XSETLSTREAM (str_obj, str);
2347 mark_object (str_obj);
2348 if (str->imp->marker)
2349 return (str->imp->marker) (str_obj);
2354 /* Read SIZE bytes of data and store it into DATA. We are a decoding stream
2355 so we read data from the other end, decode it, and store it into DATA. */
2357 static Lstream_data_count
2358 decoding_reader (Lstream *stream, unsigned char *data, Lstream_data_count size)
2360 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2361 unsigned char *orig_data = data;
2362 Lstream_data_count read_size;
2363 int error_occurred = 0;
2365 /* We need to interface to mule_decode(), which expects to take some
2366 amount of data and store the result into a Dynarr. We have
2367 mule_decode() store into str->runoff, and take data from there
2370 /* We loop until we have enough data, reading chunks from the other
2371 end and decoding it. */
2374 /* Take data from the runoff if we can. Make sure to take at
2375 most SIZE bytes, and delete the data from the runoff. */
2376 if (Dynarr_length (str->runoff) > 0)
2378 Lstream_data_count chunk = min (size, (Lstream_data_count) Dynarr_length (str->runoff));
2379 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2380 Dynarr_delete_many (str->runoff, 0, chunk);
2386 break; /* No more room for data */
2388 if (str->flags & CODING_STATE_END)
2389 /* This means that on the previous iteration, we hit the EOF on
2390 the other end. We loop once more so that mule_decode() can
2391 output any final stuff it may be holding, or any "go back
2392 to a sane state" escape sequences. (This latter makes sense
2393 during encoding.) */
2396 /* Exhausted the runoff, so get some more. DATA has at least
2397 SIZE bytes left of storage in it, so it's OK to read directly
2398 into it. (We'll be overwriting above, after we've decoded it
2399 into the runoff.) */
2400 read_size = Lstream_read (str->other_end, data, size);
2407 /* There might be some more end data produced in the translation.
2408 See the comment above. */
2409 str->flags |= CODING_STATE_END;
2410 mule_decode (stream, (Extbyte *) data, str->runoff, read_size);
2413 if (data - orig_data == 0)
2414 return error_occurred ? -1 : 0;
2416 return data - orig_data;
2419 static Lstream_data_count
2420 decoding_writer (Lstream *stream, const unsigned char *data, Lstream_data_count size)
2422 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2423 Lstream_data_count retval;
2425 /* Decode all our data into the runoff, and then attempt to write
2426 it all out to the other end. Remove whatever chunk we succeeded
2428 mule_decode (stream, (Extbyte *) data, str->runoff, size);
2429 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
2430 Dynarr_length (str->runoff));
2432 Dynarr_delete_many (str->runoff, 0, retval);
2433 /* Do NOT return retval. The return value indicates how much
2434 of the incoming data was written, not how many bytes were
2440 reset_decoding_stream (struct decoding_stream *str)
2443 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_ISO2022)
2445 Lisp_Object coding_system;
2446 XSETCODING_SYSTEM (coding_system, str->codesys);
2447 reset_iso2022 (coding_system, &str->iso2022);
2449 else if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_CCL)
2451 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_DECODE (str->codesys));
2456 str->er_counter = 0;
2457 str->combined_char_count = 0;
2458 str->combining_table = Qnil;
2460 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT
2461 || CODING_SYSTEM_EOL_TYPE (str->codesys) == EOL_AUTODETECT)
2464 str->decst.eol_type = EOL_AUTODETECT;
2465 str->decst.mask = ~0;
2467 str->flags = str->cpos = 0;
2471 decoding_rewinder (Lstream *stream)
2473 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2474 reset_decoding_stream (str);
2475 Dynarr_reset (str->runoff);
2476 return Lstream_rewind (str->other_end);
2480 decoding_seekable_p (Lstream *stream)
2482 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2483 return Lstream_seekable_p (str->other_end);
2487 decoding_flusher (Lstream *stream)
2489 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2490 return Lstream_flush (str->other_end);
2494 decoding_closer (Lstream *stream)
2496 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2497 if (stream->flags & LSTREAM_FL_WRITE)
2499 str->flags |= CODING_STATE_END;
2500 decoding_writer (stream, 0, 0);
2502 Dynarr_free (str->runoff);
2504 #ifdef ENABLE_COMPOSITE_CHARS
2505 if (str->iso2022.composite_chars)
2506 Dynarr_free (str->iso2022.composite_chars);
2509 return Lstream_close (str->other_end);
2513 decoding_stream_coding_system (Lstream *stream)
2515 Lisp_Object coding_system;
2516 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2518 XSETCODING_SYSTEM (coding_system, str->codesys);
2519 return subsidiary_coding_system (coding_system, str->eol_type);
2523 set_decoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
2525 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
2526 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2528 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
2529 str->eol_type = CODING_SYSTEM_EOL_TYPE (cs);
2530 reset_decoding_stream (str);
2533 /* WARNING WARNING WARNING WARNING!!!!! If you open up a decoding
2534 stream for writing, no automatic code detection will be performed.
2535 The reason for this is that automatic code detection requires a
2536 seekable input. Things will also fail if you open a decoding
2537 stream for reading using a non-fully-specified coding system and
2538 a non-seekable input stream. */
2541 make_decoding_stream_1 (Lstream *stream, Lisp_Object codesys,
2544 Lstream *lstr = Lstream_new (lstream_decoding, mode);
2545 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2549 str->other_end = stream;
2550 str->runoff = (unsigned_char_dynarr *) Dynarr_new (unsigned_char);
2551 str->eol_type = EOL_AUTODETECT;
2552 if (!strcmp (mode, "r")
2553 && Lstream_seekable_p (stream))
2554 /* We can determine the coding system now. */
2555 determine_real_coding_system (stream, &codesys, &str->eol_type);
2556 set_decoding_stream_coding_system (lstr, codesys);
2557 str->decst.eol_type = str->eol_type;
2558 str->decst.mask = ~0;
2559 XSETLSTREAM (obj, lstr);
2564 make_decoding_input_stream (Lstream *stream, Lisp_Object codesys)
2566 return make_decoding_stream_1 (stream, codesys, "r");
2570 make_decoding_output_stream (Lstream *stream, Lisp_Object codesys)
2572 return make_decoding_stream_1 (stream, codesys, "w");
2575 /* Note: the decode_coding_* functions all take the same
2576 arguments as mule_decode(), which is to say some SRC data of
2577 size N, which is to be stored into dynamic array DST.
2578 DECODING is the stream within which the decoding is
2579 taking place, but no data is actually read from or
2580 written to that stream; that is handled in decoding_reader()
2581 or decoding_writer(). This allows the same functions to
2582 be used for both reading and writing. */
2585 mule_decode (Lstream *decoding, const Extbyte *src,
2586 unsigned_char_dynarr *dst, Lstream_data_count n)
2588 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
2590 /* If necessary, do encoding-detection now. We do this when
2591 we're a writing stream or a non-seekable reading stream,
2592 meaning that we can't just process the whole input,
2593 rewind, and start over. */
2595 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT ||
2596 str->eol_type == EOL_AUTODETECT)
2598 Lisp_Object codesys;
2600 XSETCODING_SYSTEM (codesys, str->codesys);
2601 detect_coding_type (&str->decst, src, n,
2602 CODING_SYSTEM_TYPE (str->codesys) !=
2603 CODESYS_AUTODETECT);
2604 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT &&
2605 str->decst.mask != ~0)
2606 /* #### This is cheesy. What we really ought to do is
2607 buffer up a certain amount of data so as to get a
2608 less random result. */
2609 codesys = coding_system_from_mask (str->decst.mask);
2610 str->eol_type = str->decst.eol_type;
2611 if (XCODING_SYSTEM (codesys) != str->codesys)
2613 /* Preserve the CODING_STATE_END flag in case it was set.
2614 If we erase it, bad things might happen. */
2615 int was_end = str->flags & CODING_STATE_END;
2616 set_decoding_stream_coding_system (decoding, codesys);
2618 str->flags |= CODING_STATE_END;
2622 switch (CODING_SYSTEM_TYPE (str->codesys))
2625 case CODESYS_INTERNAL:
2626 Dynarr_add_many (dst, src, n);
2629 case CODESYS_AUTODETECT:
2630 /* If we got this far and still haven't decided on the coding
2631 system, then do no conversion. */
2632 case CODESYS_NO_CONVERSION:
2633 decode_coding_no_conversion (decoding, src, dst, n);
2636 case CODESYS_SHIFT_JIS:
2637 decode_coding_sjis (decoding, src, dst, n);
2640 decode_coding_big5 (decoding, src, dst, n);
2643 decode_coding_ucs4 (decoding, src, dst, n);
2646 decode_coding_utf16 (decoding, src, dst, n);
2649 decode_coding_utf8 (decoding, src, dst, n);
2652 str->ccl.last_block = str->flags & CODING_STATE_END;
2653 /* When applying ccl program to stream, MUST NOT set NULL
2655 ccl_driver (&str->ccl, (src ? (unsigned char *)src : (unsigned char*)""),
2656 dst, n, 0, CCL_MODE_DECODING);
2658 case CODESYS_ISO2022:
2659 decode_coding_iso2022 (decoding, src, dst, n);
2667 DEFUN ("decode-coding-region", Fdecode_coding_region, 3, 4, 0, /*
2668 Decode the text between START and END which is encoded in CODING-SYSTEM.
2669 This is useful if you've read in encoded text from a file without decoding
2670 it (e.g. you read in a JIS-formatted file but used the `binary' or
2671 `no-conversion' coding system, so that it shows up as "^[$B!<!+^[(B").
2672 Return length of decoded text.
2673 BUFFER defaults to the current buffer if unspecified.
2675 (start, end, coding_system, buffer))
2678 struct buffer *buf = decode_buffer (buffer, 0);
2679 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2680 Lstream *istr, *ostr;
2681 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2683 get_buffer_range_char (buf, start, end, &b, &e, 0);
2685 barf_if_buffer_read_only (buf, b, e);
2687 coding_system = Fget_coding_system (coding_system);
2688 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2689 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
2690 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
2692 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
2693 Fget_coding_system (Qbinary));
2694 istr = XLSTREAM (instream);
2695 ostr = XLSTREAM (outstream);
2696 GCPRO4 (instream, lb_outstream, de_outstream, outstream);
2698 /* The chain of streams looks like this:
2700 [BUFFER] <----- send through
2701 ------> [ENCODE AS BINARY]
2702 ------> [DECODE AS SPECIFIED]
2708 char tempbuf[1024]; /* some random amount */
2709 Bufpos newpos, even_newer_pos;
2710 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
2711 Lstream_data_count size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
2715 newpos = lisp_buffer_stream_startpos (istr);
2716 Lstream_write (ostr, tempbuf, size_in_bytes);
2717 even_newer_pos = lisp_buffer_stream_startpos (istr);
2718 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
2721 Lstream_close (istr);
2722 Lstream_close (ostr);
2724 Lstream_delete (istr);
2725 Lstream_delete (ostr);
2726 Lstream_delete (XLSTREAM (de_outstream));
2727 Lstream_delete (XLSTREAM (lb_outstream));
2732 /************************************************************************/
2733 /* Converting to an external encoding ("encoding") */
2734 /************************************************************************/
2736 /* An encoding stream is an output stream. When you create the
2737 stream, you specify the coding system that governs the encoding
2738 and another stream that the resulting encoded data is to be
2739 sent to, and then start sending data to it. */
2741 #define ENCODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, encoding)
2743 struct encoding_stream
2745 /* Coding system that governs the conversion. */
2746 Lisp_Coding_System *codesys;
2748 /* Stream that we read the encoded data from or
2749 write the decoded data to. */
2752 /* If we are reading, then we can return only a fixed amount of
2753 data, so if the conversion resulted in too much data, we store it
2754 here for retrieval the next time around. */
2755 unsigned_char_dynarr *runoff;
2757 /* FLAGS holds flags indicating the current state of the encoding.
2758 Some of these flags are dependent on the coding system. */
2761 /* CH holds a partially built-up character. Since we only deal
2762 with one- and two-byte characters at the moment, we only use
2763 this to store the first byte of a two-byte character. */
2766 /* Additional information used by the ISO2022 encoder. */
2769 /* CHARSET holds the character sets currently assigned to the G0
2770 through G3 registers. It is initialized from the array
2771 INITIAL_CHARSET in CODESYS. */
2772 Lisp_Object charset[4];
2774 /* Which registers are currently invoked into the left (GL) and
2775 right (GR) halves of the 8-bit encoding space? */
2776 int register_left, register_right;
2778 /* Whether we need to explicitly designate the charset in the
2779 G? register before using it. It is initialized from the
2780 array FORCE_CHARSET_ON_OUTPUT in CODESYS. */
2781 unsigned char force_charset_on_output[4];
2783 /* Other state variables that need to be preserved across
2785 Lisp_Object current_charset;
2787 int current_char_boundary;
2790 void (*encode_char) (struct encoding_stream *str, Emchar c,
2791 unsigned_char_dynarr *dst, unsigned int *flags);
2792 void (*finish) (struct encoding_stream *str,
2793 unsigned_char_dynarr *dst, unsigned int *flags);
2795 /* Additional information (the state of the running CCL program)
2796 used by the CCL encoder. */
2797 struct ccl_program ccl;
2801 static Lstream_data_count encoding_reader (Lstream *stream, unsigned char *data, Lstream_data_count size);
2802 static Lstream_data_count encoding_writer (Lstream *stream, const unsigned char *data,
2803 Lstream_data_count size);
2804 static int encoding_rewinder (Lstream *stream);
2805 static int encoding_seekable_p (Lstream *stream);
2806 static int encoding_flusher (Lstream *stream);
2807 static int encoding_closer (Lstream *stream);
2809 static Lisp_Object encoding_marker (Lisp_Object stream);
2811 DEFINE_LSTREAM_IMPLEMENTATION ("encoding", lstream_encoding,
2812 sizeof (struct encoding_stream));
2815 encoding_marker (Lisp_Object stream)
2817 Lstream *str = ENCODING_STREAM_DATA (XLSTREAM (stream))->other_end;
2818 Lisp_Object str_obj;
2820 /* We do not need to mark the coding systems or charsets stored
2821 within the stream because they are stored in a global list
2822 and automatically marked. */
2824 XSETLSTREAM (str_obj, str);
2825 mark_object (str_obj);
2826 if (str->imp->marker)
2827 return (str->imp->marker) (str_obj);
2832 /* Read SIZE bytes of data and store it into DATA. We are a encoding stream
2833 so we read data from the other end, encode it, and store it into DATA. */
2835 static Lstream_data_count
2836 encoding_reader (Lstream *stream, unsigned char *data, Lstream_data_count size)
2838 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2839 unsigned char *orig_data = data;
2840 Lstream_data_count read_size;
2841 int error_occurred = 0;
2843 /* We need to interface to mule_encode(), which expects to take some
2844 amount of data and store the result into a Dynarr. We have
2845 mule_encode() store into str->runoff, and take data from there
2848 /* We loop until we have enough data, reading chunks from the other
2849 end and encoding it. */
2852 /* Take data from the runoff if we can. Make sure to take at
2853 most SIZE bytes, and delete the data from the runoff. */
2854 if (Dynarr_length (str->runoff) > 0)
2856 int chunk = min ((int) size, Dynarr_length (str->runoff));
2857 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2858 Dynarr_delete_many (str->runoff, 0, chunk);
2864 break; /* No more room for data */
2866 if (str->flags & CODING_STATE_END)
2867 /* This means that on the previous iteration, we hit the EOF on
2868 the other end. We loop once more so that mule_encode() can
2869 output any final stuff it may be holding, or any "go back
2870 to a sane state" escape sequences. (This latter makes sense
2871 during encoding.) */
2874 /* Exhausted the runoff, so get some more. DATA at least SIZE bytes
2875 left of storage in it, so it's OK to read directly into it.
2876 (We'll be overwriting above, after we've encoded it into the
2878 read_size = Lstream_read (str->other_end, data, size);
2885 /* There might be some more end data produced in the translation.
2886 See the comment above. */
2887 str->flags |= CODING_STATE_END;
2888 mule_encode (stream, data, str->runoff, read_size);
2891 if (data == orig_data)
2892 return error_occurred ? -1 : 0;
2894 return data - orig_data;
2897 static Lstream_data_count
2898 encoding_writer (Lstream *stream, const unsigned char *data, Lstream_data_count size)
2900 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2901 Lstream_data_count retval;
2903 /* Encode all our data into the runoff, and then attempt to write
2904 it all out to the other end. Remove whatever chunk we succeeded
2906 mule_encode (stream, data, str->runoff, size);
2907 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
2908 Dynarr_length (str->runoff));
2910 Dynarr_delete_many (str->runoff, 0, retval);
2911 /* Do NOT return retval. The return value indicates how much
2912 of the incoming data was written, not how many bytes were
2918 reset_encoding_stream (struct encoding_stream *str)
2921 switch (CODING_SYSTEM_TYPE (str->codesys))
2923 case CODESYS_ISO2022:
2927 str->encode_char = &char_encode_iso2022;
2928 str->finish = &char_finish_iso2022;
2929 for (i = 0; i < 4; i++)
2931 str->iso2022.charset[i] =
2932 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, i);
2933 str->iso2022.force_charset_on_output[i] =
2934 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (str->codesys, i);
2936 str->iso2022.register_left = 0;
2937 str->iso2022.register_right = 1;
2938 str->iso2022.current_charset = Qnil;
2939 str->iso2022.current_half = 0;
2943 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_ENCODE (str->codesys));
2946 str->encode_char = &char_encode_utf8;
2947 str->finish = &char_finish_utf8;
2950 str->encode_char = &char_encode_utf16;
2951 str->finish = &char_finish_utf16;
2954 str->encode_char = &char_encode_ucs4;
2955 str->finish = &char_finish_ucs4;
2957 case CODESYS_SHIFT_JIS:
2958 str->encode_char = &char_encode_shift_jis;
2959 str->finish = &char_finish_shift_jis;
2962 str->encode_char = &char_encode_big5;
2963 str->finish = &char_finish_big5;
2969 str->iso2022.current_char_boundary = 0;
2970 str->flags = str->ch = 0;
2974 encoding_rewinder (Lstream *stream)
2976 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2977 reset_encoding_stream (str);
2978 Dynarr_reset (str->runoff);
2979 return Lstream_rewind (str->other_end);
2983 encoding_seekable_p (Lstream *stream)
2985 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2986 return Lstream_seekable_p (str->other_end);
2990 encoding_flusher (Lstream *stream)
2992 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2993 return Lstream_flush (str->other_end);
2997 encoding_closer (Lstream *stream)
2999 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3000 if (stream->flags & LSTREAM_FL_WRITE)
3002 str->flags |= CODING_STATE_END;
3003 encoding_writer (stream, 0, 0);
3005 Dynarr_free (str->runoff);
3006 return Lstream_close (str->other_end);
3010 encoding_stream_coding_system (Lstream *stream)
3012 Lisp_Object coding_system;
3013 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
3015 XSETCODING_SYSTEM (coding_system, str->codesys);
3016 return coding_system;
3020 set_encoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
3022 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
3023 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
3025 reset_encoding_stream (str);
3029 make_encoding_stream_1 (Lstream *stream, Lisp_Object codesys,
3032 Lstream *lstr = Lstream_new (lstream_encoding, mode);
3033 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
3037 str->runoff = Dynarr_new (unsigned_char);
3038 str->other_end = stream;
3039 set_encoding_stream_coding_system (lstr, codesys);
3040 XSETLSTREAM (obj, lstr);
3045 make_encoding_input_stream (Lstream *stream, Lisp_Object codesys)
3047 return make_encoding_stream_1 (stream, codesys, "r");
3051 make_encoding_output_stream (Lstream *stream, Lisp_Object codesys)
3053 return make_encoding_stream_1 (stream, codesys, "w");
3056 /* Convert N bytes of internally-formatted data stored in SRC to an
3057 external format, according to the encoding stream ENCODING.
3058 Store the encoded data into DST. */
3061 mule_encode (Lstream *encoding, const Bufbyte *src,
3062 unsigned_char_dynarr *dst, Lstream_data_count n)
3064 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3066 switch (CODING_SYSTEM_TYPE (str->codesys))
3069 case CODESYS_INTERNAL:
3070 Dynarr_add_many (dst, src, n);
3073 case CODESYS_AUTODETECT:
3074 /* If we got this far and still haven't decided on the coding
3075 system, then do no conversion. */
3076 case CODESYS_NO_CONVERSION:
3077 encode_coding_no_conversion (encoding, src, dst, n);
3081 str->ccl.last_block = str->flags & CODING_STATE_END;
3082 /* When applying ccl program to stream, MUST NOT set NULL
3084 ccl_driver (&str->ccl, ((src) ? src : (unsigned char*)""),
3085 dst, n, 0, CCL_MODE_ENCODING);
3089 text_encode_generic (encoding, src, dst, n);
3093 DEFUN ("encode-coding-region", Fencode_coding_region, 3, 4, 0, /*
3094 Encode the text between START and END using CODING-SYSTEM.
3095 This will, for example, convert Japanese characters into stuff such as
3096 "^[$B!<!+^[(B" if you use the JIS encoding. Return length of encoded
3097 text. BUFFER defaults to the current buffer if unspecified.
3099 (start, end, coding_system, buffer))
3102 struct buffer *buf = decode_buffer (buffer, 0);
3103 Lisp_Object instream, lb_outstream, de_outstream, outstream;
3104 Lstream *istr, *ostr;
3105 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
3107 get_buffer_range_char (buf, start, end, &b, &e, 0);
3109 barf_if_buffer_read_only (buf, b, e);
3111 coding_system = Fget_coding_system (coding_system);
3112 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
3113 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
3114 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
3115 Fget_coding_system (Qbinary));
3116 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
3118 istr = XLSTREAM (instream);
3119 ostr = XLSTREAM (outstream);
3120 GCPRO4 (instream, outstream, de_outstream, lb_outstream);
3121 /* The chain of streams looks like this:
3123 [BUFFER] <----- send through
3124 ------> [ENCODE AS SPECIFIED]
3125 ------> [DECODE AS BINARY]
3130 char tempbuf[1024]; /* some random amount */
3131 Bufpos newpos, even_newer_pos;
3132 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
3133 Lstream_data_count size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
3137 newpos = lisp_buffer_stream_startpos (istr);
3138 Lstream_write (ostr, tempbuf, size_in_bytes);
3139 even_newer_pos = lisp_buffer_stream_startpos (istr);
3140 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
3146 lisp_buffer_stream_startpos (XLSTREAM (instream)) - b;
3147 Lstream_close (istr);
3148 Lstream_close (ostr);
3150 Lstream_delete (istr);
3151 Lstream_delete (ostr);
3152 Lstream_delete (XLSTREAM (de_outstream));
3153 Lstream_delete (XLSTREAM (lb_outstream));
3154 return make_int (retlen);
3161 text_encode_generic (Lstream *encoding, const Bufbyte *src,
3162 unsigned_char_dynarr *dst, Lstream_data_count n)
3165 unsigned char char_boundary;
3166 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3167 unsigned int flags = str->flags;
3168 Emchar ch = str->ch;
3170 char_boundary = str->iso2022.current_char_boundary;
3176 if (char_boundary == 0)
3204 (*str->encode_char) (str, c, dst, &flags);
3206 else if (char_boundary == 1)
3208 (*str->encode_char) (str, (ch << 6) | (c & 0x3f), dst, &flags);
3214 ch = (ch << 6) | (c & 0x3f);
3219 if ((char_boundary == 0) && (flags & CODING_STATE_END))
3221 (*str->finish) (str, dst, &flags);
3226 str->iso2022.current_char_boundary = char_boundary;
3231 /************************************************************************/
3232 /* entity reference */
3233 /************************************************************************/
3236 decode_flush_er_chars (struct decoding_stream *str, unsigned_char_dynarr* dst);
3238 decode_flush_er_chars (struct decoding_stream *str, unsigned_char_dynarr* dst)
3240 if ( str->er_counter > 0)
3242 Dynarr_add_many (dst, str->er_buf, str->er_counter);
3243 str->er_counter = 0;
3247 void decode_add_er_char (struct decoding_stream *str, Emchar character,
3248 unsigned_char_dynarr* dst);
3250 decode_add_er_char (struct decoding_stream *str, Emchar c,
3251 unsigned_char_dynarr* dst)
3253 if (str->er_counter == 0)
3255 if (CODING_SYSTEM_USE_ENTITY_REFERENCE (str->codesys)
3258 str->er_buf[0] = '&';
3262 DECODE_ADD_UCS_CHAR (c, dst);
3266 Lisp_Object string = make_string (str->er_buf,
3268 Lisp_Object rest = Vcoded_charset_entity_reference_alist;
3273 Lisp_Object char_type;
3276 while (!NILP (rest))
3282 char_type = XCDR (ccs);
3287 if (NILP (ccs = Ffind_charset (ccs)))
3302 pat = concat3 (build_string ("^&"),
3303 pat, build_string ("\\([0-9]+\\)$"));
3306 else if (EQ (ret, Qx))
3308 pat = concat3 (build_string ("^&"),
3309 pat, build_string ("\\([0-9a-f]+\\)$"));
3312 else if (EQ (ret, QX))
3314 pat = concat3 (build_string ("^&"),
3315 pat, build_string ("\\([0-9A-F]+\\)$"));
3321 if (!NILP (Fstring_match (pat, string, Qnil, Qnil)))
3324 = XINT (Fstring_to_number
3325 (Fsubstring (string,
3326 Fmatch_beginning (make_int (1)),
3327 Fmatch_end (make_int (1))),
3331 ? DECODE_CHAR (ccs, code)
3332 : decode_builtin_char (ccs, code);
3334 DECODE_ADD_UCS_CHAR (chr, dst);
3339 if (!NILP (Fstring_match (build_string ("^&MCS-\\([0-9A-F]+\\)$"),
3340 string, Qnil, Qnil)))
3343 = XINT (Fstring_to_number
3344 (Fsubstring (string,
3345 Fmatch_beginning (make_int (1)),
3346 Fmatch_end (make_int (1))),
3349 DECODE_ADD_UCS_CHAR (code, dst);
3353 Dynarr_add_many (dst, str->er_buf, str->er_counter);
3354 Dynarr_add (dst, ';');
3357 str->er_counter = 0;
3359 else if ( (str->er_counter >= 16) || (c >= 0x7F) )
3361 Dynarr_add_many (dst, str->er_buf, str->er_counter);
3362 str->er_counter = 0;
3363 DECODE_ADD_UCS_CHAR (c, dst);
3366 str->er_buf[str->er_counter++] = c;
3369 void char_encode_as_entity_reference (Emchar ch, char* buf);
3371 char_encode_as_entity_reference (Emchar ch, char* buf)
3373 Lisp_Object rest = Vcoded_charset_entity_reference_alist;
3376 Lisp_Object char_type;
3377 int format_columns, idx;
3380 while (!NILP (rest))
3386 char_type = XCDR (ccs);
3391 if (!NILP (ccs = Ffind_charset (ccs)))
3393 int code_point = charset_code_point (ccs, ch, 0);
3395 if ( (code_point >= 0)
3396 && (NILP (char_type)
3397 || DECODE_CHAR (ccs, code_point) != ch) )
3403 if (STRINGP (ret) && ((idx = XSTRING_LENGTH (ret)) <= 6))
3406 strncpy (&format[1], XSTRING_DATA (ret), idx);
3416 format[idx++] = '%';
3417 format_columns = XINT (ret);
3418 if ( (2 <= format_columns) && (format_columns <= 8) )
3420 format [idx++] = '0';
3421 format [idx++] = '0' + format_columns;
3428 format [idx++] = 'd';
3429 else if (EQ (ret, Qx))
3430 format [idx++] = 'x';
3431 else if (EQ (ret, QX))
3432 format [idx++] = 'X';
3435 format [idx++] = ';';
3438 sprintf (buf, format, code_point);
3444 sprintf (buf, "&MCS-%08X;", ch);
3448 /************************************************************************/
3449 /* character composition */
3450 /************************************************************************/
3451 extern Lisp_Object Qcomposition;
3454 COMPOSE_FLUSH_CHARS (struct decoding_stream *str, unsigned_char_dynarr* dst);
3456 COMPOSE_FLUSH_CHARS (struct decoding_stream *str, unsigned_char_dynarr* dst)
3460 for (i = 0; i < str->combined_char_count; i++)
3461 decode_add_er_char (str, str->combined_chars[i], dst);
3462 str->combined_char_count = 0;
3463 str->combining_table = Qnil;
3466 void COMPOSE_ADD_CHAR (struct decoding_stream *str, Emchar character,
3467 unsigned_char_dynarr* dst);
3469 COMPOSE_ADD_CHAR (struct decoding_stream *str,
3470 Emchar character, unsigned_char_dynarr* dst)
3472 if (CODING_SYSTEM_DISABLE_COMPOSITION (str->codesys))
3473 decode_add_er_char (str, character, dst);
3474 else if (!CONSP (str->combining_table))
3477 = Fget_char_attribute (make_char (character), Qcomposition, Qnil);
3480 decode_add_er_char (str, character, dst);
3483 str->combined_chars[0] = character;
3484 str->combined_char_count = 1;
3485 str->combining_table = ret;
3491 = Fcdr (Fassq (make_char (character), str->combining_table));
3495 Emchar char2 = XCHARVAL (ret);
3496 ret = Fget_char_attribute (make_char (character), Qcomposition,
3500 decode_add_er_char (str, character, dst);
3501 str->combined_char_count = 0;
3502 str->combining_table = Qnil;
3506 str->combined_chars[0] = char2;
3507 str->combined_char_count = 1;
3508 str->combining_table = ret;
3513 COMPOSE_FLUSH_CHARS (str, dst);
3514 decode_add_er_char (str, character, dst);
3518 #else /* not UTF2000 */
3519 #define COMPOSE_FLUSH_CHARS(str, dst)
3520 #define COMPOSE_ADD_CHAR(str, ch, dst) DECODE_ADD_UCS_CHAR (ch, dst)
3521 #endif /* UTF2000 */
3524 /************************************************************************/
3525 /* Shift-JIS methods */
3526 /************************************************************************/
3528 /* Shift-JIS is a coding system encoding three character sets: ASCII, right
3529 half of JISX0201-Kana, and JISX0208. An ASCII character is encoded
3530 as is. A character of JISX0201-Kana (DIMENSION1_CHARS94 character set) is
3531 encoded by "position-code + 0x80". A character of JISX0208
3532 (DIMENSION2_CHARS94 character set) is encoded in 2-byte but two
3533 position-codes are divided and shifted so that it fit in the range
3536 --- CODE RANGE of Shift-JIS ---
3537 (character set) (range)
3539 JISX0201-Kana 0xA0 .. 0xDF
3540 JISX0208 (1st byte) 0x80 .. 0x9F and 0xE0 .. 0xEF
3541 (2nd byte) 0x40 .. 0x7E and 0x80 .. 0xFC
3542 -------------------------------
3546 /* Is this the first byte of a Shift-JIS two-byte char? */
3548 #define BYTE_SJIS_TWO_BYTE_1_P(c) \
3549 (((c) >= 0x81 && (c) <= 0x9F) || ((c) >= 0xE0 && (c) <= 0xEF))
3551 /* Is this the second byte of a Shift-JIS two-byte char? */
3553 #define BYTE_SJIS_TWO_BYTE_2_P(c) \
3554 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0x80 && (c) <= 0xFC))
3556 #define BYTE_SJIS_KATAKANA_P(c) \
3557 ((c) >= 0xA1 && (c) <= 0xDF)
3560 detect_coding_sjis (struct detection_state *st, const Extbyte *src, Lstream_data_count n)
3564 unsigned char c = *(unsigned char *)src++;
3565 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
3567 if (st->shift_jis.in_second_byte)
3569 st->shift_jis.in_second_byte = 0;
3573 else if ((c >= 0x80 && c < 0xA0) || c >= 0xE0)
3574 st->shift_jis.in_second_byte = 1;
3576 return CODING_CATEGORY_SHIFT_JIS_MASK;
3579 /* Convert Shift-JIS data to internal format. */
3582 decode_coding_sjis (Lstream *decoding, const Extbyte *src,
3583 unsigned_char_dynarr *dst, Lstream_data_count n)
3585 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3586 unsigned int flags = str->flags;
3587 unsigned int cpos = str->cpos;
3588 eol_type_t eol_type = str->eol_type;
3592 unsigned char c = *(unsigned char *)src++;
3596 /* Previous character was first byte of Shift-JIS Kanji char. */
3597 if (BYTE_SJIS_TWO_BYTE_2_P (c))
3599 unsigned char e1, e2;
3601 DECODE_SJIS (cpos, c, e1, e2);
3603 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_japanese_jisx0208,
3607 Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208);
3608 Dynarr_add (dst, e1);
3609 Dynarr_add (dst, e2);
3614 DECODE_ADD_BINARY_CHAR (cpos, dst);
3615 DECODE_ADD_BINARY_CHAR (c, dst);
3621 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3622 if (BYTE_SJIS_TWO_BYTE_1_P (c))
3624 else if (BYTE_SJIS_KATAKANA_P (c))
3627 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_katakana_jisx0201,
3630 Dynarr_add (dst, LEADING_BYTE_KATAKANA_JISX0201);
3631 Dynarr_add (dst, c);
3636 DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_latin_jisx0201,
3640 DECODE_ADD_BINARY_CHAR (c, dst);
3642 label_continue_loop:;
3645 DECODE_HANDLE_END_OF_CONVERSION (flags, cpos, dst);
3651 /* Convert internal character representation to Shift_JIS. */
3654 char_encode_shift_jis (struct encoding_stream *str, Emchar ch,
3655 unsigned_char_dynarr *dst, unsigned int *flags)
3657 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3661 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3662 Dynarr_add (dst, '\r');
3663 if (eol_type != EOL_CR)
3664 Dynarr_add (dst, ch);
3668 unsigned int s1, s2;
3670 int code_point = charset_code_point (Vcharset_latin_jisx0201, ch, 0);
3672 if (code_point >= 0)
3673 Dynarr_add (dst, code_point);
3674 else if ((code_point
3675 = charset_code_point (Vcharset_japanese_jisx0208_1990, ch, 0))
3678 ENCODE_SJIS ((code_point >> 8) | 0x80,
3679 (code_point & 0xFF) | 0x80, s1, s2);
3680 Dynarr_add (dst, s1);
3681 Dynarr_add (dst, s2);
3683 else if ((code_point
3684 = charset_code_point (Vcharset_katakana_jisx0201, ch, 0))
3686 Dynarr_add (dst, code_point | 0x80);
3687 else if ((code_point
3688 = charset_code_point (Vcharset_japanese_jisx0208, ch, 0))
3691 ENCODE_SJIS ((code_point >> 8) | 0x80,
3692 (code_point & 0xFF) | 0x80, s1, s2);
3693 Dynarr_add (dst, s1);
3694 Dynarr_add (dst, s2);
3696 else if ((code_point = charset_code_point (Vcharset_ascii, ch, 0))
3698 Dynarr_add (dst, code_point);
3700 Dynarr_add (dst, '?');
3702 Lisp_Object charset;
3703 unsigned int c1, c2;
3705 BREAKUP_CHAR (ch, charset, c1, c2);
3707 if (EQ(charset, Vcharset_katakana_jisx0201))
3709 Dynarr_add (dst, c1 | 0x80);
3713 Dynarr_add (dst, c1);
3715 else if (EQ(charset, Vcharset_japanese_jisx0208))
3717 ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
3718 Dynarr_add (dst, s1);
3719 Dynarr_add (dst, s2);
3722 Dynarr_add (dst, '?');
3728 char_finish_shift_jis (struct encoding_stream *str, unsigned_char_dynarr *dst,
3729 unsigned int *flags)
3733 DEFUN ("decode-shift-jis-char", Fdecode_shift_jis_char, 1, 1, 0, /*
3734 Decode a JISX0208 character of Shift-JIS coding-system.
3735 CODE is the character code in Shift-JIS as a cons of type bytes.
3736 Return the corresponding character.
3740 unsigned char c1, c2, s1, s2;
3743 CHECK_INT (XCAR (code));
3744 CHECK_INT (XCDR (code));
3745 s1 = XINT (XCAR (code));
3746 s2 = XINT (XCDR (code));
3747 if (BYTE_SJIS_TWO_BYTE_1_P (s1) &&
3748 BYTE_SJIS_TWO_BYTE_2_P (s2))
3750 DECODE_SJIS (s1, s2, c1, c2);
3751 return make_char (MAKE_CHAR (Vcharset_japanese_jisx0208,
3752 c1 & 0x7F, c2 & 0x7F));
3758 DEFUN ("encode-shift-jis-char", Fencode_shift_jis_char, 1, 1, 0, /*
3759 Encode a JISX0208 character CHARACTER to SHIFT-JIS coding-system.
3760 Return the corresponding character code in SHIFT-JIS as a cons of two bytes.
3764 Lisp_Object charset;
3767 CHECK_CHAR_COERCE_INT (character);
3768 BREAKUP_CHAR (XCHAR (character), charset, c1, c2);
3769 if (EQ (charset, Vcharset_japanese_jisx0208))
3771 ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
3772 return Fcons (make_int (s1), make_int (s2));
3779 /************************************************************************/
3781 /************************************************************************/
3783 /* BIG5 is a coding system encoding two character sets: ASCII and
3784 Big5. An ASCII character is encoded as is. Big5 is a two-byte
3785 character set and is encoded in two-byte.
3787 --- CODE RANGE of BIG5 ---
3788 (character set) (range)
3790 Big5 (1st byte) 0xA1 .. 0xFE
3791 (2nd byte) 0x40 .. 0x7E and 0xA1 .. 0xFE
3792 --------------------------
3794 Since the number of characters in Big5 is larger than maximum
3795 characters in Emacs' charset (96x96), it can't be handled as one
3796 charset. So, in Emacs, Big5 is divided into two: `charset-big5-1'
3797 and `charset-big5-2'. Both <type>s are DIMENSION2_CHARS94. The former
3798 contains frequently used characters and the latter contains less
3799 frequently used characters. */
3802 #define BYTE_BIG5_TWO_BYTE_1_P(c) \
3803 ((c) >= 0x81 && (c) <= 0xFE)
3805 #define BYTE_BIG5_TWO_BYTE_1_P(c) \
3806 ((c) >= 0xA1 && (c) <= 0xFE)
3809 /* Is this the second byte of a Shift-JIS two-byte char? */
3811 #define BYTE_BIG5_TWO_BYTE_2_P(c) \
3812 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0xA1 && (c) <= 0xFE))
3814 /* Number of Big5 characters which have the same code in 1st byte. */
3816 #define BIG5_SAME_ROW (0xFF - 0xA1 + 0x7F - 0x40)
3818 /* Code conversion macros. These are macros because they are used in
3819 inner loops during code conversion.
3821 Note that temporary variables in macros introduce the classic
3822 dynamic-scoping problems with variable names. We use capital-
3823 lettered variables in the assumption that XEmacs does not use
3824 capital letters in variables except in a very formalized way
3827 /* Convert Big5 code (b1, b2) into its internal string representation
3830 /* There is a much simpler way to split the Big5 charset into two.
3831 For the moment I'm going to leave the algorithm as-is because it
3832 claims to separate out the most-used characters into a single
3833 charset, which perhaps will lead to optimizations in various
3836 The way the algorithm works is something like this:
3838 Big5 can be viewed as a 94x157 charset, where the row is
3839 encoded into the bytes 0xA1 .. 0xFE and the column is encoded
3840 into the bytes 0x40 .. 0x7E and 0xA1 .. 0xFE. As for frequency,
3841 the split between low and high column numbers is apparently
3842 meaningless; ascending rows produce less and less frequent chars.
3843 Therefore, we assign the lower half of rows (0xA1 .. 0xC8) to
3844 the first charset, and the upper half (0xC9 .. 0xFE) to the
3845 second. To do the conversion, we convert the character into
3846 a single number where 0 .. 156 is the first row, 157 .. 313
3847 is the second, etc. That way, the characters are ordered by
3848 decreasing frequency. Then we just chop the space in two
3849 and coerce the result into a 94x94 space.
3852 #define DECODE_BIG5(b1, b2, lb, c1, c2) do \
3854 int B1 = b1, B2 = b2; \
3856 = (B1 - 0xA1) * BIG5_SAME_ROW + B2 - (B2 < 0x7F ? 0x40 : 0x62); \
3860 lb = LEADING_BYTE_CHINESE_BIG5_1; \
3864 lb = LEADING_BYTE_CHINESE_BIG5_2; \
3865 I -= (BIG5_SAME_ROW) * (0xC9 - 0xA1); \
3867 c1 = I / (0xFF - 0xA1) + 0xA1; \
3868 c2 = I % (0xFF - 0xA1) + 0xA1; \
3871 /* Convert the internal string representation of a Big5 character
3872 (lb, c1, c2) into Big5 code (b1, b2). */
3874 #define ENCODE_BIG5(lb, c1, c2, b1, b2) do \
3876 unsigned int I = ((c1) - 0xA1) * (0xFF - 0xA1) + ((c2) - 0xA1); \
3878 if (lb == LEADING_BYTE_CHINESE_BIG5_2) \
3880 I += BIG5_SAME_ROW * (0xC9 - 0xA1); \
3882 b1 = I / BIG5_SAME_ROW + 0xA1; \
3883 b2 = I % BIG5_SAME_ROW; \
3884 b2 += b2 < 0x3F ? 0x40 : 0x62; \
3888 detect_coding_big5 (struct detection_state *st, const Extbyte *src, Lstream_data_count n)
3892 unsigned char c = *(unsigned char *)src++;
3893 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO
3895 || (c >= 0x80 && c <= 0xA0)
3899 if (st->big5.in_second_byte)
3901 st->big5.in_second_byte = 0;
3902 if (c < 0x40 || (c >= 0x80 && c <= 0xA0))
3912 st->big5.in_second_byte = 1;
3914 return CODING_CATEGORY_BIG5_MASK;
3917 /* Convert Big5 data to internal format. */
3920 decode_coding_big5 (Lstream *decoding, const Extbyte *src,
3921 unsigned_char_dynarr *dst, Lstream_data_count n)
3923 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3924 unsigned int flags = str->flags;
3925 unsigned int cpos = str->cpos;
3926 eol_type_t eol_type = str->eol_type;
3929 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (DECODING_STREAM_DATA
3930 (decoding)->codesys, 1);
3935 unsigned char c = *(unsigned char *)src++;
3938 /* Previous character was first byte of Big5 char. */
3939 if (BYTE_BIG5_TWO_BYTE_2_P (c))
3942 int code_point = (cpos << 8) | c;
3943 Emchar char_id = decode_defined_char (ccs, code_point);
3946 char_id = DECODE_CHAR (Vcharset_chinese_big5, code_point);
3947 DECODE_ADD_UCS_CHAR (char_id, dst);
3949 unsigned char b1, b2, b3;
3950 DECODE_BIG5 (cpos, c, b1, b2, b3);
3951 Dynarr_add (dst, b1);
3952 Dynarr_add (dst, b2);
3953 Dynarr_add (dst, b3);
3958 DECODE_ADD_BINARY_CHAR (cpos, dst);
3959 DECODE_ADD_BINARY_CHAR (c, dst);
3965 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3966 if (BYTE_BIG5_TWO_BYTE_1_P (c))
3968 decode_flush_er_chars (str, dst);
3973 decode_flush_er_chars (str, dst);
3974 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3975 DECODE_ADD_BINARY_CHAR (c, dst);
3979 /* DECODE_ADD_BINARY_CHAR (c, dst); */
3980 decode_add_er_char (str, c, dst);
3983 label_continue_loop:;
3986 /* DECODE_HANDLE_END_OF_CONVERSION (flags, cpos, dst); */
3987 if (flags & CODING_STATE_END)
3989 decode_flush_er_chars (str, dst);
3990 DECODE_OUTPUT_PARTIAL_CHAR (cpos);
3991 if (flags & CODING_STATE_CR)
3992 Dynarr_add (dst, '\r');
3999 /* Convert internally-formatted data to Big5. */
4002 char_encode_big5 (struct encoding_stream *str, Emchar ch,
4003 unsigned_char_dynarr *dst, unsigned int *flags)
4005 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
4009 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
4010 Dynarr_add (dst, '\r');
4011 if (eol_type != EOL_CR)
4012 Dynarr_add (dst, ch);
4019 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, 1);
4021 if ((code_point = charset_code_point (Vcharset_ascii, ch, 0)) >= 0)
4022 Dynarr_add (dst, code_point);
4023 else if ((code_point = charset_code_point (ccs, ch, 0)) >= 0)
4025 Dynarr_add (dst, code_point >> 8);
4026 Dynarr_add (dst, code_point & 0xFF);
4028 else if ((code_point
4029 = charset_code_point (Vcharset_chinese_big5, ch, 0)) >= 0)
4031 Dynarr_add (dst, code_point >> 8);
4032 Dynarr_add (dst, code_point & 0xFF);
4034 else if ((code_point
4035 = charset_code_point (Vcharset_chinese_big5_1, ch, 0)) >= 0)
4038 = ((code_point >> 8) - 33) * (0xFF - 0xA1)
4039 + ((code_point & 0xFF) - 33);
4040 unsigned char b1 = I / BIG5_SAME_ROW + 0xA1;
4041 unsigned char b2 = I % BIG5_SAME_ROW;
4043 b2 += b2 < 0x3F ? 0x40 : 0x62;
4044 Dynarr_add (dst, b1);
4045 Dynarr_add (dst, b2);
4047 else if ((code_point
4048 = charset_code_point (Vcharset_chinese_big5_2, ch, 0)) >= 0)
4051 = ((code_point >> 8) - 33) * (0xFF - 0xA1)
4052 + ((code_point & 0xFF) - 33);
4053 unsigned char b1, b2;
4055 I += BIG5_SAME_ROW * (0xC9 - 0xA1);
4056 b1 = I / BIG5_SAME_ROW + 0xA1;
4057 b2 = I % BIG5_SAME_ROW;
4058 b2 += b2 < 0x3F ? 0x40 : 0x62;
4059 Dynarr_add (dst, b1);
4060 Dynarr_add (dst, b2);
4062 else if (CODING_SYSTEM_USE_ENTITY_REFERENCE (str->codesys))
4066 char_encode_as_entity_reference (ch, buf);
4067 Dynarr_add_many (dst, buf, strlen (buf));
4070 Dynarr_add (dst, '?');
4077 char_finish_big5 (struct encoding_stream *str, unsigned_char_dynarr *dst,
4078 unsigned int *flags)
4083 DEFUN ("decode-big5-char", Fdecode_big5_char, 1, 1, 0, /*
4084 Decode a Big5 character CODE of BIG5 coding-system.
4085 CODE is the character code in BIG5, a cons of two integers.
4086 Return the corresponding character.
4090 unsigned char c1, c2, b1, b2;
4093 CHECK_INT (XCAR (code));
4094 CHECK_INT (XCDR (code));
4095 b1 = XINT (XCAR (code));
4096 b2 = XINT (XCDR (code));
4097 if (BYTE_BIG5_TWO_BYTE_1_P (b1) &&
4098 BYTE_BIG5_TWO_BYTE_2_P (b2))
4100 Charset_ID leading_byte;
4101 Lisp_Object charset;
4102 DECODE_BIG5 (b1, b2, leading_byte, c1, c2);
4103 charset = CHARSET_BY_LEADING_BYTE (leading_byte);
4104 return make_char (MAKE_CHAR (charset, c1 & 0x7F, c2 & 0x7F));
4110 DEFUN ("encode-big5-char", Fencode_big5_char, 1, 1, 0, /*
4111 Encode the Big5 character CHARACTER in the BIG5 coding-system.
4112 Return the corresponding character code in Big5.
4116 Lisp_Object charset;
4119 CHECK_CHAR_COERCE_INT (character);
4120 BREAKUP_CHAR (XCHAR (character), charset, c1, c2);
4121 if (EQ (charset, Vcharset_chinese_big5_1) ||
4122 EQ (charset, Vcharset_chinese_big5_2))
4124 ENCODE_BIG5 (XCHARSET_LEADING_BYTE (charset), c1 | 0x80, c2 | 0x80,
4126 return Fcons (make_int (b1), make_int (b2));
4133 /************************************************************************/
4135 /************************************************************************/
4138 detect_coding_ucs4 (struct detection_state *st, const Extbyte *src, Lstream_data_count n)
4142 unsigned char c = *(unsigned char *)src++;
4143 switch (st->ucs4.in_byte)
4152 st->ucs4.in_byte = 0;
4158 return CODING_CATEGORY_UCS4_MASK;
4162 decode_coding_ucs4 (Lstream *decoding, const Extbyte *src,
4163 unsigned_char_dynarr *dst, Lstream_data_count n)
4165 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
4166 unsigned int flags = str->flags;
4167 unsigned int cpos = str->cpos;
4168 unsigned char counter = str->counter;
4172 unsigned char c = *(unsigned char *)src++;
4180 DECODE_ADD_UCS_CHAR ((cpos << 8) | c, dst);
4185 cpos = ( cpos << 8 ) | c;
4189 if (counter & CODING_STATE_END)
4190 DECODE_OUTPUT_PARTIAL_CHAR (cpos);
4194 str->counter = counter;
4198 char_encode_ucs4 (struct encoding_stream *str, Emchar ch,
4199 unsigned_char_dynarr *dst, unsigned int *flags)
4201 Dynarr_add (dst, ch >> 24);
4202 Dynarr_add (dst, ch >> 16);
4203 Dynarr_add (dst, ch >> 8);
4204 Dynarr_add (dst, ch );
4208 char_finish_ucs4 (struct encoding_stream *str, unsigned_char_dynarr *dst,
4209 unsigned int *flags)
4214 /************************************************************************/
4215 /* UTF-16 methods */
4216 /************************************************************************/
4219 detect_coding_utf16 (struct detection_state *st, const Extbyte *src, Lstream_data_count n)
4221 return CODING_CATEGORY_UTF16_MASK;
4225 decode_coding_utf16 (Lstream *decoding, const Extbyte *src,
4226 unsigned_char_dynarr *dst, Lstream_data_count n)
4228 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
4229 unsigned int flags = str->flags;
4230 unsigned int cpos = str->cpos;
4231 unsigned char counter = str->counter & 3;
4232 unsigned char byte_order = str->counter >> 2;
4233 eol_type_t eol_type = str->eol_type;
4237 unsigned char c = *(unsigned char *)src++;
4243 else if (counter == 1)
4247 if (byte_order == 0)
4248 code = (c << 8) | cpos;
4250 code = (cpos << 8) | c;
4253 code = ((code & 0xFF) << 8) | (code >> 8);
4254 if ( byte_order == 0 )
4259 if ( (0xD800 <= code) && (code <= 0xDBFF) )
4270 DECODE_HANDLE_EOL_TYPE (eol_type, code, flags, dst);
4271 DECODE_ADD_UCS_CHAR (code, dst);
4275 else if (counter == 2)
4277 cpos = (cpos << 8) | c;
4285 ? (c << 8) | (cpos & 0xFF)
4286 : ((cpos & 0xFF) << 8) | c;
4288 DECODE_ADD_UCS_CHAR ((x - 0xD800) * 0x400 + (y - 0xDC00)
4293 label_continue_loop:;
4295 if (counter & CODING_STATE_END)
4296 DECODE_OUTPUT_PARTIAL_CHAR (cpos);
4300 str->counter = (byte_order << 2) | counter;
4304 char_encode_utf16 (struct encoding_stream *str, Emchar ch,
4305 unsigned_char_dynarr *dst, unsigned int *flags)
4309 Dynarr_add (dst, ch);
4310 Dynarr_add (dst, ch >> 8);
4314 int y = ((ch - 0x10000) / 0x400) + 0xD800;
4315 int z = ((ch - 0x10000) % 0x400) + 0xDC00;
4317 Dynarr_add (dst, y);
4318 Dynarr_add (dst, y >> 8);
4319 Dynarr_add (dst, z);
4320 Dynarr_add (dst, z >> 8);
4325 char_finish_utf16 (struct encoding_stream *str, unsigned_char_dynarr *dst,
4326 unsigned int *flags)
4331 /************************************************************************/
4333 /************************************************************************/
4336 detect_coding_utf8 (struct detection_state *st, const Extbyte *src, Lstream_data_count n)
4340 unsigned char c = *(unsigned char *)src++;
4341 switch (st->utf8.in_byte)
4344 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
4347 st->utf8.in_byte = 5;
4349 st->utf8.in_byte = 4;
4351 st->utf8.in_byte = 3;
4353 st->utf8.in_byte = 2;
4355 st->utf8.in_byte = 1;
4360 if ((c & 0xc0) != 0x80)
4366 return CODING_CATEGORY_UTF8_MASK;
4370 decode_output_utf8_partial_char (unsigned char counter,
4372 unsigned_char_dynarr *dst)
4375 DECODE_ADD_BINARY_CHAR ( (cpos|0xFC), dst);
4376 else if (counter == 4)
4378 if (cpos < (1 << 6))
4379 DECODE_ADD_BINARY_CHAR ( (cpos|0xF8), dst);
4382 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xFC), dst);
4383 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4386 else if (counter == 3)
4388 if (cpos < (1 << 6))
4389 DECODE_ADD_BINARY_CHAR ( (cpos|0xF0), dst);
4390 else if (cpos < (1 << 12))
4392 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xF8), dst);
4393 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4397 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 12)|0xFC), dst);
4398 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4399 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4402 else if (counter == 2)
4404 if (cpos < (1 << 6))
4405 DECODE_ADD_BINARY_CHAR ( (cpos|0xE0), dst);
4406 else if (cpos < (1 << 12))
4408 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xF0), dst);
4409 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4411 else if (cpos < (1 << 18))
4413 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 12)|0xF8), dst);
4414 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4415 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4419 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 18)|0xFC), dst);
4420 DECODE_ADD_BINARY_CHAR ( (((cpos >> 12)&0x3F)|0x80), dst);
4421 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4422 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4427 if (cpos < (1 << 6))
4428 DECODE_ADD_BINARY_CHAR ( (cpos|0xC0), dst);
4429 else if (cpos < (1 << 12))
4431 DECODE_ADD_BINARY_CHAR ( ((cpos >> 6)|0xE0), dst);
4432 DECODE_ADD_BINARY_CHAR ( ((cpos&0x3F)|0x80), dst);
4434 else if (cpos < (1 << 18))
4436 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 12)|0xF0), dst);
4437 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4438 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4440 else if (cpos < (1 << 24))
4442 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 18)|0xF8), dst);
4443 DECODE_ADD_BINARY_CHAR ( (((cpos >> 12)&0x3F)|0x80), dst);
4444 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4445 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4449 DECODE_ADD_BINARY_CHAR ( ( (cpos >> 24)|0xFC), dst);
4450 DECODE_ADD_BINARY_CHAR ( (((cpos >> 18)&0x3F)|0x80), dst);
4451 DECODE_ADD_BINARY_CHAR ( (((cpos >> 12)&0x3F)|0x80), dst);
4452 DECODE_ADD_BINARY_CHAR ( (((cpos >> 6)&0x3F)|0x80), dst);
4453 DECODE_ADD_BINARY_CHAR ( ( (cpos &0x3F)|0x80), dst);
4459 decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
4460 unsigned_char_dynarr *dst, Lstream_data_count n)
4462 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
4463 unsigned int flags = str->flags;
4464 unsigned int cpos = str->cpos;
4465 eol_type_t eol_type = str->eol_type;
4466 unsigned char counter = str->counter;
4469 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (DECODING_STREAM_DATA
4470 (decoding)->codesys, 0);
4475 unsigned char c = *(unsigned char *)src++;
4480 COMPOSE_FLUSH_CHARS (str, dst);
4481 decode_flush_er_chars (str, dst);
4482 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
4483 DECODE_ADD_UCS_CHAR (c, dst);
4485 else if ( c < 0xC0 )
4486 /* decode_add_er_char (str, c, dst); */
4487 COMPOSE_ADD_CHAR (str, c, dst);
4490 /* decode_flush_er_chars (str, dst); */
4496 else if ( c < 0xF0 )
4501 else if ( c < 0xF8 )
4506 else if ( c < 0xFC )
4518 else if ( (c & 0xC0) == 0x80 )
4520 cpos = ( cpos << 6 ) | ( c & 0x3f );
4527 char_id = decode_defined_char (ccs, cpos);
4534 COMPOSE_ADD_CHAR (str, char_id, dst);
4543 COMPOSE_FLUSH_CHARS (str, dst);
4544 decode_flush_er_chars (str, dst);
4545 decode_output_utf8_partial_char (counter, cpos, dst);
4546 DECODE_ADD_BINARY_CHAR (c, dst);
4550 label_continue_loop:;
4553 if (flags & CODING_STATE_END)
4555 COMPOSE_FLUSH_CHARS (str, dst);
4556 decode_flush_er_chars (str, dst);
4559 decode_output_utf8_partial_char (counter, cpos, dst);
4566 str->counter = counter;
4570 char_encode_utf8 (struct encoding_stream *str, Emchar ch,
4571 unsigned_char_dynarr *dst, unsigned int *flags)
4573 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
4577 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
4578 Dynarr_add (dst, '\r');
4579 if (eol_type != EOL_CR)
4580 Dynarr_add (dst, ch);
4582 else if (ch <= 0x7f)
4584 Dynarr_add (dst, ch);
4589 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, 0);
4590 int code_point = charset_code_point (ucs_ccs, ch, 0);
4592 if ( (code_point < 0) || (code_point > 0x10FFFF) )
4595 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, 1);
4599 && INTP (ret = Fget_char_attribute (make_char (ch),
4601 code_point = XINT (ret);
4602 else if ( !NILP (map =
4603 CODING_SYSTEM_ISO2022_INITIAL_CHARSET
4605 && INTP (ret = Fget_char_attribute (make_char (ch),
4607 code_point = XINT (ret);
4608 else if (CODING_SYSTEM_USE_ENTITY_REFERENCE (str->codesys))
4612 char_encode_as_entity_reference (ch, buf);
4613 Dynarr_add_many (dst, buf, strlen (buf));
4619 if (code_point <= 0x7ff)
4621 Dynarr_add (dst, (code_point >> 6) | 0xc0);
4622 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4624 else if (code_point <= 0xffff)
4626 Dynarr_add (dst, (code_point >> 12) | 0xe0);
4627 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4628 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4630 else if (code_point <= 0x1fffff)
4632 Dynarr_add (dst, (code_point >> 18) | 0xf0);
4633 Dynarr_add (dst, ((code_point >> 12) & 0x3f) | 0x80);
4634 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4635 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4637 else if (code_point <= 0x3ffffff)
4639 Dynarr_add (dst, (code_point >> 24) | 0xf8);
4640 Dynarr_add (dst, ((code_point >> 18) & 0x3f) | 0x80);
4641 Dynarr_add (dst, ((code_point >> 12) & 0x3f) | 0x80);
4642 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4643 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4647 Dynarr_add (dst, (code_point >> 30) | 0xfc);
4648 Dynarr_add (dst, ((code_point >> 24) & 0x3f) | 0x80);
4649 Dynarr_add (dst, ((code_point >> 18) & 0x3f) | 0x80);
4650 Dynarr_add (dst, ((code_point >> 12) & 0x3f) | 0x80);
4651 Dynarr_add (dst, ((code_point >> 6) & 0x3f) | 0x80);
4652 Dynarr_add (dst, (code_point & 0x3f) | 0x80);
4658 char_finish_utf8 (struct encoding_stream *str, unsigned_char_dynarr *dst,
4659 unsigned int *flags)
4664 /************************************************************************/
4665 /* ISO2022 methods */
4666 /************************************************************************/
4668 /* The following note describes the coding system ISO2022 briefly.
4669 Since the intention of this note is to help understand the
4670 functions in this file, some parts are NOT ACCURATE or OVERLY
4671 SIMPLIFIED. For thorough understanding, please refer to the
4672 original document of ISO2022.
4674 ISO2022 provides many mechanisms to encode several character sets
4675 in 7-bit and 8-bit environments. For 7-bit environments, all text
4676 is encoded using bytes less than 128. This may make the encoded
4677 text a little bit longer, but the text passes more easily through
4678 several gateways, some of which strip off MSB (Most Signigant Bit).
4680 There are two kinds of character sets: control character set and
4681 graphic character set. The former contains control characters such
4682 as `newline' and `escape' to provide control functions (control
4683 functions are also provided by escape sequences). The latter
4684 contains graphic characters such as 'A' and '-'. Emacs recognizes
4685 two control character sets and many graphic character sets.
4687 Graphic character sets are classified into one of the following
4688 four classes, according to the number of bytes (DIMENSION) and
4689 number of characters in one dimension (CHARS) of the set:
4690 - DIMENSION1_CHARS94
4691 - DIMENSION1_CHARS96
4692 - DIMENSION2_CHARS94
4693 - DIMENSION2_CHARS96
4695 In addition, each character set is assigned an identification tag,
4696 unique for each set, called "final character" (denoted as <F>
4697 hereafter). The <F> of each character set is decided by ECMA(*)
4698 when it is registered in ISO. The code range of <F> is 0x30..0x7F
4699 (0x30..0x3F are for private use only).
4701 Note (*): ECMA = European Computer Manufacturers Association
4703 Here are examples of graphic character set [NAME(<F>)]:
4704 o DIMENSION1_CHARS94 -- ASCII('B'), right-half-of-JISX0201('I'), ...
4705 o DIMENSION1_CHARS96 -- right-half-of-ISO8859-1('A'), ...
4706 o DIMENSION2_CHARS94 -- GB2312('A'), JISX0208('B'), ...
4707 o DIMENSION2_CHARS96 -- none for the moment
4709 A code area (1 byte = 8 bits) is divided into 4 areas, C0, GL, C1, and GR.
4710 C0 [0x00..0x1F] -- control character plane 0
4711 GL [0x20..0x7F] -- graphic character plane 0
4712 C1 [0x80..0x9F] -- control character plane 1
4713 GR [0xA0..0xFF] -- graphic character plane 1
4715 A control character set is directly designated and invoked to C0 or
4716 C1 by an escape sequence. The most common case is that:
4717 - ISO646's control character set is designated/invoked to C0, and
4718 - ISO6429's control character set is designated/invoked to C1,
4719 and usually these designations/invocations are omitted in encoded
4720 text. In a 7-bit environment, only C0 can be used, and a control
4721 character for C1 is encoded by an appropriate escape sequence to
4722 fit into the environment. All control characters for C1 are
4723 defined to have corresponding escape sequences.
4725 A graphic character set is at first designated to one of four
4726 graphic registers (G0 through G3), then these graphic registers are
4727 invoked to GL or GR. These designations and invocations can be
4728 done independently. The most common case is that G0 is invoked to
4729 GL, G1 is invoked to GR, and ASCII is designated to G0. Usually
4730 these invocations and designations are omitted in encoded text.
4731 In a 7-bit environment, only GL can be used.
4733 When a graphic character set of CHARS94 is invoked to GL, codes
4734 0x20 and 0x7F of the GL area work as control characters SPACE and
4735 DEL respectively, and codes 0xA0 and 0xFF of the GR area should not
4738 There are two ways of invocation: locking-shift and single-shift.
4739 With locking-shift, the invocation lasts until the next different
4740 invocation, whereas with single-shift, the invocation affects the
4741 following character only and doesn't affect the locking-shift
4742 state. Invocations are done by the following control characters or
4745 ----------------------------------------------------------------------
4746 abbrev function cntrl escape seq description
4747 ----------------------------------------------------------------------
4748 SI/LS0 (shift-in) 0x0F none invoke G0 into GL
4749 SO/LS1 (shift-out) 0x0E none invoke G1 into GL
4750 LS2 (locking-shift-2) none ESC 'n' invoke G2 into GL
4751 LS3 (locking-shift-3) none ESC 'o' invoke G3 into GL
4752 LS1R (locking-shift-1 right) none ESC '~' invoke G1 into GR (*)
4753 LS2R (locking-shift-2 right) none ESC '}' invoke G2 into GR (*)
4754 LS3R (locking-shift 3 right) none ESC '|' invoke G3 into GR (*)
4755 SS2 (single-shift-2) 0x8E ESC 'N' invoke G2 for one char
4756 SS3 (single-shift-3) 0x8F ESC 'O' invoke G3 for one char
4757 ----------------------------------------------------------------------
4758 (*) These are not used by any known coding system.
4760 Control characters for these functions are defined by macros
4761 ISO_CODE_XXX in `coding.h'.
4763 Designations are done by the following escape sequences:
4764 ----------------------------------------------------------------------
4765 escape sequence description
4766 ----------------------------------------------------------------------
4767 ESC '(' <F> designate DIMENSION1_CHARS94<F> to G0
4768 ESC ')' <F> designate DIMENSION1_CHARS94<F> to G1
4769 ESC '*' <F> designate DIMENSION1_CHARS94<F> to G2
4770 ESC '+' <F> designate DIMENSION1_CHARS94<F> to G3
4771 ESC ',' <F> designate DIMENSION1_CHARS96<F> to G0 (*)
4772 ESC '-' <F> designate DIMENSION1_CHARS96<F> to G1
4773 ESC '.' <F> designate DIMENSION1_CHARS96<F> to G2
4774 ESC '/' <F> designate DIMENSION1_CHARS96<F> to G3
4775 ESC '$' '(' <F> designate DIMENSION2_CHARS94<F> to G0 (**)
4776 ESC '$' ')' <F> designate DIMENSION2_CHARS94<F> to G1
4777 ESC '$' '*' <F> designate DIMENSION2_CHARS94<F> to G2
4778 ESC '$' '+' <F> designate DIMENSION2_CHARS94<F> to G3
4779 ESC '$' ',' <F> designate DIMENSION2_CHARS96<F> to G0 (*)
4780 ESC '$' '-' <F> designate DIMENSION2_CHARS96<F> to G1
4781 ESC '$' '.' <F> designate DIMENSION2_CHARS96<F> to G2
4782 ESC '$' '/' <F> designate DIMENSION2_CHARS96<F> to G3
4783 ----------------------------------------------------------------------
4785 In this list, "DIMENSION1_CHARS94<F>" means a graphic character set
4786 of dimension 1, chars 94, and final character <F>, etc...
4788 Note (*): Although these designations are not allowed in ISO2022,
4789 Emacs accepts them on decoding, and produces them on encoding
4790 CHARS96 character sets in a coding system which is characterized as
4791 7-bit environment, non-locking-shift, and non-single-shift.
4793 Note (**): If <F> is '@', 'A', or 'B', the intermediate character
4794 '(' can be omitted. We refer to this as "short-form" hereafter.
4796 Now you may notice that there are a lot of ways for encoding the
4797 same multilingual text in ISO2022. Actually, there exist many
4798 coding systems such as Compound Text (used in X11's inter client
4799 communication, ISO-2022-JP (used in Japanese internet), ISO-2022-KR
4800 (used in Korean internet), EUC (Extended UNIX Code, used in Asian
4801 localized platforms), and all of these are variants of ISO2022.
4803 In addition to the above, Emacs handles two more kinds of escape
4804 sequences: ISO6429's direction specification and Emacs' private
4805 sequence for specifying character composition.
4807 ISO6429's direction specification takes the following form:
4808 o CSI ']' -- end of the current direction
4809 o CSI '0' ']' -- end of the current direction
4810 o CSI '1' ']' -- start of left-to-right text
4811 o CSI '2' ']' -- start of right-to-left text
4812 The control character CSI (0x9B: control sequence introducer) is
4813 abbreviated to the escape sequence ESC '[' in a 7-bit environment.
4815 Character composition specification takes the following form:
4816 o ESC '0' -- start character composition
4817 o ESC '1' -- end character composition
4818 Since these are not standard escape sequences of any ISO standard,
4819 their use with these meanings is restricted to Emacs only. */
4822 reset_iso2022 (Lisp_Object coding_system, struct iso2022_decoder *iso)
4826 for (i = 0; i < 4; i++)
4828 if (!NILP (coding_system))
4830 XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, i);
4832 iso->charset[i] = Qt;
4833 iso->invalid_designated[i] = 0;
4835 iso->esc = ISO_ESC_NOTHING;
4836 iso->esc_bytes_index = 0;
4837 iso->register_left = 0;
4838 iso->register_right = 1;
4839 iso->switched_dir_and_no_valid_charset_yet = 0;
4840 iso->invalid_switch_dir = 0;
4841 iso->output_direction_sequence = 0;
4842 iso->output_literally = 0;
4843 #ifdef ENABLE_COMPOSITE_CHARS
4844 if (iso->composite_chars)
4845 Dynarr_reset (iso->composite_chars);
4850 fit_to_be_escape_quoted (unsigned char c)
4867 /* Parse one byte of an ISO2022 escape sequence.
4868 If the result is an invalid escape sequence, return 0 and
4869 do not change anything in STR. Otherwise, if the result is
4870 an incomplete escape sequence, update ISO2022.ESC and
4871 ISO2022.ESC_BYTES and return -1. Otherwise, update
4872 all the state variables (but not ISO2022.ESC_BYTES) and
4875 If CHECK_INVALID_CHARSETS is non-zero, check for designation
4876 or invocation of an invalid character set and treat that as
4877 an unrecognized escape sequence. */
4880 parse_iso2022_esc (Lisp_Object codesys, struct iso2022_decoder *iso,
4881 unsigned char c, unsigned int *flags,
4882 int check_invalid_charsets)
4884 /* (1) If we're at the end of a designation sequence, CS is the
4885 charset being designated and REG is the register to designate
4888 (2) If we're at the end of a locking-shift sequence, REG is
4889 the register to invoke and HALF (0 == left, 1 == right) is
4890 the half to invoke it into.
4892 (3) If we're at the end of a single-shift sequence, REG is
4893 the register to invoke. */
4894 Lisp_Object cs = Qnil;
4897 /* NOTE: This code does goto's all over the fucking place.
4898 The reason for this is that we're basically implementing
4899 a state machine here, and hierarchical languages like C
4900 don't really provide a clean way of doing this. */
4902 if (! (*flags & CODING_STATE_ESCAPE))
4903 /* At beginning of escape sequence; we need to reset our
4904 escape-state variables. */
4905 iso->esc = ISO_ESC_NOTHING;
4907 iso->output_literally = 0;
4908 iso->output_direction_sequence = 0;
4912 case ISO_ESC_NOTHING:
4913 iso->esc_bytes_index = 0;
4916 case ISO_CODE_ESC: /* Start escape sequence */
4917 *flags |= CODING_STATE_ESCAPE;
4921 case ISO_CODE_CSI: /* ISO6429 (specifying directionality) */
4922 *flags |= CODING_STATE_ESCAPE;
4923 iso->esc = ISO_ESC_5_11;
4926 case ISO_CODE_SO: /* locking shift 1 */
4929 case ISO_CODE_SI: /* locking shift 0 */
4933 case ISO_CODE_SS2: /* single shift */
4936 case ISO_CODE_SS3: /* single shift */
4940 default: /* Other control characters */
4947 /**** single shift ****/
4949 case 'N': /* single shift 2 */
4952 case 'O': /* single shift 3 */
4956 /**** locking shift ****/
4958 case '~': /* locking shift 1 right */
4961 case 'n': /* locking shift 2 */
4964 case '}': /* locking shift 2 right */
4967 case 'o': /* locking shift 3 */
4970 case '|': /* locking shift 3 right */
4974 #ifdef ENABLE_COMPOSITE_CHARS
4975 /**** composite ****/
4978 iso->esc = ISO_ESC_START_COMPOSITE;
4979 *flags = (*flags & CODING_STATE_ISO2022_LOCK) |
4980 CODING_STATE_COMPOSITE;
4984 iso->esc = ISO_ESC_END_COMPOSITE;
4985 *flags = (*flags & CODING_STATE_ISO2022_LOCK) &
4986 ~CODING_STATE_COMPOSITE;
4988 #endif /* ENABLE_COMPOSITE_CHARS */
4990 /**** directionality ****/
4993 iso->esc = ISO_ESC_5_11;
4996 /**** designation ****/
4998 case '$': /* multibyte charset prefix */
4999 iso->esc = ISO_ESC_2_4;
5003 if (0x28 <= c && c <= 0x2F)
5005 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_8);
5009 /* This function is called with CODESYS equal to nil when
5010 doing coding-system detection. */
5012 && XCODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
5013 && fit_to_be_escape_quoted (c))
5015 iso->esc = ISO_ESC_LITERAL;
5016 *flags &= CODING_STATE_ISO2022_LOCK;
5026 /**** directionality ****/
5028 case ISO_ESC_5_11: /* ISO6429 direction control */
5031 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
5032 goto directionality;
5034 if (c == '0') iso->esc = ISO_ESC_5_11_0;
5035 else if (c == '1') iso->esc = ISO_ESC_5_11_1;
5036 else if (c == '2') iso->esc = ISO_ESC_5_11_2;
5040 case ISO_ESC_5_11_0:
5043 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
5044 goto directionality;
5048 case ISO_ESC_5_11_1:
5051 *flags = (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
5052 goto directionality;
5056 case ISO_ESC_5_11_2:
5059 *flags = (*flags & CODING_STATE_ISO2022_LOCK) | CODING_STATE_R2L;
5060 goto directionality;
5065 iso->esc = ISO_ESC_DIRECTIONALITY;
5066 /* Various junk here to attempt to preserve the direction sequences
5067 literally in the text if they would otherwise be swallowed due
5068 to invalid designations that don't show up as actual charset
5069 changes in the text. */
5070 if (iso->invalid_switch_dir)
5072 /* We already inserted a direction switch literally into the
5073 text. We assume (#### this may not be right) that the
5074 next direction switch is the one going the other way,
5075 and we need to output that literally as well. */
5076 iso->output_literally = 1;
5077 iso->invalid_switch_dir = 0;
5083 /* If we are in the thrall of an invalid designation,
5084 then stick the directionality sequence literally into the
5085 output stream so it ends up in the original text again. */
5086 for (jj = 0; jj < 4; jj++)
5087 if (iso->invalid_designated[jj])
5091 iso->output_literally = 1;
5092 iso->invalid_switch_dir = 1;
5095 /* Indicate that we haven't yet seen a valid designation,
5096 so that if a switch-dir is directly followed by an
5097 invalid designation, both get inserted literally. */
5098 iso->switched_dir_and_no_valid_charset_yet = 1;
5103 /**** designation ****/
5106 if (0x28 <= c && c <= 0x2F)
5108 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_4_8);
5111 if (0x40 <= c && c <= 0x42)
5114 cs = CHARSET_BY_ATTRIBUTES (94, -1, c,
5115 *flags & CODING_STATE_R2L ?
5116 CHARSET_RIGHT_TO_LEFT :
5117 CHARSET_LEFT_TO_RIGHT);
5128 if (c < '0' || c > '~')
5129 return 0; /* bad final byte */
5131 if (iso->esc >= ISO_ESC_2_8 &&
5132 iso->esc <= ISO_ESC_2_15)
5134 chars = (iso->esc >= ISO_ESC_2_12) ? 96 : 94;
5135 single = 1; /* single-byte */
5136 reg = (iso->esc - ISO_ESC_2_8) & 3;
5138 else if (iso->esc >= ISO_ESC_2_4_8 &&
5139 iso->esc <= ISO_ESC_2_4_15)
5141 chars = (iso->esc >= ISO_ESC_2_4_12) ? 96 : 94;
5142 single = -1; /* multi-byte */
5143 reg = (iso->esc - ISO_ESC_2_4_8) & 3;
5147 /* Can this ever be reached? -slb */
5151 cs = CHARSET_BY_ATTRIBUTES (chars, single, c,
5152 *flags & CODING_STATE_R2L ?
5153 CHARSET_RIGHT_TO_LEFT :
5154 CHARSET_LEFT_TO_RIGHT);
5160 iso->esc_bytes[iso->esc_bytes_index++] = (unsigned char) c;
5164 if (check_invalid_charsets && !CHARSETP (iso->charset[reg]))
5165 /* can't invoke something that ain't there. */
5167 iso->esc = ISO_ESC_SINGLE_SHIFT;
5168 *flags &= CODING_STATE_ISO2022_LOCK;
5170 *flags |= CODING_STATE_SS2;
5172 *flags |= CODING_STATE_SS3;
5176 if (check_invalid_charsets &&
5177 !CHARSETP (iso->charset[reg]))
5178 /* can't invoke something that ain't there. */
5181 iso->register_right = reg;
5183 iso->register_left = reg;
5184 *flags &= CODING_STATE_ISO2022_LOCK;
5185 iso->esc = ISO_ESC_LOCKING_SHIFT;
5189 if (NILP (cs) && check_invalid_charsets)
5191 iso->invalid_designated[reg] = 1;
5192 iso->charset[reg] = Vcharset_ascii;
5193 iso->esc = ISO_ESC_DESIGNATE;
5194 *flags &= CODING_STATE_ISO2022_LOCK;
5195 iso->output_literally = 1;
5196 if (iso->switched_dir_and_no_valid_charset_yet)
5198 /* We encountered a switch-direction followed by an
5199 invalid designation. Ensure that the switch-direction
5200 gets outputted; otherwise it will probably get eaten
5201 when the text is written out again. */
5202 iso->switched_dir_and_no_valid_charset_yet = 0;
5203 iso->output_direction_sequence = 1;
5204 /* And make sure that the switch-dir going the other
5205 way gets outputted, as well. */
5206 iso->invalid_switch_dir = 1;
5210 /* This function is called with CODESYS equal to nil when
5211 doing coding-system detection. */
5212 if (!NILP (codesys))
5214 charset_conversion_spec_dynarr *dyn =
5215 XCODING_SYSTEM (codesys)->iso2022.input_conv;
5221 for (i = 0; i < Dynarr_length (dyn); i++)
5223 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
5224 if (EQ (cs, spec->from_charset))
5225 cs = spec->to_charset;
5230 iso->charset[reg] = cs;
5231 iso->esc = ISO_ESC_DESIGNATE;
5232 *flags &= CODING_STATE_ISO2022_LOCK;
5233 if (iso->invalid_designated[reg])
5235 iso->invalid_designated[reg] = 0;
5236 iso->output_literally = 1;
5238 if (iso->switched_dir_and_no_valid_charset_yet)
5239 iso->switched_dir_and_no_valid_charset_yet = 0;
5244 detect_coding_iso2022 (struct detection_state *st, const Extbyte *src, Lstream_data_count n)
5248 /* #### There are serious deficiencies in the recognition mechanism
5249 here. This needs to be much smarter if it's going to cut it.
5250 The sequence "\xff\x0f" is currently detected as LOCK_SHIFT while
5251 it should be detected as Latin-1.
5252 All the ISO2022 stuff in this file should be synced up with the
5253 code from FSF Emacs-20.4, in which Mule should be more or less stable.
5254 Perhaps we should wait till R2L works in FSF Emacs? */
5256 if (!st->iso2022.initted)
5258 reset_iso2022 (Qnil, &st->iso2022.iso);
5259 st->iso2022.mask = (CODING_CATEGORY_ISO_7_MASK |
5260 CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
5261 CODING_CATEGORY_ISO_8_1_MASK |
5262 CODING_CATEGORY_ISO_8_2_MASK |
5263 CODING_CATEGORY_ISO_LOCK_SHIFT_MASK);
5264 st->iso2022.flags = 0;
5265 st->iso2022.high_byte_count = 0;
5266 st->iso2022.saw_single_shift = 0;
5267 st->iso2022.initted = 1;
5270 mask = st->iso2022.mask;
5274 unsigned char c = *(unsigned char *)src++;
5277 mask &= ~CODING_CATEGORY_ISO_7_MASK;
5278 st->iso2022.high_byte_count++;
5282 if (st->iso2022.high_byte_count && !st->iso2022.saw_single_shift)
5284 if (st->iso2022.high_byte_count & 1)
5285 /* odd number of high bytes; assume not iso-8-2 */
5286 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
5288 st->iso2022.high_byte_count = 0;
5289 st->iso2022.saw_single_shift = 0;
5291 mask &= ~CODING_CATEGORY_ISO_7_MASK;
5293 if (!(st->iso2022.flags & CODING_STATE_ESCAPE)
5294 && (BYTE_C0_P (c) || BYTE_C1_P (c)))
5295 { /* control chars */
5298 /* Allow and ignore control characters that you might
5299 reasonably see in a text file */
5304 case 8: /* backspace */
5305 case 11: /* vertical tab */
5306 case 12: /* form feed */
5307 case 26: /* MS-DOS C-z junk */
5308 case 31: /* '^_' -- for info */
5309 goto label_continue_loop;
5316 if ((st->iso2022.flags & CODING_STATE_ESCAPE) || BYTE_C0_P (c)
5319 if (parse_iso2022_esc (Qnil, &st->iso2022.iso, c,
5320 &st->iso2022.flags, 0))
5322 switch (st->iso2022.iso.esc)
5324 case ISO_ESC_DESIGNATE:
5325 mask &= ~CODING_CATEGORY_ISO_8_1_MASK;
5326 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
5328 case ISO_ESC_LOCKING_SHIFT:
5329 mask = CODING_CATEGORY_ISO_LOCK_SHIFT_MASK;
5330 goto ran_out_of_chars;
5331 case ISO_ESC_SINGLE_SHIFT:
5332 mask &= ~CODING_CATEGORY_ISO_8_DESIGNATE_MASK;
5333 st->iso2022.saw_single_shift = 1;
5342 goto ran_out_of_chars;
5345 label_continue_loop:;
5354 postprocess_iso2022_mask (int mask)
5356 /* #### kind of cheesy */
5357 /* If seven-bit ISO is allowed, then assume that the encoding is
5358 entirely seven-bit and turn off the eight-bit ones. */
5359 if (mask & CODING_CATEGORY_ISO_7_MASK)
5360 mask &= ~ (CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
5361 CODING_CATEGORY_ISO_8_1_MASK |
5362 CODING_CATEGORY_ISO_8_2_MASK);
5366 /* If FLAGS is a null pointer or specifies right-to-left motion,
5367 output a switch-dir-to-left-to-right sequence to DST.
5368 Also update FLAGS if it is not a null pointer.
5369 If INTERNAL_P is set, we are outputting in internal format and
5370 need to handle the CSI differently. */
5373 restore_left_to_right_direction (Lisp_Coding_System *codesys,
5374 unsigned_char_dynarr *dst,
5375 unsigned int *flags,
5378 if (!flags || (*flags & CODING_STATE_R2L))
5380 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
5382 Dynarr_add (dst, ISO_CODE_ESC);
5383 Dynarr_add (dst, '[');
5385 else if (internal_p)
5386 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
5388 Dynarr_add (dst, ISO_CODE_CSI);
5389 Dynarr_add (dst, '0');
5390 Dynarr_add (dst, ']');
5392 *flags &= ~CODING_STATE_R2L;
5396 /* If FLAGS is a null pointer or specifies a direction different from
5397 DIRECTION (which should be either CHARSET_RIGHT_TO_LEFT or
5398 CHARSET_LEFT_TO_RIGHT), output the appropriate switch-dir escape
5399 sequence to DST. Also update FLAGS if it is not a null pointer.
5400 If INTERNAL_P is set, we are outputting in internal format and
5401 need to handle the CSI differently. */
5404 ensure_correct_direction (int direction, Lisp_Coding_System *codesys,
5405 unsigned_char_dynarr *dst, unsigned int *flags,
5408 if ((!flags || (*flags & CODING_STATE_R2L)) &&
5409 direction == CHARSET_LEFT_TO_RIGHT)
5410 restore_left_to_right_direction (codesys, dst, flags, internal_p);
5411 else if (!CODING_SYSTEM_ISO2022_NO_ISO6429 (codesys)
5412 && (!flags || !(*flags & CODING_STATE_R2L)) &&
5413 direction == CHARSET_RIGHT_TO_LEFT)
5415 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
5417 Dynarr_add (dst, ISO_CODE_ESC);
5418 Dynarr_add (dst, '[');
5420 else if (internal_p)
5421 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
5423 Dynarr_add (dst, ISO_CODE_CSI);
5424 Dynarr_add (dst, '2');
5425 Dynarr_add (dst, ']');
5427 *flags |= CODING_STATE_R2L;
5431 /* Convert ISO2022-format data to internal format. */
5434 decode_coding_iso2022 (Lstream *decoding, const Extbyte *src,
5435 unsigned_char_dynarr *dst, Lstream_data_count n)
5437 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
5438 unsigned int flags = str->flags;
5439 unsigned int cpos = str->cpos;
5440 unsigned char counter = str->counter;
5441 eol_type_t eol_type = str->eol_type;
5442 #ifdef ENABLE_COMPOSITE_CHARS
5443 unsigned_char_dynarr *real_dst = dst;
5445 Lisp_Object coding_system;
5447 XSETCODING_SYSTEM (coding_system, str->codesys);
5449 #ifdef ENABLE_COMPOSITE_CHARS
5450 if (flags & CODING_STATE_COMPOSITE)
5451 dst = str->iso2022.composite_chars;
5452 #endif /* ENABLE_COMPOSITE_CHARS */
5456 unsigned char c = *(unsigned char *)src++;
5457 if (flags & CODING_STATE_ESCAPE)
5458 { /* Within ESC sequence */
5459 int retval = parse_iso2022_esc (coding_system, &str->iso2022,
5464 switch (str->iso2022.esc)
5466 #ifdef ENABLE_COMPOSITE_CHARS
5467 case ISO_ESC_START_COMPOSITE:
5468 if (str->iso2022.composite_chars)
5469 Dynarr_reset (str->iso2022.composite_chars);
5471 str->iso2022.composite_chars = Dynarr_new (unsigned_char);
5472 dst = str->iso2022.composite_chars;
5474 case ISO_ESC_END_COMPOSITE:
5476 Bufbyte comstr[MAX_EMCHAR_LEN];
5478 Emchar emch = lookup_composite_char (Dynarr_atp (dst, 0),
5479 Dynarr_length (dst));
5481 len = set_charptr_emchar (comstr, emch);
5482 Dynarr_add_many (dst, comstr, len);
5485 #endif /* ENABLE_COMPOSITE_CHARS */
5487 case ISO_ESC_LITERAL:
5488 COMPOSE_FLUSH_CHARS (str, dst);
5489 decode_flush_er_chars (str, dst);
5490 DECODE_ADD_BINARY_CHAR (c, dst);
5494 /* Everything else handled already */
5499 /* Attempted error recovery. */
5500 if (str->iso2022.output_direction_sequence)
5501 ensure_correct_direction (flags & CODING_STATE_R2L ?
5502 CHARSET_RIGHT_TO_LEFT :
5503 CHARSET_LEFT_TO_RIGHT,
5504 str->codesys, dst, 0, 1);
5505 /* More error recovery. */
5506 if (!retval || str->iso2022.output_literally)
5508 /* Output the (possibly invalid) sequence */
5510 COMPOSE_FLUSH_CHARS (str, dst);
5511 decode_flush_er_chars (str, dst);
5512 for (i = 0; i < str->iso2022.esc_bytes_index; i++)
5513 DECODE_ADD_BINARY_CHAR (str->iso2022.esc_bytes[i], dst);
5514 flags &= CODING_STATE_ISO2022_LOCK;
5516 n++, src--;/* Repeat the loop with the same character. */
5519 /* No sense in reprocessing the final byte of the
5520 escape sequence; it could mess things up anyway.
5522 COMPOSE_FLUSH_CHARS (str, dst);
5523 decode_flush_er_chars (str, dst);
5524 DECODE_ADD_BINARY_CHAR (c, dst);
5530 else if (BYTE_C0_P (c) || BYTE_C1_P (c))
5531 { /* Control characters */
5533 /***** Error-handling *****/
5535 /* If we were in the middle of a character, dump out the
5536 partial character. */
5539 COMPOSE_FLUSH_CHARS (str, dst);
5540 decode_flush_er_chars (str, dst);
5544 DECODE_ADD_BINARY_CHAR
5545 ((unsigned char)(cpos >> (counter * 8)), dst);
5550 /* If we just saw a single-shift character, dump it out.
5551 This may dump out the wrong sort of single-shift character,
5552 but least it will give an indication that something went
5554 if (flags & CODING_STATE_SS2)
5556 COMPOSE_FLUSH_CHARS (str, dst);
5557 decode_flush_er_chars (str, dst);
5558 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS2, dst);
5559 flags &= ~CODING_STATE_SS2;
5561 if (flags & CODING_STATE_SS3)
5563 COMPOSE_FLUSH_CHARS (str, dst);
5564 decode_flush_er_chars (str, dst);
5565 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS3, dst);
5566 flags &= ~CODING_STATE_SS3;
5569 /***** Now handle the control characters. *****/
5575 COMPOSE_FLUSH_CHARS (str, dst);
5576 decode_flush_er_chars (str, dst);
5577 if (eol_type == EOL_CR)
5578 Dynarr_add (dst, '\n');
5579 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR)
5580 Dynarr_add (dst, c);
5582 flags |= CODING_STATE_CR;
5583 goto label_continue_loop;
5585 else if (flags & CODING_STATE_CR)
5586 { /* eol_type == CODING_SYSTEM_EOL_CRLF */
5588 Dynarr_add (dst, '\r');
5589 flags &= ~CODING_STATE_CR;
5592 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
5595 flags &= CODING_STATE_ISO2022_LOCK;
5597 if (!parse_iso2022_esc (coding_system, &str->iso2022, c, &flags, 1))
5599 COMPOSE_FLUSH_CHARS (str, dst);
5600 decode_flush_er_chars (str, dst);
5601 DECODE_ADD_BINARY_CHAR (c, dst);
5605 { /* Graphic characters */
5606 Lisp_Object charset;
5615 COMPOSE_FLUSH_CHARS (str, dst);
5616 decode_flush_er_chars (str, dst);
5617 if (eol_type == EOL_CR)
5618 Dynarr_add (dst, '\n');
5619 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR)
5620 Dynarr_add (dst, c);
5622 flags |= CODING_STATE_CR;
5623 goto label_continue_loop;
5625 else if (flags & CODING_STATE_CR)
5626 { /* eol_type == CODING_SYSTEM_EOL_CRLF */
5628 Dynarr_add (dst, '\r');
5629 flags &= ~CODING_STATE_CR;
5632 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
5635 /* Now determine the charset. */
5636 reg = ((flags & CODING_STATE_SS2) ? 2
5637 : (flags & CODING_STATE_SS3) ? 3
5638 : !BYTE_ASCII_P (c) ? str->iso2022.register_right
5639 : str->iso2022.register_left);
5640 charset = str->iso2022.charset[reg];
5642 /* Error checking: */
5643 if (! CHARSETP (charset)
5644 || str->iso2022.invalid_designated[reg]
5645 || (((c & 0x7F) == ' ' || (c & 0x7F) == ISO_CODE_DEL)
5646 && XCHARSET_CHARS (charset) == 94))
5647 /* Mrmph. We are trying to invoke a register that has no
5648 or an invalid charset in it, or trying to add a character
5649 outside the range of the charset. Insert that char literally
5650 to preserve it for the output. */
5652 COMPOSE_FLUSH_CHARS (str, dst);
5653 decode_flush_er_chars (str, dst);
5657 DECODE_ADD_BINARY_CHAR
5658 ((unsigned char)(cpos >> (counter * 8)), dst);
5661 DECODE_ADD_BINARY_CHAR (c, dst);
5666 /* Things are probably hunky-dorey. */
5668 /* Fetch reverse charset, maybe. */
5669 if (((flags & CODING_STATE_R2L) &&
5670 XCHARSET_DIRECTION (charset) == CHARSET_LEFT_TO_RIGHT)
5672 (!(flags & CODING_STATE_R2L) &&
5673 XCHARSET_DIRECTION (charset) == CHARSET_RIGHT_TO_LEFT))
5675 Lisp_Object new_charset =
5676 XCHARSET_REVERSE_DIRECTION_CHARSET (charset);
5677 if (!NILP (new_charset))
5678 charset = new_charset;
5683 if (XCHARSET_DIMENSION (charset) == counter)
5685 COMPOSE_ADD_CHAR (str,
5686 DECODE_CHAR (charset,
5687 ((cpos & 0x7F7F7F) << 8)
5694 cpos = (cpos << 8) | c;
5696 lb = XCHARSET_LEADING_BYTE (charset);
5697 switch (XCHARSET_REP_BYTES (charset))
5700 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5701 Dynarr_add (dst, c & 0x7F);
5704 case 2: /* one-byte official */
5705 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5706 Dynarr_add (dst, lb);
5707 Dynarr_add (dst, c | 0x80);
5710 case 3: /* one-byte private or two-byte official */
5711 if (XCHARSET_PRIVATE_P (charset))
5713 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5714 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_1);
5715 Dynarr_add (dst, lb);
5716 Dynarr_add (dst, c | 0x80);
5722 Dynarr_add (dst, lb);
5723 Dynarr_add (dst, ch | 0x80);
5724 Dynarr_add (dst, c | 0x80);
5732 default: /* two-byte private */
5735 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_2);
5736 Dynarr_add (dst, lb);
5737 Dynarr_add (dst, ch | 0x80);
5738 Dynarr_add (dst, c | 0x80);
5748 flags &= CODING_STATE_ISO2022_LOCK;
5751 label_continue_loop:;
5754 if (flags & CODING_STATE_END)
5756 COMPOSE_FLUSH_CHARS (str, dst);
5757 decode_flush_er_chars (str, dst);
5758 DECODE_OUTPUT_PARTIAL_CHAR (cpos);
5762 str->counter = counter;
5766 /***** ISO2022 encoder *****/
5768 /* Designate CHARSET into register REG. */
5771 iso2022_designate (Lisp_Object charset, unsigned char reg,
5772 struct encoding_stream *str, unsigned_char_dynarr *dst)
5774 static const char inter94[] = "()*+";
5775 static const char inter96[] = ",-./";
5776 unsigned short chars;
5777 unsigned char dimension;
5778 unsigned char final;
5779 Lisp_Object old_charset = str->iso2022.charset[reg];
5781 str->iso2022.charset[reg] = charset;
5782 if (!CHARSETP (charset))
5783 /* charset might be an initial nil or t. */
5785 chars = XCHARSET_CHARS (charset);
5786 dimension = XCHARSET_DIMENSION (charset);
5787 final = XCHARSET_FINAL (charset);
5788 if (!str->iso2022.force_charset_on_output[reg] &&
5789 CHARSETP (old_charset) &&
5790 XCHARSET_CHARS (old_charset) == chars &&
5791 XCHARSET_DIMENSION (old_charset) == dimension &&
5792 XCHARSET_FINAL (old_charset) == final)
5795 str->iso2022.force_charset_on_output[reg] = 0;
5798 charset_conversion_spec_dynarr *dyn =
5799 str->codesys->iso2022.output_conv;
5805 for (i = 0; i < Dynarr_length (dyn); i++)
5807 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
5808 if (EQ (charset, spec->from_charset))
5809 charset = spec->to_charset;
5814 Dynarr_add (dst, ISO_CODE_ESC);
5819 Dynarr_add (dst, inter94[reg]);
5822 Dynarr_add (dst, '$');
5824 || !(CODING_SYSTEM_ISO2022_SHORT (str->codesys))
5827 Dynarr_add (dst, inter94[reg]);
5832 Dynarr_add (dst, inter96[reg]);
5835 Dynarr_add (dst, '$');
5836 Dynarr_add (dst, inter96[reg]);
5840 Dynarr_add (dst, final);
5844 ensure_normal_shift (struct encoding_stream *str, unsigned_char_dynarr *dst)
5846 if (str->iso2022.register_left != 0)
5848 Dynarr_add (dst, ISO_CODE_SI);
5849 str->iso2022.register_left = 0;
5854 ensure_shift_out (struct encoding_stream *str, unsigned_char_dynarr *dst)
5856 if (str->iso2022.register_left != 1)
5858 Dynarr_add (dst, ISO_CODE_SO);
5859 str->iso2022.register_left = 1;
5864 char_encode_iso2022 (struct encoding_stream *str, Emchar ch,
5865 unsigned_char_dynarr *dst, unsigned int *flags)
5867 unsigned char charmask;
5868 Lisp_Coding_System* codesys = str->codesys;
5869 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
5871 Lisp_Object charset = str->iso2022.current_charset;
5872 int half = str->iso2022.current_half;
5873 int code_point = -1;
5877 restore_left_to_right_direction (codesys, dst, flags, 0);
5879 /* Make sure G0 contains ASCII */
5880 if ((ch > ' ' && ch < ISO_CODE_DEL)
5881 || !CODING_SYSTEM_ISO2022_NO_ASCII_CNTL (codesys))
5883 ensure_normal_shift (str, dst);
5884 iso2022_designate (Vcharset_ascii, 0, str, dst);
5887 /* If necessary, restore everything to the default state
5889 if (ch == '\n' && !(CODING_SYSTEM_ISO2022_NO_ASCII_EOL (codesys)))
5891 restore_left_to_right_direction (codesys, dst, flags, 0);
5893 ensure_normal_shift (str, dst);
5895 for (i = 0; i < 4; i++)
5897 Lisp_Object initial_charset =
5898 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
5899 iso2022_designate (initial_charset, i, str, dst);
5904 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5905 Dynarr_add (dst, '\r');
5906 if (eol_type != EOL_CR)
5907 Dynarr_add (dst, ch);
5911 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
5912 && fit_to_be_escape_quoted (ch))
5913 Dynarr_add (dst, ISO_CODE_ESC);
5914 Dynarr_add (dst, ch);
5917 else if ( (0x80 <= ch) && (ch <= 0x9f) )
5919 charmask = (half == 0 ? 0x00 : 0x80);
5921 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
5922 && fit_to_be_escape_quoted (ch))
5923 Dynarr_add (dst, ISO_CODE_ESC);
5924 /* you asked for it ... */
5925 Dynarr_add (dst, ch);
5931 /* Now determine which register to use. */
5933 for (i = 0; i < 4; i++)
5935 if ((CHARSETP (charset = str->iso2022.charset[i])
5936 && ((code_point = charset_code_point (charset, ch, 0)) >= 0))
5940 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i))
5941 && ((code_point = charset_code_point (charset, ch, 0)) >= 0)))
5949 Lisp_Object original_default_coded_charset_priority_list
5950 = Vdefault_coded_charset_priority_list;
5952 while (!EQ (Vdefault_coded_charset_priority_list, Qnil))
5954 code_point = ENCODE_CHAR (ch, charset);
5955 if (XCHARSET_FINAL (charset))
5957 Vdefault_coded_charset_priority_list
5958 = Fcdr (Fmemq (XCHARSET_NAME (charset),
5959 Vdefault_coded_charset_priority_list));
5961 code_point = ENCODE_CHAR (ch, charset);
5962 if (!XCHARSET_FINAL (charset))
5964 charset = Vcharset_ascii;
5968 Vdefault_coded_charset_priority_list
5969 = original_default_coded_charset_priority_list;
5971 ensure_correct_direction (XCHARSET_DIRECTION (charset),
5972 codesys, dst, flags, 0);
5976 if (XCHARSET_GRAPHIC (charset) != 0)
5978 if (!NILP (str->iso2022.charset[1]) &&
5979 (!CODING_SYSTEM_ISO2022_SEVEN (codesys)
5980 || CODING_SYSTEM_ISO2022_LOCK_SHIFT (codesys)))
5982 else if (!NILP (str->iso2022.charset[2]))
5984 else if (!NILP (str->iso2022.charset[3]))
5993 iso2022_designate (charset, reg, str, dst);
5995 /* Now invoke that register. */
5999 ensure_normal_shift (str, dst);
6003 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
6005 ensure_shift_out (str, dst);
6012 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
6014 Dynarr_add (dst, ISO_CODE_ESC);
6015 Dynarr_add (dst, 'N');
6020 Dynarr_add (dst, ISO_CODE_SS2);
6025 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
6027 Dynarr_add (dst, ISO_CODE_ESC);
6028 Dynarr_add (dst, 'O');
6033 Dynarr_add (dst, ISO_CODE_SS3);
6041 charmask = (half == 0 ? 0x00 : 0x80);
6043 switch (XCHARSET_DIMENSION (charset))
6046 Dynarr_add (dst, (code_point & 0xFF) | charmask);
6049 Dynarr_add (dst, ((code_point >> 8) & 0xFF) | charmask);
6050 Dynarr_add (dst, ( code_point & 0xFF) | charmask);
6053 Dynarr_add (dst, ((code_point >> 16) & 0xFF) | charmask);
6054 Dynarr_add (dst, ((code_point >> 8) & 0xFF) | charmask);
6055 Dynarr_add (dst, ( code_point & 0xFF) | charmask);
6058 Dynarr_add (dst, ((code_point >> 24) & 0xFF) | charmask);
6059 Dynarr_add (dst, ((code_point >> 16) & 0xFF) | charmask);
6060 Dynarr_add (dst, ((code_point >> 8) & 0xFF) | charmask);
6061 Dynarr_add (dst, ( code_point & 0xFF) | charmask);
6067 str->iso2022.current_charset = charset;
6068 str->iso2022.current_half = half;
6072 char_finish_iso2022 (struct encoding_stream *str, unsigned_char_dynarr *dst,
6073 unsigned int *flags)
6075 Lisp_Coding_System* codesys = str->codesys;
6078 restore_left_to_right_direction (codesys, dst, flags, 0);
6079 ensure_normal_shift (str, dst);
6080 for (i = 0; i < 4; i++)
6082 Lisp_Object initial_charset
6083 = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
6084 iso2022_designate (initial_charset, i, str, dst);
6089 /************************************************************************/
6090 /* No-conversion methods */
6091 /************************************************************************/
6093 /* This is used when reading in "binary" files -- i.e. files that may
6094 contain all 256 possible byte values and that are not to be
6095 interpreted as being in any particular decoding. */
6097 decode_coding_no_conversion (Lstream *decoding, const Extbyte *src,
6098 unsigned_char_dynarr *dst, Lstream_data_count n)
6100 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
6101 unsigned int flags = str->flags;
6102 unsigned int cpos = str->cpos;
6103 eol_type_t eol_type = str->eol_type;
6107 unsigned char c = *(unsigned char *)src++;
6109 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
6110 DECODE_ADD_BINARY_CHAR (c, dst);
6111 label_continue_loop:;
6114 DECODE_HANDLE_END_OF_CONVERSION (flags, cpos, dst);
6121 encode_coding_no_conversion (Lstream *encoding, const Bufbyte *src,
6122 unsigned_char_dynarr *dst, Lstream_data_count n)
6125 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
6126 unsigned int flags = str->flags;
6127 unsigned int ch = str->ch;
6128 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
6130 unsigned char char_boundary = str->iso2022.current_char_boundary;
6137 if (char_boundary == 0)
6143 else if ( c >= 0xf8 )
6148 else if ( c >= 0xf0 )
6153 else if ( c >= 0xe0 )
6158 else if ( c >= 0xc0 )
6168 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
6169 Dynarr_add (dst, '\r');
6170 if (eol_type != EOL_CR)
6171 Dynarr_add (dst, c);
6174 Dynarr_add (dst, c);
6177 else if (char_boundary == 1)
6179 ch = ( ch << 6 ) | ( c & 0x3f );
6180 Dynarr_add (dst, ch & 0xff);
6185 ch = ( ch << 6 ) | ( c & 0x3f );
6188 #else /* not UTF2000 */
6191 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
6192 Dynarr_add (dst, '\r');
6193 if (eol_type != EOL_CR)
6194 Dynarr_add (dst, '\n');
6197 else if (BYTE_ASCII_P (c))
6200 Dynarr_add (dst, c);
6202 else if (BUFBYTE_LEADING_BYTE_P (c))
6205 if (c == LEADING_BYTE_LATIN_ISO8859_1 ||
6206 c == LEADING_BYTE_CONTROL_1)
6209 Dynarr_add (dst, '~'); /* untranslatable character */
6213 if (ch == LEADING_BYTE_LATIN_ISO8859_1)
6214 Dynarr_add (dst, c);
6215 else if (ch == LEADING_BYTE_CONTROL_1)
6218 Dynarr_add (dst, c - 0x20);
6220 /* else it should be the second or third byte of an
6221 untranslatable character, so ignore it */
6224 #endif /* not UTF2000 */
6230 str->iso2022.current_char_boundary = char_boundary;
6236 /************************************************************************/
6237 /* Initialization */
6238 /************************************************************************/
6241 syms_of_file_coding (void)
6243 INIT_LRECORD_IMPLEMENTATION (coding_system);
6245 DEFERROR_STANDARD (Qcoding_system_error, Qio_error);
6247 DEFSUBR (Fcoding_system_p);
6248 DEFSUBR (Ffind_coding_system);
6249 DEFSUBR (Fget_coding_system);
6250 DEFSUBR (Fcoding_system_list);
6251 DEFSUBR (Fcoding_system_name);
6252 DEFSUBR (Fmake_coding_system);
6253 DEFSUBR (Fcopy_coding_system);
6254 DEFSUBR (Fcoding_system_canonical_name_p);
6255 DEFSUBR (Fcoding_system_alias_p);
6256 DEFSUBR (Fcoding_system_aliasee);
6257 DEFSUBR (Fdefine_coding_system_alias);
6258 DEFSUBR (Fsubsidiary_coding_system);
6260 DEFSUBR (Fcoding_system_type);
6261 DEFSUBR (Fcoding_system_doc_string);
6263 DEFSUBR (Fcoding_system_charset);
6265 DEFSUBR (Fcoding_system_property);
6267 DEFSUBR (Fcoding_category_list);
6268 DEFSUBR (Fset_coding_priority_list);
6269 DEFSUBR (Fcoding_priority_list);
6270 DEFSUBR (Fset_coding_category_system);
6271 DEFSUBR (Fcoding_category_system);
6273 DEFSUBR (Fdetect_coding_region);
6274 DEFSUBR (Fdecode_coding_region);
6275 DEFSUBR (Fencode_coding_region);
6277 DEFSUBR (Fdecode_shift_jis_char);
6278 DEFSUBR (Fencode_shift_jis_char);
6279 DEFSUBR (Fdecode_big5_char);
6280 DEFSUBR (Fencode_big5_char);
6282 defsymbol (&Qcoding_systemp, "coding-system-p");
6283 defsymbol (&Qno_conversion, "no-conversion");
6284 defsymbol (&Qraw_text, "raw-text");
6286 defsymbol (&Qbig5, "big5");
6287 defsymbol (&Qshift_jis, "shift-jis");
6288 defsymbol (&Qucs4, "ucs-4");
6289 defsymbol (&Qutf8, "utf-8");
6290 defsymbol (&Qutf16, "utf-16");
6291 defsymbol (&Qccl, "ccl");
6292 defsymbol (&Qiso2022, "iso2022");
6294 defsymbol (&Qmnemonic, "mnemonic");
6295 defsymbol (&Qeol_type, "eol-type");
6296 defsymbol (&Qpost_read_conversion, "post-read-conversion");
6297 defsymbol (&Qpre_write_conversion, "pre-write-conversion");
6299 defsymbol (&Qcr, "cr");
6300 defsymbol (&Qlf, "lf");
6301 defsymbol (&Qcrlf, "crlf");
6302 defsymbol (&Qeol_cr, "eol-cr");
6303 defsymbol (&Qeol_lf, "eol-lf");
6304 defsymbol (&Qeol_crlf, "eol-crlf");
6306 defsymbol (&Qcharset_g0, "charset-g0");
6307 defsymbol (&Qcharset_g1, "charset-g1");
6308 defsymbol (&Qcharset_g2, "charset-g2");
6309 defsymbol (&Qcharset_g3, "charset-g3");
6310 defsymbol (&Qforce_g0_on_output, "force-g0-on-output");
6311 defsymbol (&Qforce_g1_on_output, "force-g1-on-output");
6312 defsymbol (&Qforce_g2_on_output, "force-g2-on-output");
6313 defsymbol (&Qforce_g3_on_output, "force-g3-on-output");
6314 defsymbol (&Qno_iso6429, "no-iso6429");
6315 defsymbol (&Qinput_charset_conversion, "input-charset-conversion");
6316 defsymbol (&Qoutput_charset_conversion, "output-charset-conversion");
6318 defsymbol (&Qshort, "short");
6319 defsymbol (&Qno_ascii_eol, "no-ascii-eol");
6320 defsymbol (&Qno_ascii_cntl, "no-ascii-cntl");
6321 defsymbol (&Qseven, "seven");
6322 defsymbol (&Qlock_shift, "lock-shift");
6323 defsymbol (&Qescape_quoted, "escape-quoted");
6326 defsymbol (&Qutf_8_mcs, "utf-8-mcs");
6327 defsymbol (&Qdisable_composition, "disable-composition");
6328 defsymbol (&Quse_entity_reference, "use-entity-reference");
6329 defsymbol (&Qd, "d");
6330 defsymbol (&Qx, "x");
6331 defsymbol (&QX, "X");
6333 defsymbol (&Qencode, "encode");
6334 defsymbol (&Qdecode, "decode");
6337 defsymbol (&coding_category_symbol[CODING_CATEGORY_SHIFT_JIS],
6339 defsymbol (&coding_category_symbol[CODING_CATEGORY_BIG5],
6341 defsymbol (&coding_category_symbol[CODING_CATEGORY_UCS4],
6343 defsymbol (&coding_category_symbol[CODING_CATEGORY_UTF16],
6345 defsymbol (&coding_category_symbol[CODING_CATEGORY_UTF8],
6347 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_7],
6349 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_DESIGNATE],
6351 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_1],
6353 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_2],
6355 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_LOCK_SHIFT],
6358 defsymbol (&coding_category_symbol[CODING_CATEGORY_NO_CONVERSION],
6363 lstream_type_create_file_coding (void)
6365 LSTREAM_HAS_METHOD (decoding, reader);
6366 LSTREAM_HAS_METHOD (decoding, writer);
6367 LSTREAM_HAS_METHOD (decoding, rewinder);
6368 LSTREAM_HAS_METHOD (decoding, seekable_p);
6369 LSTREAM_HAS_METHOD (decoding, flusher);
6370 LSTREAM_HAS_METHOD (decoding, closer);
6371 LSTREAM_HAS_METHOD (decoding, marker);
6373 LSTREAM_HAS_METHOD (encoding, reader);
6374 LSTREAM_HAS_METHOD (encoding, writer);
6375 LSTREAM_HAS_METHOD (encoding, rewinder);
6376 LSTREAM_HAS_METHOD (encoding, seekable_p);
6377 LSTREAM_HAS_METHOD (encoding, flusher);
6378 LSTREAM_HAS_METHOD (encoding, closer);
6379 LSTREAM_HAS_METHOD (encoding, marker);
6383 vars_of_file_coding (void)
6387 fcd = xnew (struct file_coding_dump);
6388 dump_add_root_struct_ptr (&fcd, &fcd_description);
6390 /* Initialize to something reasonable ... */
6391 for (i = 0; i < CODING_CATEGORY_LAST; i++)
6393 fcd->coding_category_system[i] = Qnil;
6394 fcd->coding_category_by_priority[i] = i;
6397 Fprovide (intern ("file-coding"));
6399 DEFVAR_LISP ("keyboard-coding-system", &Vkeyboard_coding_system /*
6400 Coding system used for TTY keyboard input.
6401 Not used under a windowing system.
6403 Vkeyboard_coding_system = Qnil;
6405 DEFVAR_LISP ("terminal-coding-system", &Vterminal_coding_system /*
6406 Coding system used for TTY display output.
6407 Not used under a windowing system.
6409 Vterminal_coding_system = Qnil;
6411 DEFVAR_LISP ("coding-system-for-read", &Vcoding_system_for_read /*
6412 Overriding coding system used when reading from a file or process.
6413 You should bind this variable with `let', but do not set it globally.
6414 If this is non-nil, it specifies the coding system that will be used
6415 to decode input on read operations, such as from a file or process.
6416 It overrides `buffer-file-coding-system-for-read',
6417 `insert-file-contents-pre-hook', etc. Use those variables instead of
6418 this one for permanent changes to the environment. */ );
6419 Vcoding_system_for_read = Qnil;
6421 DEFVAR_LISP ("coding-system-for-write",
6422 &Vcoding_system_for_write /*
6423 Overriding coding system used when writing to a file or process.
6424 You should bind this variable with `let', but do not set it globally.
6425 If this is non-nil, it specifies the coding system that will be used
6426 to encode output for write operations, such as to a file or process.
6427 It overrides `buffer-file-coding-system', `write-region-pre-hook', etc.
6428 Use those variables instead of this one for permanent changes to the
6430 Vcoding_system_for_write = Qnil;
6432 DEFVAR_LISP ("file-name-coding-system", &Vfile_name_coding_system /*
6433 Coding system used to convert pathnames when accessing files.
6435 Vfile_name_coding_system = Qnil;
6437 DEFVAR_LISP ("coded-charset-entity-reference-alist",
6438 &Vcoded_charset_entity_reference_alist /*
6439 Alist of coded-charset vs corresponding entity-reference.
6440 Each element looks like (CCS PREFIX CODE-COLUMNS CODE-TYPE).
6441 CCS is coded-charset.
6442 CODE-COLUMNS is columns of code-point of entity-reference.
6443 CODE-TYPE is format type of code-point of entity-reference.
6444 `d' means decimal value and `x' means hexadecimal value.
6446 Vcoded_charset_entity_reference_alist = Qnil;
6448 DEFVAR_BOOL ("enable-multibyte-characters", &enable_multibyte_characters /*
6449 Non-nil means the buffer contents are regarded as multi-byte form
6450 of characters, not a binary code. This affects the display, file I/O,
6451 and behaviors of various editing commands.
6453 Setting this to nil does not do anything.
6455 enable_multibyte_characters = 1;
6459 complex_vars_of_file_coding (void)
6461 staticpro (&Vcoding_system_hash_table);
6462 Vcoding_system_hash_table =
6463 make_lisp_hash_table (50, HASH_TABLE_NON_WEAK, HASH_TABLE_EQ);
6465 the_codesys_prop_dynarr = Dynarr_new (codesys_prop);
6466 dump_add_root_struct_ptr (&the_codesys_prop_dynarr, &codesys_prop_dynarr_description);
6468 #define DEFINE_CODESYS_PROP(Prop_Type, Sym) do \
6470 struct codesys_prop csp; \
6472 csp.prop_type = (Prop_Type); \
6473 Dynarr_add (the_codesys_prop_dynarr, csp); \
6476 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qmnemonic);
6477 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_type);
6478 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_cr);
6479 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_crlf);
6480 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_lf);
6481 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpost_read_conversion);
6482 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpre_write_conversion);
6484 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g0);
6485 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g1);
6486 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g2);
6487 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g3);
6488 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g0_on_output);
6489 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g1_on_output);
6490 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g2_on_output);
6491 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g3_on_output);
6492 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qshort);
6493 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_eol);
6494 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_cntl);
6495 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qseven);
6496 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qlock_shift);
6497 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_iso6429);
6498 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qescape_quoted);
6499 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qinput_charset_conversion);
6500 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qoutput_charset_conversion);
6502 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qencode);
6503 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qdecode);
6505 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qdisable_composition);
6506 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Quse_entity_reference);
6509 /* Need to create this here or we're really screwed. */
6511 (Qraw_text, Qno_conversion,
6512 build_string ("Raw text, which means it converts only line-break-codes."),
6513 list2 (Qmnemonic, build_string ("Raw")));
6516 (Qbinary, Qno_conversion,
6517 build_string ("Binary, which means it does not convert anything."),
6518 list4 (Qeol_type, Qlf,
6519 Qmnemonic, build_string ("Binary")));
6525 ("Coding-system of UTF-8 with Multiple Coded-character-Sets extension."),
6526 list2 (Qmnemonic, build_string ("MTF8")));
6529 Fdefine_coding_system_alias (Qno_conversion, Qraw_text);
6531 Fdefine_coding_system_alias (Qfile_name, Qbinary);
6533 Fdefine_coding_system_alias (Qterminal, Qbinary);
6534 Fdefine_coding_system_alias (Qkeyboard, Qbinary);
6536 /* Need this for bootstrapping */
6537 fcd->coding_category_system[CODING_CATEGORY_NO_CONVERSION] =
6538 Fget_coding_system (Qraw_text);
6541 fcd->coding_category_system[CODING_CATEGORY_UTF8]
6542 = Fget_coding_system (Qutf_8_mcs);
6545 #if defined(MULE) && !defined(UTF2000)
6549 for (i = 0; i < countof (fcd->ucs_to_mule_table); i++)
6550 fcd->ucs_to_mule_table[i] = Qnil;
6552 staticpro (&mule_to_ucs_table);
6553 mule_to_ucs_table = Fmake_char_table(Qgeneric);
6554 #endif /* defined(MULE) && !defined(UTF2000) */