1 /* Code conversion functions.
2 Copyright (C) 1991, 1995 Free Software Foundation, Inc.
3 Copyright (C) 1995 Sun Microsystems, Inc.
5 This file is part of XEmacs.
7 XEmacs is free software; you can redistribute it and/or modify it
8 under the terms of the GNU General Public License as published by the
9 Free Software Foundation; either version 2, or (at your option) any
12 XEmacs is distributed in the hope that it will be useful, but WITHOUT
13 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with XEmacs; see the file COPYING. If not, write to
19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 Boston, MA 02111-1307, USA. */
22 /* Synched up with: Mule 2.3. Not in FSF. */
24 /* Rewritten by Ben Wing <ben@xemacs.org>. */
38 #include "file-coding.h"
40 Lisp_Object Qcoding_system_error;
42 Lisp_Object Vkeyboard_coding_system;
43 Lisp_Object Vterminal_coding_system;
44 Lisp_Object Vcoding_system_for_read;
45 Lisp_Object Vcoding_system_for_write;
46 Lisp_Object Vfile_name_coding_system;
48 /* Table of symbols identifying each coding category. */
49 Lisp_Object coding_category_symbol[CODING_CATEGORY_LAST + 1];
53 struct file_coding_dump {
54 /* Coding system currently associated with each coding category. */
55 Lisp_Object coding_category_system[CODING_CATEGORY_LAST + 1];
57 /* Table of all coding categories in decreasing order of priority.
58 This describes a permutation of the possible coding categories. */
59 int coding_category_by_priority[CODING_CATEGORY_LAST + 1];
62 Lisp_Object ucs_to_mule_table[65536];
66 static const struct lrecord_description fcd_description_1[] = {
67 { XD_LISP_OBJECT_ARRAY, offsetof (struct file_coding_dump, coding_category_system), CODING_CATEGORY_LAST + 1 },
69 { XD_LISP_OBJECT_ARRAY, offsetof (struct file_coding_dump, ucs_to_mule_table), 65536 },
74 static const struct struct_description fcd_description = {
75 sizeof (struct file_coding_dump),
79 Lisp_Object mule_to_ucs_table;
81 Lisp_Object Qcoding_systemp;
83 Lisp_Object Qraw_text, Qno_conversion, Qccl, Qiso2022;
84 /* Qinternal in general.c */
86 Lisp_Object Qmnemonic, Qeol_type;
87 Lisp_Object Qcr, Qcrlf, Qlf;
88 Lisp_Object Qeol_cr, Qeol_crlf, Qeol_lf;
89 Lisp_Object Qpost_read_conversion;
90 Lisp_Object Qpre_write_conversion;
93 Lisp_Object Qucs4, Qutf8;
94 Lisp_Object Qbig5, Qshift_jis;
95 Lisp_Object Qcharset_g0, Qcharset_g1, Qcharset_g2, Qcharset_g3;
96 Lisp_Object Qforce_g0_on_output, Qforce_g1_on_output;
97 Lisp_Object Qforce_g2_on_output, Qforce_g3_on_output;
98 Lisp_Object Qno_iso6429;
99 Lisp_Object Qinput_charset_conversion, Qoutput_charset_conversion;
100 Lisp_Object Qescape_quoted;
101 Lisp_Object Qshort, Qno_ascii_eol, Qno_ascii_cntl, Qseven, Qlock_shift;
103 Lisp_Object Qencode, Qdecode;
105 Lisp_Object Vcoding_system_hash_table;
107 int enable_multibyte_characters;
110 /* Additional information used by the ISO2022 decoder and detector. */
111 struct iso2022_decoder
113 /* CHARSET holds the character sets currently assigned to the G0
114 through G3 variables. It is initialized from the array
115 INITIAL_CHARSET in CODESYS. */
116 Lisp_Object charset[4];
118 /* Which registers are currently invoked into the left (GL) and
119 right (GR) halves of the 8-bit encoding space? */
120 int register_left, register_right;
122 /* ISO_ESC holds a value indicating part of an escape sequence
123 that has already been seen. */
124 enum iso_esc_flag esc;
126 /* This records the bytes we've seen so far in an escape sequence,
127 in case the sequence is invalid (we spit out the bytes unchanged). */
128 unsigned char esc_bytes[8];
130 /* Index for next byte to store in ISO escape sequence. */
133 #ifdef ENABLE_COMPOSITE_CHARS
134 /* Stuff seen so far when composing a string. */
135 unsigned_char_dynarr *composite_chars;
138 /* If we saw an invalid designation sequence for a particular
139 register, we flag it here and switch to ASCII. The next time we
140 see a valid designation for this register, we turn off the flag
141 and do the designation normally, but pretend the sequence was
142 invalid. The effect of all this is that (most of the time) the
143 escape sequences for both the switch to the unknown charset, and
144 the switch back to the known charset, get inserted literally into
145 the buffer and saved out as such. The hope is that we can
146 preserve the escape sequences so that the resulting written out
147 file makes sense. If we don't do any of this, the designation
148 to the invalid charset will be preserved but that switch back
149 to the known charset will probably get eaten because it was
150 the same charset that was already present in the register. */
151 unsigned char invalid_designated[4];
153 /* We try to do similar things as above for direction-switching
154 sequences. If we encountered a direction switch while an
155 invalid designation was present, or an invalid designation
156 just after a direction switch (i.e. no valid designation
157 encountered yet), we insert the direction-switch escape
158 sequence literally into the output stream, and later on
159 insert the corresponding direction-restoring escape sequence
161 unsigned int switched_dir_and_no_valid_charset_yet :1;
162 unsigned int invalid_switch_dir :1;
164 /* Tells the decoder to output the escape sequence literally
165 even though it was valid. Used in the games we play to
166 avoid lossage when we encounter invalid designations. */
167 unsigned int output_literally :1;
168 /* We encountered a direction switch followed by an invalid
169 designation. We didn't output the direction switch
170 literally because we didn't know about the invalid designation;
171 but we have to do so now. */
172 unsigned int output_direction_sequence :1;
175 EXFUN (Fcopy_coding_system, 2);
177 struct detection_state;
178 static int detect_coding_sjis (struct detection_state *st,
179 const unsigned char *src,
181 static void decode_coding_sjis (Lstream *decoding,
182 const unsigned char *src,
183 unsigned_char_dynarr *dst,
185 static void encode_coding_sjis (Lstream *encoding,
186 const unsigned char *src,
187 unsigned_char_dynarr *dst,
189 static int detect_coding_big5 (struct detection_state *st,
190 const unsigned char *src,
192 static void decode_coding_big5 (Lstream *decoding,
193 const unsigned char *src,
194 unsigned_char_dynarr *dst, unsigned int n);
195 static void encode_coding_big5 (Lstream *encoding,
196 const unsigned char *src,
197 unsigned_char_dynarr *dst, unsigned int n);
198 static int detect_coding_ucs4 (struct detection_state *st,
199 const unsigned char *src,
201 static void decode_coding_ucs4 (Lstream *decoding,
202 const unsigned char *src,
203 unsigned_char_dynarr *dst, unsigned int n);
204 static void encode_coding_ucs4 (Lstream *encoding,
205 const unsigned char *src,
206 unsigned_char_dynarr *dst, unsigned int n);
207 static int detect_coding_utf8 (struct detection_state *st,
208 const unsigned char *src,
210 static void decode_coding_utf8 (Lstream *decoding,
211 const unsigned char *src,
212 unsigned_char_dynarr *dst, unsigned int n);
213 static void encode_coding_utf8 (Lstream *encoding,
214 const unsigned char *src,
215 unsigned_char_dynarr *dst, unsigned int n);
216 static int postprocess_iso2022_mask (int mask);
217 static void reset_iso2022 (Lisp_Object coding_system,
218 struct iso2022_decoder *iso);
219 static int detect_coding_iso2022 (struct detection_state *st,
220 const unsigned char *src,
222 static void decode_coding_iso2022 (Lstream *decoding,
223 const unsigned char *src,
224 unsigned_char_dynarr *dst, unsigned int n);
225 static void encode_coding_iso2022 (Lstream *encoding,
226 const unsigned char *src,
227 unsigned_char_dynarr *dst, unsigned int n);
229 static void decode_coding_no_conversion (Lstream *decoding,
230 const unsigned char *src,
231 unsigned_char_dynarr *dst,
233 static void encode_coding_no_conversion (Lstream *encoding,
234 const unsigned char *src,
235 unsigned_char_dynarr *dst,
237 static void mule_decode (Lstream *decoding, const unsigned char *src,
238 unsigned_char_dynarr *dst, unsigned int n);
239 static void mule_encode (Lstream *encoding, const unsigned char *src,
240 unsigned_char_dynarr *dst, unsigned int n);
242 typedef struct codesys_prop codesys_prop;
251 Dynarr_declare (codesys_prop);
252 } codesys_prop_dynarr;
254 static const struct lrecord_description codesys_prop_description_1[] = {
255 { XD_LISP_OBJECT, offsetof (codesys_prop, sym) },
259 static const struct struct_description codesys_prop_description = {
260 sizeof (codesys_prop),
261 codesys_prop_description_1
264 static const struct lrecord_description codesys_prop_dynarr_description_1[] = {
265 XD_DYNARR_DESC (codesys_prop_dynarr, &codesys_prop_description),
269 static const struct struct_description codesys_prop_dynarr_description = {
270 sizeof (codesys_prop_dynarr),
271 codesys_prop_dynarr_description_1
274 codesys_prop_dynarr *the_codesys_prop_dynarr;
276 enum codesys_prop_enum
279 CODESYS_PROP_ISO2022,
284 /************************************************************************/
285 /* Coding system functions */
286 /************************************************************************/
288 static Lisp_Object mark_coding_system (Lisp_Object);
289 static void print_coding_system (Lisp_Object, Lisp_Object, int);
290 static void finalize_coding_system (void *header, int for_disksave);
293 static const struct lrecord_description ccs_description_1[] = {
294 { XD_LISP_OBJECT, offsetof (charset_conversion_spec, from_charset) },
295 { XD_LISP_OBJECT, offsetof (charset_conversion_spec, to_charset) },
299 static const struct struct_description ccs_description = {
300 sizeof (charset_conversion_spec),
304 static const struct lrecord_description ccsd_description_1[] = {
305 XD_DYNARR_DESC (charset_conversion_spec_dynarr, &ccs_description),
309 static const struct struct_description ccsd_description = {
310 sizeof (charset_conversion_spec_dynarr),
315 static const struct lrecord_description coding_system_description[] = {
316 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, name) },
317 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, doc_string) },
318 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, mnemonic) },
319 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, post_read_conversion) },
320 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, pre_write_conversion) },
321 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_lf) },
322 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_crlf) },
323 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, eol_cr) },
325 { XD_LISP_OBJECT_ARRAY, offsetof (Lisp_Coding_System, iso2022.initial_charset), 4 },
326 { XD_STRUCT_PTR, offsetof (Lisp_Coding_System, iso2022.input_conv), 1, &ccsd_description },
327 { XD_STRUCT_PTR, offsetof (Lisp_Coding_System, iso2022.output_conv), 1, &ccsd_description },
328 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccl.decode) },
329 { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, ccl.encode) },
334 DEFINE_LRECORD_IMPLEMENTATION ("coding-system", coding_system,
335 mark_coding_system, print_coding_system,
336 finalize_coding_system,
337 0, 0, coding_system_description,
341 mark_coding_system (Lisp_Object obj)
343 Lisp_Coding_System *codesys = XCODING_SYSTEM (obj);
345 mark_object (CODING_SYSTEM_NAME (codesys));
346 mark_object (CODING_SYSTEM_DOC_STRING (codesys));
347 mark_object (CODING_SYSTEM_MNEMONIC (codesys));
348 mark_object (CODING_SYSTEM_EOL_LF (codesys));
349 mark_object (CODING_SYSTEM_EOL_CRLF (codesys));
350 mark_object (CODING_SYSTEM_EOL_CR (codesys));
352 switch (CODING_SYSTEM_TYPE (codesys))
356 case CODESYS_ISO2022:
357 for (i = 0; i < 4; i++)
358 mark_object (CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i));
359 if (codesys->iso2022.input_conv)
361 for (i = 0; i < Dynarr_length (codesys->iso2022.input_conv); i++)
363 struct charset_conversion_spec *ccs =
364 Dynarr_atp (codesys->iso2022.input_conv, i);
365 mark_object (ccs->from_charset);
366 mark_object (ccs->to_charset);
369 if (codesys->iso2022.output_conv)
371 for (i = 0; i < Dynarr_length (codesys->iso2022.output_conv); i++)
373 struct charset_conversion_spec *ccs =
374 Dynarr_atp (codesys->iso2022.output_conv, i);
375 mark_object (ccs->from_charset);
376 mark_object (ccs->to_charset);
382 mark_object (CODING_SYSTEM_CCL_DECODE (codesys));
383 mark_object (CODING_SYSTEM_CCL_ENCODE (codesys));
390 mark_object (CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys));
391 return CODING_SYSTEM_POST_READ_CONVERSION (codesys);
395 print_coding_system (Lisp_Object obj, Lisp_Object printcharfun,
398 Lisp_Coding_System *c = XCODING_SYSTEM (obj);
400 error ("printing unreadable object #<coding_system 0x%x>",
403 write_c_string ("#<coding_system ", printcharfun);
404 print_internal (c->name, printcharfun, 1);
405 write_c_string (">", printcharfun);
409 finalize_coding_system (void *header, int for_disksave)
411 Lisp_Coding_System *c = (Lisp_Coding_System *) header;
412 /* Since coding systems never go away, this function is not
413 necessary. But it would be necessary if we changed things
414 so that coding systems could go away. */
415 if (!for_disksave) /* see comment in lstream.c */
417 switch (CODING_SYSTEM_TYPE (c))
420 case CODESYS_ISO2022:
421 if (c->iso2022.input_conv)
423 Dynarr_free (c->iso2022.input_conv);
424 c->iso2022.input_conv = 0;
426 if (c->iso2022.output_conv)
428 Dynarr_free (c->iso2022.output_conv);
429 c->iso2022.output_conv = 0;
440 symbol_to_eol_type (Lisp_Object symbol)
442 CHECK_SYMBOL (symbol);
443 if (NILP (symbol)) return EOL_AUTODETECT;
444 if (EQ (symbol, Qlf)) return EOL_LF;
445 if (EQ (symbol, Qcrlf)) return EOL_CRLF;
446 if (EQ (symbol, Qcr)) return EOL_CR;
448 signal_simple_error ("Unrecognized eol type", symbol);
449 return EOL_AUTODETECT; /* not reached */
453 eol_type_to_symbol (eol_type_t type)
458 case EOL_LF: return Qlf;
459 case EOL_CRLF: return Qcrlf;
460 case EOL_CR: return Qcr;
461 case EOL_AUTODETECT: return Qnil;
466 setup_eol_coding_systems (Lisp_Coding_System *codesys)
468 Lisp_Object codesys_obj;
469 int len = string_length (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name);
470 char *codesys_name = (char *) alloca (len + 7);
472 char *codesys_mnemonic=0;
474 Lisp_Object codesys_name_sym, sub_codesys_obj;
478 XSETCODING_SYSTEM (codesys_obj, codesys);
480 memcpy (codesys_name,
481 string_data (XSYMBOL (CODING_SYSTEM_NAME (codesys))->name), len);
483 if (STRINGP (CODING_SYSTEM_MNEMONIC (codesys)))
485 mlen = XSTRING_LENGTH (CODING_SYSTEM_MNEMONIC (codesys));
486 codesys_mnemonic = (char *) alloca (mlen + 7);
487 memcpy (codesys_mnemonic,
488 XSTRING_DATA (CODING_SYSTEM_MNEMONIC (codesys)), mlen);
491 #define DEFINE_SUB_CODESYS(op_sys, op_sys_abbr, Type) do { \
492 strcpy (codesys_name + len, "-" op_sys); \
494 strcpy (codesys_mnemonic + mlen, op_sys_abbr); \
495 codesys_name_sym = intern (codesys_name); \
496 sub_codesys_obj = Fcopy_coding_system (codesys_obj, codesys_name_sym); \
497 XCODING_SYSTEM_EOL_TYPE (sub_codesys_obj) = Type; \
499 XCODING_SYSTEM_MNEMONIC(sub_codesys_obj) = \
500 build_string (codesys_mnemonic); \
501 CODING_SYSTEM_##Type (codesys) = sub_codesys_obj; \
504 DEFINE_SUB_CODESYS("unix", "", EOL_LF);
505 DEFINE_SUB_CODESYS("dos", ":T", EOL_CRLF);
506 DEFINE_SUB_CODESYS("mac", ":t", EOL_CR);
509 DEFUN ("coding-system-p", Fcoding_system_p, 1, 1, 0, /*
510 Return t if OBJECT is a coding system.
511 A coding system is an object that defines how text containing multiple
512 character sets is encoded into a stream of (typically 8-bit) bytes.
513 The coding system is used to decode the stream into a series of
514 characters (which may be from multiple charsets) when the text is read
515 from a file or process, and is used to encode the text back into the
516 same format when it is written out to a file or process.
518 For example, many ISO2022-compliant coding systems (such as Compound
519 Text, which is used for inter-client data under the X Window System)
520 use escape sequences to switch between different charsets -- Japanese
521 Kanji, for example, is invoked with "ESC $ ( B"; ASCII is invoked
522 with "ESC ( B"; and Cyrillic is invoked with "ESC - L". See
523 `make-coding-system' for more information.
525 Coding systems are normally identified using a symbol, and the
526 symbol is accepted in place of the actual coding system object whenever
527 a coding system is called for. (This is similar to how faces work.)
531 return CODING_SYSTEMP (object) ? Qt : Qnil;
534 DEFUN ("find-coding-system", Ffind_coding_system, 1, 1, 0, /*
535 Retrieve the coding system of the given name.
537 If CODING-SYSTEM-OR-NAME is a coding-system object, it is simply
538 returned. Otherwise, CODING-SYSTEM-OR-NAME should be a symbol.
539 If there is no such coding system, nil is returned. Otherwise the
540 associated coding system object is returned.
542 (coding_system_or_name))
544 if (NILP (coding_system_or_name))
545 coding_system_or_name = Qbinary;
546 else if (CODING_SYSTEMP (coding_system_or_name))
547 return coding_system_or_name;
549 CHECK_SYMBOL (coding_system_or_name);
553 coding_system_or_name =
554 Fgethash (coding_system_or_name, Vcoding_system_hash_table, Qnil);
556 if (CODING_SYSTEMP (coding_system_or_name) || NILP (coding_system_or_name))
557 return coding_system_or_name;
561 DEFUN ("get-coding-system", Fget_coding_system, 1, 1, 0, /*
562 Retrieve the coding system of the given name.
563 Same as `find-coding-system' except that if there is no such
564 coding system, an error is signaled instead of returning nil.
568 Lisp_Object coding_system = Ffind_coding_system (name);
570 if (NILP (coding_system))
571 signal_simple_error ("No such coding system", name);
572 return coding_system;
575 /* We store the coding systems in hash tables with the names as the key and the
576 actual coding system object as the value. Occasionally we need to use them
577 in a list format. These routines provide us with that. */
578 struct coding_system_list_closure
580 Lisp_Object *coding_system_list;
584 add_coding_system_to_list_mapper (Lisp_Object key, Lisp_Object value,
585 void *coding_system_list_closure)
587 /* This function can GC */
588 struct coding_system_list_closure *cscl =
589 (struct coding_system_list_closure *) coding_system_list_closure;
590 Lisp_Object *coding_system_list = cscl->coding_system_list;
592 *coding_system_list = Fcons (key, *coding_system_list);
596 DEFUN ("coding-system-list", Fcoding_system_list, 0, 0, 0, /*
597 Return a list of the names of all defined coding systems.
601 Lisp_Object coding_system_list = Qnil;
603 struct coding_system_list_closure coding_system_list_closure;
605 GCPRO1 (coding_system_list);
606 coding_system_list_closure.coding_system_list = &coding_system_list;
607 elisp_maphash (add_coding_system_to_list_mapper, Vcoding_system_hash_table,
608 &coding_system_list_closure);
611 return coding_system_list;
614 DEFUN ("coding-system-name", Fcoding_system_name, 1, 1, 0, /*
615 Return the name of the given coding system.
619 coding_system = Fget_coding_system (coding_system);
620 return XCODING_SYSTEM_NAME (coding_system);
623 static Lisp_Coding_System *
624 allocate_coding_system (enum coding_system_type type, Lisp_Object name)
626 Lisp_Coding_System *codesys =
627 alloc_lcrecord_type (Lisp_Coding_System, &lrecord_coding_system);
629 zero_lcrecord (codesys);
630 CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = Qnil;
631 CODING_SYSTEM_POST_READ_CONVERSION (codesys) = Qnil;
632 CODING_SYSTEM_EOL_TYPE (codesys) = EOL_AUTODETECT;
633 CODING_SYSTEM_EOL_CRLF (codesys) = Qnil;
634 CODING_SYSTEM_EOL_CR (codesys) = Qnil;
635 CODING_SYSTEM_EOL_LF (codesys) = Qnil;
636 CODING_SYSTEM_TYPE (codesys) = type;
637 CODING_SYSTEM_MNEMONIC (codesys) = Qnil;
639 if (type == CODESYS_ISO2022)
642 for (i = 0; i < 4; i++)
643 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i) = Qnil;
645 else if (type == CODESYS_CCL)
647 CODING_SYSTEM_CCL_DECODE (codesys) = Qnil;
648 CODING_SYSTEM_CCL_ENCODE (codesys) = Qnil;
651 CODING_SYSTEM_NAME (codesys) = name;
657 /* Given a list of charset conversion specs as specified in a Lisp
658 program, parse it into STORE_HERE. */
661 parse_charset_conversion_specs (charset_conversion_spec_dynarr *store_here,
662 Lisp_Object spec_list)
666 EXTERNAL_LIST_LOOP (rest, spec_list)
668 Lisp_Object car = XCAR (rest);
669 Lisp_Object from, to;
670 struct charset_conversion_spec spec;
672 if (!CONSP (car) || !CONSP (XCDR (car)) || !NILP (XCDR (XCDR (car))))
673 signal_simple_error ("Invalid charset conversion spec", car);
674 from = Fget_charset (XCAR (car));
675 to = Fget_charset (XCAR (XCDR (car)));
676 if (XCHARSET_TYPE (from) != XCHARSET_TYPE (to))
677 signal_simple_error_2
678 ("Attempted conversion between different charset types",
680 spec.from_charset = from;
681 spec.to_charset = to;
683 Dynarr_add (store_here, spec);
687 /* Given a dynarr LOAD_HERE of internally-stored charset conversion
688 specs, return the equivalent as the Lisp programmer would see it.
690 If LOAD_HERE is 0, return Qnil. */
693 unparse_charset_conversion_specs (charset_conversion_spec_dynarr *load_here)
700 for (i = 0, result = Qnil; i < Dynarr_length (load_here); i++)
702 struct charset_conversion_spec *ccs = Dynarr_atp (load_here, i);
703 result = Fcons (list2 (ccs->from_charset, ccs->to_charset), result);
706 return Fnreverse (result);
711 DEFUN ("make-coding-system", Fmake_coding_system, 2, 4, 0, /*
712 Register symbol NAME as a coding system.
714 TYPE describes the conversion method used and should be one of
717 Automatic conversion. XEmacs attempts to detect the coding system
720 No conversion. Use this for binary files and such. On output,
721 graphic characters that are not in ASCII or Latin-1 will be
722 replaced by a ?. (For a no-conversion-encoded buffer, these
723 characters will only be present if you explicitly insert them.)
725 Shift-JIS (a Japanese encoding commonly used in PC operating systems).
727 ISO 10646 UCS-4 encoding.
729 ISO 10646 UTF-8 encoding.
731 Any ISO2022-compliant encoding. Among other things, this includes
732 JIS (the Japanese encoding commonly used for e-mail), EUC (the
733 standard Unix encoding for Japanese and other languages), and
734 Compound Text (the encoding used in X11). You can specify more
735 specific information about the conversion with the FLAGS argument.
737 Big5 (the encoding commonly used for Taiwanese).
739 The conversion is performed using a user-written pseudo-code
740 program. CCL (Code Conversion Language) is the name of this
743 Write out or read in the raw contents of the memory representing
744 the buffer's text. This is primarily useful for debugging
745 purposes, and is only enabled when XEmacs has been compiled with
746 DEBUG_XEMACS defined (via the --debug configure option).
747 WARNING: Reading in a file using 'internal conversion can result
748 in an internal inconsistency in the memory representing a
749 buffer's text, which will produce unpredictable results and may
750 cause XEmacs to crash. Under normal circumstances you should
751 never use 'internal conversion.
753 DOC-STRING is a string describing the coding system.
755 PROPS is a property list, describing the specific nature of the
756 character set. Recognized properties are:
759 String to be displayed in the modeline when this coding system is
763 End-of-line conversion to be used. It should be one of
766 Automatically detect the end-of-line type (LF, CRLF,
767 or CR). Also generate subsidiary coding systems named
768 `NAME-unix', `NAME-dos', and `NAME-mac', that are
769 identical to this coding system but have an EOL-TYPE
770 value of 'lf, 'crlf, and 'cr, respectively.
772 The end of a line is marked externally using ASCII LF.
773 Since this is also the way that XEmacs represents an
774 end-of-line internally, specifying this option results
775 in no end-of-line conversion. This is the standard
776 format for Unix text files.
778 The end of a line is marked externally using ASCII
779 CRLF. This is the standard format for MS-DOS text
782 The end of a line is marked externally using ASCII CR.
783 This is the standard format for Macintosh text files.
785 Automatically detect the end-of-line type but do not
786 generate subsidiary coding systems. (This value is
787 converted to nil when stored internally, and
788 `coding-system-property' will return nil.)
790 'post-read-conversion
791 Function called after a file has been read in, to perform the
792 decoding. Called with two arguments, BEG and END, denoting
793 a region of the current buffer to be decoded.
795 'pre-write-conversion
796 Function called before a file is written out, to perform the
797 encoding. Called with two arguments, BEG and END, denoting
798 a region of the current buffer to be encoded.
801 The following additional properties are recognized if TYPE is 'iso2022:
807 The character set initially designated to the G0 - G3 registers.
808 The value should be one of
810 -- A charset object (designate that character set)
811 -- nil (do not ever use this register)
812 -- t (no character set is initially designated to
813 the register, but may be later on; this automatically
814 sets the corresponding `force-g*-on-output' property)
820 If non-nil, send an explicit designation sequence on output before
821 using the specified register.
824 If non-nil, use the short forms "ESC $ @", "ESC $ A", and
825 "ESC $ B" on output in place of the full designation sequences
826 "ESC $ ( @", "ESC $ ( A", and "ESC $ ( B".
829 If non-nil, don't designate ASCII to G0 at each end of line on output.
830 Setting this to non-nil also suppresses other state-resetting that
831 normally happens at the end of a line.
834 If non-nil, don't designate ASCII to G0 before control chars on output.
837 If non-nil, use 7-bit environment on output. Otherwise, use 8-bit
841 If non-nil, use locking-shift (SO/SI) instead of single-shift
842 or designation by escape sequence.
845 If non-nil, don't use ISO6429's direction specification.
848 If non-nil, literal control characters that are the same as
849 the beginning of a recognized ISO2022 or ISO6429 escape sequence
850 (in particular, ESC (0x1B), SO (0x0E), SI (0x0F), SS2 (0x8E),
851 SS3 (0x8F), and CSI (0x9B)) are "quoted" with an escape character
852 so that they can be properly distinguished from an escape sequence.
853 (Note that doing this results in a non-portable encoding.) This
854 encoding flag is used for byte-compiled files. Note that ESC
855 is a good choice for a quoting character because there are no
856 escape sequences whose second byte is a character from the Control-0
857 or Control-1 character sets; this is explicitly disallowed by the
860 'input-charset-conversion
861 A list of conversion specifications, specifying conversion of
862 characters in one charset to another when decoding is performed.
863 Each specification is a list of two elements: the source charset,
864 and the destination charset.
866 'output-charset-conversion
867 A list of conversion specifications, specifying conversion of
868 characters in one charset to another when encoding is performed.
869 The form of each specification is the same as for
870 'input-charset-conversion.
873 The following additional properties are recognized (and required)
877 CCL program used for decoding (converting to internal format).
880 CCL program used for encoding (converting to external format).
882 (name, type, doc_string, props))
884 Lisp_Coding_System *codesys;
885 Lisp_Object rest, key, value;
886 enum coding_system_type ty;
887 int need_to_setup_eol_systems = 1;
889 /* Convert type to constant */
890 if (NILP (type) || EQ (type, Qundecided))
891 { ty = CODESYS_AUTODETECT; }
893 else if (EQ (type, Qshift_jis)) { ty = CODESYS_SHIFT_JIS; }
894 else if (EQ (type, Qiso2022)) { ty = CODESYS_ISO2022; }
895 else if (EQ (type, Qbig5)) { ty = CODESYS_BIG5; }
896 else if (EQ (type, Qucs4)) { ty = CODESYS_UCS4; }
897 else if (EQ (type, Qutf8)) { ty = CODESYS_UTF8; }
898 else if (EQ (type, Qccl)) { ty = CODESYS_CCL; }
900 else if (EQ (type, Qno_conversion)) { ty = CODESYS_NO_CONVERSION; }
902 else if (EQ (type, Qinternal)) { ty = CODESYS_INTERNAL; }
905 signal_simple_error ("Invalid coding system type", type);
909 codesys = allocate_coding_system (ty, name);
911 if (NILP (doc_string))
912 doc_string = build_string ("");
914 CHECK_STRING (doc_string);
915 CODING_SYSTEM_DOC_STRING (codesys) = doc_string;
917 EXTERNAL_PROPERTY_LIST_LOOP (rest, key, value, props)
919 if (EQ (key, Qmnemonic))
922 CHECK_STRING (value);
923 CODING_SYSTEM_MNEMONIC (codesys) = value;
926 else if (EQ (key, Qeol_type))
928 need_to_setup_eol_systems = NILP (value);
931 CODING_SYSTEM_EOL_TYPE (codesys) = symbol_to_eol_type (value);
934 else if (EQ (key, Qpost_read_conversion)) CODING_SYSTEM_POST_READ_CONVERSION (codesys) = value;
935 else if (EQ (key, Qpre_write_conversion)) CODING_SYSTEM_PRE_WRITE_CONVERSION (codesys) = value;
937 else if (ty == CODESYS_ISO2022)
939 #define FROB_INITIAL_CHARSET(charset_num) \
940 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, charset_num) = \
941 ((EQ (value, Qt) || EQ (value, Qnil)) ? value : Fget_charset (value))
943 if (EQ (key, Qcharset_g0)) FROB_INITIAL_CHARSET (0);
944 else if (EQ (key, Qcharset_g1)) FROB_INITIAL_CHARSET (1);
945 else if (EQ (key, Qcharset_g2)) FROB_INITIAL_CHARSET (2);
946 else if (EQ (key, Qcharset_g3)) FROB_INITIAL_CHARSET (3);
948 #define FROB_FORCE_CHARSET(charset_num) \
949 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (codesys, charset_num) = !NILP (value)
951 else if (EQ (key, Qforce_g0_on_output)) FROB_FORCE_CHARSET (0);
952 else if (EQ (key, Qforce_g1_on_output)) FROB_FORCE_CHARSET (1);
953 else if (EQ (key, Qforce_g2_on_output)) FROB_FORCE_CHARSET (2);
954 else if (EQ (key, Qforce_g3_on_output)) FROB_FORCE_CHARSET (3);
956 #define FROB_BOOLEAN_PROPERTY(prop) \
957 CODING_SYSTEM_ISO2022_##prop (codesys) = !NILP (value)
959 else if (EQ (key, Qshort)) FROB_BOOLEAN_PROPERTY (SHORT);
960 else if (EQ (key, Qno_ascii_eol)) FROB_BOOLEAN_PROPERTY (NO_ASCII_EOL);
961 else if (EQ (key, Qno_ascii_cntl)) FROB_BOOLEAN_PROPERTY (NO_ASCII_CNTL);
962 else if (EQ (key, Qseven)) FROB_BOOLEAN_PROPERTY (SEVEN);
963 else if (EQ (key, Qlock_shift)) FROB_BOOLEAN_PROPERTY (LOCK_SHIFT);
964 else if (EQ (key, Qno_iso6429)) FROB_BOOLEAN_PROPERTY (NO_ISO6429);
965 else if (EQ (key, Qescape_quoted)) FROB_BOOLEAN_PROPERTY (ESCAPE_QUOTED);
967 else if (EQ (key, Qinput_charset_conversion))
969 codesys->iso2022.input_conv =
970 Dynarr_new (charset_conversion_spec);
971 parse_charset_conversion_specs (codesys->iso2022.input_conv,
974 else if (EQ (key, Qoutput_charset_conversion))
976 codesys->iso2022.output_conv =
977 Dynarr_new (charset_conversion_spec);
978 parse_charset_conversion_specs (codesys->iso2022.output_conv,
982 signal_simple_error ("Unrecognized property", key);
984 else if (EQ (type, Qccl))
986 if (EQ (key, Qdecode))
988 CHECK_VECTOR (value);
989 CODING_SYSTEM_CCL_DECODE (codesys) = value;
991 else if (EQ (key, Qencode))
993 CHECK_VECTOR (value);
994 CODING_SYSTEM_CCL_ENCODE (codesys) = value;
997 signal_simple_error ("Unrecognized property", key);
1001 signal_simple_error ("Unrecognized property", key);
1004 if (need_to_setup_eol_systems)
1005 setup_eol_coding_systems (codesys);
1008 Lisp_Object codesys_obj;
1009 XSETCODING_SYSTEM (codesys_obj, codesys);
1010 Fputhash (name, codesys_obj, Vcoding_system_hash_table);
1015 DEFUN ("copy-coding-system", Fcopy_coding_system, 2, 2, 0, /*
1016 Copy OLD-CODING-SYSTEM to NEW-NAME.
1017 If NEW-NAME does not name an existing coding system, a new one will
1020 (old_coding_system, new_name))
1022 Lisp_Object new_coding_system;
1023 old_coding_system = Fget_coding_system (old_coding_system);
1024 new_coding_system = Ffind_coding_system (new_name);
1025 if (NILP (new_coding_system))
1027 XSETCODING_SYSTEM (new_coding_system,
1028 allocate_coding_system
1029 (XCODING_SYSTEM_TYPE (old_coding_system),
1031 Fputhash (new_name, new_coding_system, Vcoding_system_hash_table);
1035 Lisp_Coding_System *to = XCODING_SYSTEM (new_coding_system);
1036 Lisp_Coding_System *from = XCODING_SYSTEM (old_coding_system);
1037 memcpy (((char *) to ) + sizeof (to->header),
1038 ((char *) from) + sizeof (from->header),
1039 sizeof (*from) - sizeof (from->header));
1040 to->name = new_name;
1042 return new_coding_system;
1045 DEFUN ("coding-system-canonical-name-p", Fcoding_system_canonical_name_p, 1, 1, 0, /*
1046 Return t if OBJECT names a coding system, and is not a coding system alias.
1050 return CODING_SYSTEMP (Fgethash (object, Vcoding_system_hash_table, Qnil))
1054 DEFUN ("coding-system-alias-p", Fcoding_system_alias_p, 1, 1, 0, /*
1055 Return t if OBJECT is a coding system alias.
1056 All coding system aliases are created by `define-coding-system-alias'.
1060 return SYMBOLP (Fgethash (object, Vcoding_system_hash_table, Qzero))
1064 DEFUN ("coding-system-aliasee", Fcoding_system_aliasee, 1, 1, 0, /*
1065 Return the coding-system symbol for which symbol ALIAS is an alias.
1069 Lisp_Object aliasee = Fgethash (alias, Vcoding_system_hash_table, Qnil);
1070 if (SYMBOLP (aliasee))
1073 signal_simple_error ("Symbol is not a coding system alias", alias);
1077 append_suffix_to_symbol (Lisp_Object symbol, const char *ascii_string)
1079 return Fintern (concat2 (Fsymbol_name (symbol), build_string (ascii_string)),
1083 /* A maphash function, for removing dangling coding system aliases. */
1085 dangling_coding_system_alias_p (Lisp_Object alias,
1086 Lisp_Object aliasee,
1087 void *dangling_aliases)
1089 if (SYMBOLP (aliasee)
1090 && NILP (Fgethash (aliasee, Vcoding_system_hash_table, Qnil)))
1092 (*(int *) dangling_aliases)++;
1099 DEFUN ("define-coding-system-alias", Fdefine_coding_system_alias, 2, 2, 0, /*
1100 Define symbol ALIAS as an alias for coding system ALIASEE.
1102 You can use this function to redefine an alias that has already been defined,
1103 but you cannot redefine a name which is the canonical name for a coding system.
1104 \(a canonical name of a coding system is what is returned when you call
1105 `coding-system-name' on a coding system).
1107 ALIASEE itself can be an alias, which allows you to define nested aliases.
1109 You are forbidden, however, from creating alias loops or `dangling' aliases.
1110 These will be detected, and an error will be signaled if you attempt to do so.
1112 If ALIASEE is nil, then ALIAS will simply be undefined.
1114 See also `coding-system-alias-p', `coding-system-aliasee',
1115 and `coding-system-canonical-name-p'.
1119 Lisp_Object real_coding_system, probe;
1121 CHECK_SYMBOL (alias);
1123 if (!NILP (Fcoding_system_canonical_name_p (alias)))
1125 ("Symbol is the canonical name of a coding system and cannot be redefined",
1130 Lisp_Object subsidiary_unix = append_suffix_to_symbol (alias, "-unix");
1131 Lisp_Object subsidiary_dos = append_suffix_to_symbol (alias, "-dos");
1132 Lisp_Object subsidiary_mac = append_suffix_to_symbol (alias, "-mac");
1134 Fremhash (alias, Vcoding_system_hash_table);
1136 /* Undefine subsidiary aliases,
1137 presumably created by a previous call to this function */
1138 if (! NILP (Fcoding_system_alias_p (subsidiary_unix)) &&
1139 ! NILP (Fcoding_system_alias_p (subsidiary_dos)) &&
1140 ! NILP (Fcoding_system_alias_p (subsidiary_mac)))
1142 Fdefine_coding_system_alias (subsidiary_unix, Qnil);
1143 Fdefine_coding_system_alias (subsidiary_dos, Qnil);
1144 Fdefine_coding_system_alias (subsidiary_mac, Qnil);
1147 /* Undefine dangling coding system aliases. */
1149 int dangling_aliases;
1152 dangling_aliases = 0;
1153 elisp_map_remhash (dangling_coding_system_alias_p,
1154 Vcoding_system_hash_table,
1156 } while (dangling_aliases > 0);
1162 if (CODING_SYSTEMP (aliasee))
1163 aliasee = XCODING_SYSTEM_NAME (aliasee);
1165 /* Checks that aliasee names a coding-system */
1166 real_coding_system = Fget_coding_system (aliasee);
1168 /* Check for coding system alias loops */
1169 if (EQ (alias, aliasee))
1170 alias_loop: signal_simple_error_2
1171 ("Attempt to create a coding system alias loop", alias, aliasee);
1173 for (probe = aliasee;
1175 probe = Fgethash (probe, Vcoding_system_hash_table, Qzero))
1177 if (EQ (probe, alias))
1181 Fputhash (alias, aliasee, Vcoding_system_hash_table);
1183 /* Set up aliases for subsidiaries.
1184 #### There must be a better way to handle subsidiary coding systems. */
1186 static const char *suffixes[] = { "-unix", "-dos", "-mac" };
1188 for (i = 0; i < countof (suffixes); i++)
1190 Lisp_Object alias_subsidiary =
1191 append_suffix_to_symbol (alias, suffixes[i]);
1192 Lisp_Object aliasee_subsidiary =
1193 append_suffix_to_symbol (aliasee, suffixes[i]);
1195 if (! NILP (Ffind_coding_system (aliasee_subsidiary)))
1196 Fdefine_coding_system_alias (alias_subsidiary, aliasee_subsidiary);
1199 /* FSF return value is a vector of [ALIAS-unix ALIAS-dos ALIAS-mac],
1200 but it doesn't look intentional, so I'd rather return something
1201 meaningful or nothing at all. */
1206 subsidiary_coding_system (Lisp_Object coding_system, eol_type_t type)
1208 Lisp_Coding_System *cs = XCODING_SYSTEM (coding_system);
1209 Lisp_Object new_coding_system;
1211 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
1212 return coding_system;
1216 case EOL_AUTODETECT: return coding_system;
1217 case EOL_LF: new_coding_system = CODING_SYSTEM_EOL_LF (cs); break;
1218 case EOL_CR: new_coding_system = CODING_SYSTEM_EOL_CR (cs); break;
1219 case EOL_CRLF: new_coding_system = CODING_SYSTEM_EOL_CRLF (cs); break;
1223 return NILP (new_coding_system) ? coding_system : new_coding_system;
1226 DEFUN ("subsidiary-coding-system", Fsubsidiary_coding_system, 2, 2, 0, /*
1227 Return the subsidiary coding system of CODING-SYSTEM with eol type EOL-TYPE.
1229 (coding_system, eol_type))
1231 coding_system = Fget_coding_system (coding_system);
1233 return subsidiary_coding_system (coding_system,
1234 symbol_to_eol_type (eol_type));
1238 /************************************************************************/
1239 /* Coding system accessors */
1240 /************************************************************************/
1242 DEFUN ("coding-system-doc-string", Fcoding_system_doc_string, 1, 1, 0, /*
1243 Return the doc string for CODING-SYSTEM.
1247 coding_system = Fget_coding_system (coding_system);
1248 return XCODING_SYSTEM_DOC_STRING (coding_system);
1251 DEFUN ("coding-system-type", Fcoding_system_type, 1, 1, 0, /*
1252 Return the type of CODING-SYSTEM.
1256 switch (XCODING_SYSTEM_TYPE (Fget_coding_system (coding_system)))
1259 case CODESYS_AUTODETECT: return Qundecided;
1261 case CODESYS_SHIFT_JIS: return Qshift_jis;
1262 case CODESYS_ISO2022: return Qiso2022;
1263 case CODESYS_BIG5: return Qbig5;
1264 case CODESYS_UCS4: return Qucs4;
1265 case CODESYS_UTF8: return Qutf8;
1266 case CODESYS_CCL: return Qccl;
1268 case CODESYS_NO_CONVERSION: return Qno_conversion;
1270 case CODESYS_INTERNAL: return Qinternal;
1277 Lisp_Object coding_system_charset (Lisp_Object coding_system, int gnum)
1280 = XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, gnum);
1282 return CHARSETP (cs) ? XCHARSET_NAME (cs) : Qnil;
1285 DEFUN ("coding-system-charset", Fcoding_system_charset, 2, 2, 0, /*
1286 Return initial charset of CODING-SYSTEM designated to GNUM.
1289 (coding_system, gnum))
1291 coding_system = Fget_coding_system (coding_system);
1294 return coding_system_charset (coding_system, XINT (gnum));
1298 DEFUN ("coding-system-property", Fcoding_system_property, 2, 2, 0, /*
1299 Return the PROP property of CODING-SYSTEM.
1301 (coding_system, prop))
1304 enum coding_system_type type;
1306 coding_system = Fget_coding_system (coding_system);
1307 CHECK_SYMBOL (prop);
1308 type = XCODING_SYSTEM_TYPE (coding_system);
1310 for (i = 0; !ok && i < Dynarr_length (the_codesys_prop_dynarr); i++)
1311 if (EQ (Dynarr_at (the_codesys_prop_dynarr, i).sym, prop))
1314 switch (Dynarr_at (the_codesys_prop_dynarr, i).prop_type)
1316 case CODESYS_PROP_ALL_OK:
1319 case CODESYS_PROP_ISO2022:
1320 if (type != CODESYS_ISO2022)
1322 ("Property only valid in ISO2022 coding systems",
1326 case CODESYS_PROP_CCL:
1327 if (type != CODESYS_CCL)
1329 ("Property only valid in CCL coding systems",
1339 signal_simple_error ("Unrecognized property", prop);
1341 if (EQ (prop, Qname))
1342 return XCODING_SYSTEM_NAME (coding_system);
1343 else if (EQ (prop, Qtype))
1344 return Fcoding_system_type (coding_system);
1345 else if (EQ (prop, Qdoc_string))
1346 return XCODING_SYSTEM_DOC_STRING (coding_system);
1347 else if (EQ (prop, Qmnemonic))
1348 return XCODING_SYSTEM_MNEMONIC (coding_system);
1349 else if (EQ (prop, Qeol_type))
1350 return eol_type_to_symbol (XCODING_SYSTEM_EOL_TYPE (coding_system));
1351 else if (EQ (prop, Qeol_lf))
1352 return XCODING_SYSTEM_EOL_LF (coding_system);
1353 else if (EQ (prop, Qeol_crlf))
1354 return XCODING_SYSTEM_EOL_CRLF (coding_system);
1355 else if (EQ (prop, Qeol_cr))
1356 return XCODING_SYSTEM_EOL_CR (coding_system);
1357 else if (EQ (prop, Qpost_read_conversion))
1358 return XCODING_SYSTEM_POST_READ_CONVERSION (coding_system);
1359 else if (EQ (prop, Qpre_write_conversion))
1360 return XCODING_SYSTEM_PRE_WRITE_CONVERSION (coding_system);
1362 else if (type == CODESYS_ISO2022)
1364 if (EQ (prop, Qcharset_g0))
1365 return coding_system_charset (coding_system, 0);
1366 else if (EQ (prop, Qcharset_g1))
1367 return coding_system_charset (coding_system, 1);
1368 else if (EQ (prop, Qcharset_g2))
1369 return coding_system_charset (coding_system, 2);
1370 else if (EQ (prop, Qcharset_g3))
1371 return coding_system_charset (coding_system, 3);
1373 #define FORCE_CHARSET(charset_num) \
1374 (XCODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT \
1375 (coding_system, charset_num) ? Qt : Qnil)
1377 else if (EQ (prop, Qforce_g0_on_output)) return FORCE_CHARSET (0);
1378 else if (EQ (prop, Qforce_g1_on_output)) return FORCE_CHARSET (1);
1379 else if (EQ (prop, Qforce_g2_on_output)) return FORCE_CHARSET (2);
1380 else if (EQ (prop, Qforce_g3_on_output)) return FORCE_CHARSET (3);
1382 #define LISP_BOOLEAN(prop) \
1383 (XCODING_SYSTEM_ISO2022_##prop (coding_system) ? Qt : Qnil)
1385 else if (EQ (prop, Qshort)) return LISP_BOOLEAN (SHORT);
1386 else if (EQ (prop, Qno_ascii_eol)) return LISP_BOOLEAN (NO_ASCII_EOL);
1387 else if (EQ (prop, Qno_ascii_cntl)) return LISP_BOOLEAN (NO_ASCII_CNTL);
1388 else if (EQ (prop, Qseven)) return LISP_BOOLEAN (SEVEN);
1389 else if (EQ (prop, Qlock_shift)) return LISP_BOOLEAN (LOCK_SHIFT);
1390 else if (EQ (prop, Qno_iso6429)) return LISP_BOOLEAN (NO_ISO6429);
1391 else if (EQ (prop, Qescape_quoted)) return LISP_BOOLEAN (ESCAPE_QUOTED);
1393 else if (EQ (prop, Qinput_charset_conversion))
1395 unparse_charset_conversion_specs
1396 (XCODING_SYSTEM (coding_system)->iso2022.input_conv);
1397 else if (EQ (prop, Qoutput_charset_conversion))
1399 unparse_charset_conversion_specs
1400 (XCODING_SYSTEM (coding_system)->iso2022.output_conv);
1404 else if (type == CODESYS_CCL)
1406 if (EQ (prop, Qdecode))
1407 return XCODING_SYSTEM_CCL_DECODE (coding_system);
1408 else if (EQ (prop, Qencode))
1409 return XCODING_SYSTEM_CCL_ENCODE (coding_system);
1417 return Qnil; /* not reached */
1421 /************************************************************************/
1422 /* Coding category functions */
1423 /************************************************************************/
1426 decode_coding_category (Lisp_Object symbol)
1430 CHECK_SYMBOL (symbol);
1431 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1432 if (EQ (coding_category_symbol[i], symbol))
1435 signal_simple_error ("Unrecognized coding category", symbol);
1436 return 0; /* not reached */
1439 DEFUN ("coding-category-list", Fcoding_category_list, 0, 0, 0, /*
1440 Return a list of all recognized coding categories.
1445 Lisp_Object list = Qnil;
1447 for (i = CODING_CATEGORY_LAST; i >= 0; i--)
1448 list = Fcons (coding_category_symbol[i], list);
1452 DEFUN ("set-coding-priority-list", Fset_coding_priority_list, 1, 1, 0, /*
1453 Change the priority order of the coding categories.
1454 LIST should be list of coding categories, in descending order of
1455 priority. Unspecified coding categories will be lower in priority
1456 than all specified ones, in the same relative order they were in
1461 int category_to_priority[CODING_CATEGORY_LAST + 1];
1465 /* First generate a list that maps coding categories to priorities. */
1467 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1468 category_to_priority[i] = -1;
1470 /* Highest priority comes from the specified list. */
1472 EXTERNAL_LIST_LOOP (rest, list)
1474 int cat = decode_coding_category (XCAR (rest));
1476 if (category_to_priority[cat] >= 0)
1477 signal_simple_error ("Duplicate coding category in list", XCAR (rest));
1478 category_to_priority[cat] = i++;
1481 /* Now go through the existing categories by priority to retrieve
1482 the categories not yet specified and preserve their priority
1484 for (j = 0; j <= CODING_CATEGORY_LAST; j++)
1486 int cat = fcd->coding_category_by_priority[j];
1487 if (category_to_priority[cat] < 0)
1488 category_to_priority[cat] = i++;
1491 /* Now we need to construct the inverse of the mapping we just
1494 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1495 fcd->coding_category_by_priority[category_to_priority[i]] = i;
1497 /* Phew! That was confusing. */
1501 DEFUN ("coding-priority-list", Fcoding_priority_list, 0, 0, 0, /*
1502 Return a list of coding categories in descending order of priority.
1507 Lisp_Object list = Qnil;
1509 for (i = CODING_CATEGORY_LAST; i >= 0; i--)
1510 list = Fcons (coding_category_symbol[fcd->coding_category_by_priority[i]],
1515 DEFUN ("set-coding-category-system", Fset_coding_category_system, 2, 2, 0, /*
1516 Change the coding system associated with a coding category.
1518 (coding_category, coding_system))
1520 int cat = decode_coding_category (coding_category);
1522 coding_system = Fget_coding_system (coding_system);
1523 fcd->coding_category_system[cat] = coding_system;
1527 DEFUN ("coding-category-system", Fcoding_category_system, 1, 1, 0, /*
1528 Return the coding system associated with a coding category.
1532 int cat = decode_coding_category (coding_category);
1533 Lisp_Object sys = fcd->coding_category_system[cat];
1536 return XCODING_SYSTEM_NAME (sys);
1541 /************************************************************************/
1542 /* Detecting the encoding of data */
1543 /************************************************************************/
1545 struct detection_state
1547 eol_type_t eol_type;
1583 struct iso2022_decoder iso;
1585 int high_byte_count;
1586 unsigned int saw_single_shift:1;
1599 acceptable_control_char_p (int c)
1603 /* Allow and ignore control characters that you might
1604 reasonably see in a text file */
1609 case 8: /* backspace */
1610 case 11: /* vertical tab */
1611 case 12: /* form feed */
1612 case 26: /* MS-DOS C-z junk */
1613 case 31: /* '^_' -- for info */
1621 mask_has_at_most_one_bit_p (int mask)
1623 /* Perhaps the only thing useful you learn from intensive Microsoft
1624 technical interviews */
1625 return (mask & (mask - 1)) == 0;
1629 detect_eol_type (struct detection_state *st, const unsigned char *src,
1639 if (st->eol.just_saw_cr)
1641 else if (st->eol.seen_anything)
1644 else if (st->eol.just_saw_cr)
1647 st->eol.just_saw_cr = 1;
1649 st->eol.just_saw_cr = 0;
1650 st->eol.seen_anything = 1;
1653 return EOL_AUTODETECT;
1656 /* Attempt to determine the encoding and EOL type of the given text.
1657 Before calling this function for the first type, you must initialize
1658 st->eol_type as appropriate and initialize st->mask to ~0.
1660 st->eol_type holds the determined EOL type, or EOL_AUTODETECT if
1663 st->mask holds the determined coding category mask, or ~0 if only
1664 ASCII has been seen so far.
1668 0 == st->eol_type is EOL_AUTODETECT and/or more than coding category
1669 is present in st->mask
1670 1 == definitive answers are here for both st->eol_type and st->mask
1674 detect_coding_type (struct detection_state *st, const Extbyte *src,
1675 unsigned int n, int just_do_eol)
1679 if (st->eol_type == EOL_AUTODETECT)
1680 st->eol_type = detect_eol_type (st, src, n);
1683 return st->eol_type != EOL_AUTODETECT;
1685 if (!st->seen_non_ascii)
1687 for (; n; n--, src++)
1690 if ((c < 0x20 && !acceptable_control_char_p (c)) || c >= 0x80)
1692 st->seen_non_ascii = 1;
1694 st->shift_jis.mask = ~0;
1698 st->iso2022.mask = ~0;
1708 if (!mask_has_at_most_one_bit_p (st->iso2022.mask))
1709 st->iso2022.mask = detect_coding_iso2022 (st, src, n);
1710 if (!mask_has_at_most_one_bit_p (st->shift_jis.mask))
1711 st->shift_jis.mask = detect_coding_sjis (st, src, n);
1712 if (!mask_has_at_most_one_bit_p (st->big5.mask))
1713 st->big5.mask = detect_coding_big5 (st, src, n);
1714 if (!mask_has_at_most_one_bit_p (st->utf8.mask))
1715 st->utf8.mask = detect_coding_utf8 (st, src, n);
1716 if (!mask_has_at_most_one_bit_p (st->ucs4.mask))
1717 st->ucs4.mask = detect_coding_ucs4 (st, src, n);
1720 = st->iso2022.mask | st->shift_jis.mask | st->big5.mask
1721 | st->utf8.mask | st->ucs4.mask;
1724 int retval = mask_has_at_most_one_bit_p (st->mask);
1725 st->mask |= CODING_CATEGORY_NO_CONVERSION_MASK;
1726 return retval && st->eol_type != EOL_AUTODETECT;
1731 coding_system_from_mask (int mask)
1735 /* If the file was entirely or basically ASCII, use the
1736 default value of `buffer-file-coding-system'. */
1737 Lisp_Object retval =
1738 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system;
1741 retval = Ffind_coding_system (retval);
1745 (Qbad_variable, Qwarning,
1746 "Invalid `default-buffer-file-coding-system', set to nil");
1747 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system = Qnil;
1751 retval = Fget_coding_system (Qraw_text);
1759 mask = postprocess_iso2022_mask (mask);
1761 /* Look through the coding categories by priority and find
1762 the first one that is allowed. */
1763 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
1765 cat = fcd->coding_category_by_priority[i];
1766 if ((mask & (1 << cat)) &&
1767 !NILP (fcd->coding_category_system[cat]))
1771 return fcd->coding_category_system[cat];
1773 return Fget_coding_system (Qraw_text);
1777 /* Given a seekable read stream and potential coding system and EOL type
1778 as specified, do any autodetection that is called for. If the
1779 coding system and/or EOL type are not `autodetect', they will be left
1780 alone; but this function will never return an autodetect coding system
1783 This function does not automatically fetch subsidiary coding systems;
1784 that should be unnecessary with the explicit eol-type argument. */
1786 #define LENGTH(string_constant) (sizeof (string_constant) - 1)
1789 determine_real_coding_system (Lstream *stream, Lisp_Object *codesys_in_out,
1790 eol_type_t *eol_type_in_out)
1792 struct detection_state decst;
1794 if (*eol_type_in_out == EOL_AUTODETECT)
1795 *eol_type_in_out = XCODING_SYSTEM_EOL_TYPE (*codesys_in_out);
1798 decst.eol_type = *eol_type_in_out;
1801 /* If autodetection is called for, do it now. */
1802 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT
1803 || *eol_type_in_out == EOL_AUTODETECT)
1806 Lisp_Object coding_system = Qnil;
1808 ssize_t nread = Lstream_read (stream, buf, sizeof (buf));
1811 /* Look for initial "-*-"; mode line prefix */
1813 scan_end = buf + nread - LENGTH ("-*-coding:?-*-");
1818 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-')
1820 Extbyte *local_vars_beg = p + 3;
1821 /* Look for final "-*-"; mode line suffix */
1822 for (p = local_vars_beg,
1823 scan_end = buf + nread - LENGTH ("-*-");
1828 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-')
1830 Extbyte *suffix = p;
1831 /* Look for "coding:" */
1832 for (p = local_vars_beg,
1833 scan_end = suffix - LENGTH ("coding:?");
1836 if (memcmp ("coding:", p, LENGTH ("coding:")) == 0
1837 && (p == local_vars_beg
1838 || (*(p-1) == ' ' ||
1844 p += LENGTH ("coding:");
1845 while (*p == ' ' || *p == '\t') p++;
1847 /* Get coding system name */
1848 save = *suffix; *suffix = '\0';
1849 /* Characters valid in a MIME charset name (rfc 1521),
1850 and in a Lisp symbol name. */
1851 n = strspn ( (char *) p,
1852 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1853 "abcdefghijklmnopqrstuvwxyz"
1859 save = p[n]; p[n] = '\0';
1861 Ffind_coding_system (intern ((char *) p));
1871 if (NILP (coding_system))
1874 if (detect_coding_type (&decst, buf, nread,
1875 XCODING_SYSTEM_TYPE (*codesys_in_out)
1876 != CODESYS_AUTODETECT))
1878 nread = Lstream_read (stream, buf, sizeof (buf));
1884 else if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT
1885 && XCODING_SYSTEM_EOL_TYPE (coding_system) == EOL_AUTODETECT)
1888 if (detect_coding_type (&decst, buf, nread, 1))
1890 nread = Lstream_read (stream, buf, sizeof (buf));
1896 *eol_type_in_out = decst.eol_type;
1897 if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT)
1899 if (NILP (coding_system))
1900 *codesys_in_out = coding_system_from_mask (decst.mask);
1902 *codesys_in_out = coding_system;
1906 /* If we absolutely can't determine the EOL type, just assume LF. */
1907 if (*eol_type_in_out == EOL_AUTODETECT)
1908 *eol_type_in_out = EOL_LF;
1910 Lstream_rewind (stream);
1913 DEFUN ("detect-coding-region", Fdetect_coding_region, 2, 3, 0, /*
1914 Detect coding system of the text in the region between START and END.
1915 Returned a list of possible coding systems ordered by priority.
1916 If only ASCII characters are found, it returns 'undecided or one of
1917 its subsidiary coding systems according to a detected end-of-line
1918 type. Optional arg BUFFER defaults to the current buffer.
1920 (start, end, buffer))
1922 Lisp_Object val = Qnil;
1923 struct buffer *buf = decode_buffer (buffer, 0);
1925 Lisp_Object instream, lb_instream;
1926 Lstream *istr, *lb_istr;
1927 struct detection_state decst;
1928 struct gcpro gcpro1, gcpro2;
1930 get_buffer_range_char (buf, start, end, &b, &e, 0);
1931 lb_instream = make_lisp_buffer_input_stream (buf, b, e, 0);
1932 lb_istr = XLSTREAM (lb_instream);
1933 instream = make_encoding_input_stream (lb_istr, Fget_coding_system (Qbinary));
1934 istr = XLSTREAM (instream);
1935 GCPRO2 (instream, lb_instream);
1937 decst.eol_type = EOL_AUTODETECT;
1941 unsigned char random_buffer[4096];
1942 ssize_t nread = Lstream_read (istr, random_buffer, sizeof (random_buffer));
1946 if (detect_coding_type (&decst, random_buffer, nread, 0))
1950 if (decst.mask == ~0)
1951 val = subsidiary_coding_system (Fget_coding_system (Qundecided),
1959 decst.mask = postprocess_iso2022_mask (decst.mask);
1961 for (i = CODING_CATEGORY_LAST; i >= 0; i--)
1963 int sys = fcd->coding_category_by_priority[i];
1964 if (decst.mask & (1 << sys))
1966 Lisp_Object codesys = fcd->coding_category_system[sys];
1967 if (!NILP (codesys))
1968 codesys = subsidiary_coding_system (codesys, decst.eol_type);
1969 val = Fcons (codesys, val);
1973 Lstream_close (istr);
1975 Lstream_delete (istr);
1976 Lstream_delete (lb_istr);
1981 /************************************************************************/
1982 /* Converting to internal Mule format ("decoding") */
1983 /************************************************************************/
1985 /* A decoding stream is a stream used for decoding text (i.e.
1986 converting from some external format to internal format).
1987 The decoding-stream object keeps track of the actual coding
1988 stream, the stream that is at the other end, and data that
1989 needs to be persistent across the lifetime of the stream. */
1991 /* Handle the EOL stuff related to just-read-in character C.
1992 EOL_TYPE is the EOL type of the coding stream.
1993 FLAGS is the current value of FLAGS in the coding stream, and may
1994 be modified by this macro. (The macro only looks at the
1995 CODING_STATE_CR flag.) DST is the Dynarr to which the decoded
1996 bytes are to be written. You need to also define a local goto
1997 label "label_continue_loop" that is at the end of the main
1998 character-reading loop.
2000 If C is a CR character, then this macro handles it entirely and
2001 jumps to label_continue_loop. Otherwise, this macro does not add
2002 anything to DST, and continues normally. You should continue
2003 processing C normally after this macro. */
2005 #define DECODE_HANDLE_EOL_TYPE(eol_type, c, flags, dst) \
2009 if (eol_type == EOL_CR) \
2010 Dynarr_add (dst, '\n'); \
2011 else if (eol_type != EOL_CRLF || flags & CODING_STATE_CR) \
2012 Dynarr_add (dst, c); \
2014 flags |= CODING_STATE_CR; \
2015 goto label_continue_loop; \
2017 else if (flags & CODING_STATE_CR) \
2018 { /* eol_type == CODING_SYSTEM_EOL_CRLF */ \
2020 Dynarr_add (dst, '\r'); \
2021 flags &= ~CODING_STATE_CR; \
2025 /* C should be a binary character in the range 0 - 255; convert
2026 to internal format and add to Dynarr DST. */
2028 #define DECODE_ADD_BINARY_CHAR(c, dst) \
2030 if (BYTE_ASCII_P (c)) \
2031 Dynarr_add (dst, c); \
2032 else if (BYTE_C1_P (c)) \
2034 Dynarr_add (dst, LEADING_BYTE_CONTROL_1); \
2035 Dynarr_add (dst, c + 0x20); \
2039 Dynarr_add (dst, LEADING_BYTE_LATIN_ISO8859_1); \
2040 Dynarr_add (dst, c); \
2044 #define DECODE_OUTPUT_PARTIAL_CHAR(ch) \
2048 DECODE_ADD_BINARY_CHAR (ch, dst); \
2053 #define DECODE_HANDLE_END_OF_CONVERSION(flags, ch, dst) \
2055 if (flags & CODING_STATE_END) \
2057 DECODE_OUTPUT_PARTIAL_CHAR (ch); \
2058 if (flags & CODING_STATE_CR) \
2059 Dynarr_add (dst, '\r'); \
2063 #define DECODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, decoding)
2065 struct decoding_stream
2067 /* Coding system that governs the conversion. */
2068 Lisp_Coding_System *codesys;
2070 /* Stream that we read the encoded data from or
2071 write the decoded data to. */
2074 /* If we are reading, then we can return only a fixed amount of
2075 data, so if the conversion resulted in too much data, we store it
2076 here for retrieval the next time around. */
2077 unsigned_char_dynarr *runoff;
2079 /* FLAGS holds flags indicating the current state of the decoding.
2080 Some of these flags are dependent on the coding system. */
2083 /* CH holds a partially built-up character. Since we only deal
2084 with one- and two-byte characters at the moment, we only use
2085 this to store the first byte of a two-byte character. */
2088 /* EOL_TYPE specifies the type of end-of-line conversion that
2089 currently applies. We need to keep this separate from the
2090 EOL type stored in CODESYS because the latter might indicate
2091 automatic EOL-type detection while the former will always
2092 indicate a particular EOL type. */
2093 eol_type_t eol_type;
2095 /* Additional ISO2022 information. We define the structure above
2096 because it's also needed by the detection routines. */
2097 struct iso2022_decoder iso2022;
2099 /* Additional information (the state of the running CCL program)
2100 used by the CCL decoder. */
2101 struct ccl_program ccl;
2103 /* counter for UTF-8 or UCS-4 */
2104 unsigned char counter;
2106 struct detection_state decst;
2109 static ssize_t decoding_reader (Lstream *stream,
2110 unsigned char *data, size_t size);
2111 static ssize_t decoding_writer (Lstream *stream,
2112 const unsigned char *data, size_t size);
2113 static int decoding_rewinder (Lstream *stream);
2114 static int decoding_seekable_p (Lstream *stream);
2115 static int decoding_flusher (Lstream *stream);
2116 static int decoding_closer (Lstream *stream);
2118 static Lisp_Object decoding_marker (Lisp_Object stream);
2120 DEFINE_LSTREAM_IMPLEMENTATION ("decoding", lstream_decoding,
2121 sizeof (struct decoding_stream));
2124 decoding_marker (Lisp_Object stream)
2126 Lstream *str = DECODING_STREAM_DATA (XLSTREAM (stream))->other_end;
2127 Lisp_Object str_obj;
2129 /* We do not need to mark the coding systems or charsets stored
2130 within the stream because they are stored in a global list
2131 and automatically marked. */
2133 XSETLSTREAM (str_obj, str);
2134 mark_object (str_obj);
2135 if (str->imp->marker)
2136 return (str->imp->marker) (str_obj);
2141 /* Read SIZE bytes of data and store it into DATA. We are a decoding stream
2142 so we read data from the other end, decode it, and store it into DATA. */
2145 decoding_reader (Lstream *stream, unsigned char *data, size_t size)
2147 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2148 unsigned char *orig_data = data;
2150 int error_occurred = 0;
2152 /* We need to interface to mule_decode(), which expects to take some
2153 amount of data and store the result into a Dynarr. We have
2154 mule_decode() store into str->runoff, and take data from there
2157 /* We loop until we have enough data, reading chunks from the other
2158 end and decoding it. */
2161 /* Take data from the runoff if we can. Make sure to take at
2162 most SIZE bytes, and delete the data from the runoff. */
2163 if (Dynarr_length (str->runoff) > 0)
2165 size_t chunk = min (size, (size_t) Dynarr_length (str->runoff));
2166 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2167 Dynarr_delete_many (str->runoff, 0, chunk);
2173 break; /* No more room for data */
2175 if (str->flags & CODING_STATE_END)
2176 /* This means that on the previous iteration, we hit the EOF on
2177 the other end. We loop once more so that mule_decode() can
2178 output any final stuff it may be holding, or any "go back
2179 to a sane state" escape sequences. (This latter makes sense
2180 during encoding.) */
2183 /* Exhausted the runoff, so get some more. DATA has at least
2184 SIZE bytes left of storage in it, so it's OK to read directly
2185 into it. (We'll be overwriting above, after we've decoded it
2186 into the runoff.) */
2187 read_size = Lstream_read (str->other_end, data, size);
2194 /* There might be some more end data produced in the translation.
2195 See the comment above. */
2196 str->flags |= CODING_STATE_END;
2197 mule_decode (stream, data, str->runoff, read_size);
2200 if (data - orig_data == 0)
2201 return error_occurred ? -1 : 0;
2203 return data - orig_data;
2207 decoding_writer (Lstream *stream, const unsigned char *data, size_t size)
2209 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2212 /* Decode all our data into the runoff, and then attempt to write
2213 it all out to the other end. Remove whatever chunk we succeeded
2215 mule_decode (stream, data, str->runoff, size);
2216 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
2217 Dynarr_length (str->runoff));
2219 Dynarr_delete_many (str->runoff, 0, retval);
2220 /* Do NOT return retval. The return value indicates how much
2221 of the incoming data was written, not how many bytes were
2227 reset_decoding_stream (struct decoding_stream *str)
2230 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_ISO2022)
2232 Lisp_Object coding_system;
2233 XSETCODING_SYSTEM (coding_system, str->codesys);
2234 reset_iso2022 (coding_system, &str->iso2022);
2236 else if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_CCL)
2238 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_DECODE (str->codesys));
2242 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT
2243 || CODING_SYSTEM_EOL_TYPE (str->codesys) == EOL_AUTODETECT)
2246 str->decst.eol_type = EOL_AUTODETECT;
2247 str->decst.mask = ~0;
2249 str->flags = str->ch = 0;
2253 decoding_rewinder (Lstream *stream)
2255 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2256 reset_decoding_stream (str);
2257 Dynarr_reset (str->runoff);
2258 return Lstream_rewind (str->other_end);
2262 decoding_seekable_p (Lstream *stream)
2264 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2265 return Lstream_seekable_p (str->other_end);
2269 decoding_flusher (Lstream *stream)
2271 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2272 return Lstream_flush (str->other_end);
2276 decoding_closer (Lstream *stream)
2278 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2279 if (stream->flags & LSTREAM_FL_WRITE)
2281 str->flags |= CODING_STATE_END;
2282 decoding_writer (stream, 0, 0);
2284 Dynarr_free (str->runoff);
2286 #ifdef ENABLE_COMPOSITE_CHARS
2287 if (str->iso2022.composite_chars)
2288 Dynarr_free (str->iso2022.composite_chars);
2291 return Lstream_close (str->other_end);
2295 decoding_stream_coding_system (Lstream *stream)
2297 Lisp_Object coding_system;
2298 struct decoding_stream *str = DECODING_STREAM_DATA (stream);
2300 XSETCODING_SYSTEM (coding_system, str->codesys);
2301 return subsidiary_coding_system (coding_system, str->eol_type);
2305 set_decoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
2307 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
2308 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2310 if (CODING_SYSTEM_EOL_TYPE (cs) != EOL_AUTODETECT)
2311 str->eol_type = CODING_SYSTEM_EOL_TYPE (cs);
2312 reset_decoding_stream (str);
2315 /* WARNING WARNING WARNING WARNING!!!!! If you open up a decoding
2316 stream for writing, no automatic code detection will be performed.
2317 The reason for this is that automatic code detection requires a
2318 seekable input. Things will also fail if you open a decoding
2319 stream for reading using a non-fully-specified coding system and
2320 a non-seekable input stream. */
2323 make_decoding_stream_1 (Lstream *stream, Lisp_Object codesys,
2326 Lstream *lstr = Lstream_new (lstream_decoding, mode);
2327 struct decoding_stream *str = DECODING_STREAM_DATA (lstr);
2331 str->other_end = stream;
2332 str->runoff = (unsigned_char_dynarr *) Dynarr_new (unsigned_char);
2333 str->eol_type = EOL_AUTODETECT;
2334 if (!strcmp (mode, "r")
2335 && Lstream_seekable_p (stream))
2336 /* We can determine the coding system now. */
2337 determine_real_coding_system (stream, &codesys, &str->eol_type);
2338 set_decoding_stream_coding_system (lstr, codesys);
2339 str->decst.eol_type = str->eol_type;
2340 str->decst.mask = ~0;
2341 XSETLSTREAM (obj, lstr);
2346 make_decoding_input_stream (Lstream *stream, Lisp_Object codesys)
2348 return make_decoding_stream_1 (stream, codesys, "r");
2352 make_decoding_output_stream (Lstream *stream, Lisp_Object codesys)
2354 return make_decoding_stream_1 (stream, codesys, "w");
2357 /* Note: the decode_coding_* functions all take the same
2358 arguments as mule_decode(), which is to say some SRC data of
2359 size N, which is to be stored into dynamic array DST.
2360 DECODING is the stream within which the decoding is
2361 taking place, but no data is actually read from or
2362 written to that stream; that is handled in decoding_reader()
2363 or decoding_writer(). This allows the same functions to
2364 be used for both reading and writing. */
2367 mule_decode (Lstream *decoding, const unsigned char *src,
2368 unsigned_char_dynarr *dst, unsigned int n)
2370 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
2372 /* If necessary, do encoding-detection now. We do this when
2373 we're a writing stream or a non-seekable reading stream,
2374 meaning that we can't just process the whole input,
2375 rewind, and start over. */
2377 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT ||
2378 str->eol_type == EOL_AUTODETECT)
2380 Lisp_Object codesys;
2382 XSETCODING_SYSTEM (codesys, str->codesys);
2383 detect_coding_type (&str->decst, src, n,
2384 CODING_SYSTEM_TYPE (str->codesys) !=
2385 CODESYS_AUTODETECT);
2386 if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT &&
2387 str->decst.mask != ~0)
2388 /* #### This is cheesy. What we really ought to do is
2389 buffer up a certain amount of data so as to get a
2390 less random result. */
2391 codesys = coding_system_from_mask (str->decst.mask);
2392 str->eol_type = str->decst.eol_type;
2393 if (XCODING_SYSTEM (codesys) != str->codesys)
2395 /* Preserve the CODING_STATE_END flag in case it was set.
2396 If we erase it, bad things might happen. */
2397 int was_end = str->flags & CODING_STATE_END;
2398 set_decoding_stream_coding_system (decoding, codesys);
2400 str->flags |= CODING_STATE_END;
2404 switch (CODING_SYSTEM_TYPE (str->codesys))
2407 case CODESYS_INTERNAL:
2408 Dynarr_add_many (dst, src, n);
2411 case CODESYS_AUTODETECT:
2412 /* If we got this far and still haven't decided on the coding
2413 system, then do no conversion. */
2414 case CODESYS_NO_CONVERSION:
2415 decode_coding_no_conversion (decoding, src, dst, n);
2418 case CODESYS_SHIFT_JIS:
2419 decode_coding_sjis (decoding, src, dst, n);
2422 decode_coding_big5 (decoding, src, dst, n);
2425 decode_coding_ucs4 (decoding, src, dst, n);
2428 decode_coding_utf8 (decoding, src, dst, n);
2431 str->ccl.last_block = str->flags & CODING_STATE_END;
2432 ccl_driver (&str->ccl, src, dst, n, 0, CCL_MODE_DECODING);
2434 case CODESYS_ISO2022:
2435 decode_coding_iso2022 (decoding, src, dst, n);
2443 DEFUN ("decode-coding-region", Fdecode_coding_region, 3, 4, 0, /*
2444 Decode the text between START and END which is encoded in CODING-SYSTEM.
2445 This is useful if you've read in encoded text from a file without decoding
2446 it (e.g. you read in a JIS-formatted file but used the `binary' or
2447 `no-conversion' coding system, so that it shows up as "^[$B!<!+^[(B").
2448 Return length of decoded text.
2449 BUFFER defaults to the current buffer if unspecified.
2451 (start, end, coding_system, buffer))
2454 struct buffer *buf = decode_buffer (buffer, 0);
2455 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2456 Lstream *istr, *ostr;
2457 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2459 get_buffer_range_char (buf, start, end, &b, &e, 0);
2461 barf_if_buffer_read_only (buf, b, e);
2463 coding_system = Fget_coding_system (coding_system);
2464 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2465 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
2466 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
2468 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
2469 Fget_coding_system (Qbinary));
2470 istr = XLSTREAM (instream);
2471 ostr = XLSTREAM (outstream);
2472 GCPRO4 (instream, lb_outstream, de_outstream, outstream);
2474 /* The chain of streams looks like this:
2476 [BUFFER] <----- send through
2477 ------> [ENCODE AS BINARY]
2478 ------> [DECODE AS SPECIFIED]
2484 char tempbuf[1024]; /* some random amount */
2485 Bufpos newpos, even_newer_pos;
2486 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
2487 ssize_t size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
2491 newpos = lisp_buffer_stream_startpos (istr);
2492 Lstream_write (ostr, tempbuf, size_in_bytes);
2493 even_newer_pos = lisp_buffer_stream_startpos (istr);
2494 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
2497 Lstream_close (istr);
2498 Lstream_close (ostr);
2500 Lstream_delete (istr);
2501 Lstream_delete (ostr);
2502 Lstream_delete (XLSTREAM (de_outstream));
2503 Lstream_delete (XLSTREAM (lb_outstream));
2508 /************************************************************************/
2509 /* Converting to an external encoding ("encoding") */
2510 /************************************************************************/
2512 /* An encoding stream is an output stream. When you create the
2513 stream, you specify the coding system that governs the encoding
2514 and another stream that the resulting encoded data is to be
2515 sent to, and then start sending data to it. */
2517 #define ENCODING_STREAM_DATA(stream) LSTREAM_TYPE_DATA (stream, encoding)
2519 struct encoding_stream
2521 /* Coding system that governs the conversion. */
2522 Lisp_Coding_System *codesys;
2524 /* Stream that we read the encoded data from or
2525 write the decoded data to. */
2528 /* If we are reading, then we can return only a fixed amount of
2529 data, so if the conversion resulted in too much data, we store it
2530 here for retrieval the next time around. */
2531 unsigned_char_dynarr *runoff;
2533 /* FLAGS holds flags indicating the current state of the encoding.
2534 Some of these flags are dependent on the coding system. */
2537 /* CH holds a partially built-up character. Since we only deal
2538 with one- and two-byte characters at the moment, we only use
2539 this to store the first byte of a two-byte character. */
2542 /* Additional information used by the ISO2022 encoder. */
2545 /* CHARSET holds the character sets currently assigned to the G0
2546 through G3 registers. It is initialized from the array
2547 INITIAL_CHARSET in CODESYS. */
2548 Lisp_Object charset[4];
2550 /* Which registers are currently invoked into the left (GL) and
2551 right (GR) halves of the 8-bit encoding space? */
2552 int register_left, register_right;
2554 /* Whether we need to explicitly designate the charset in the
2555 G? register before using it. It is initialized from the
2556 array FORCE_CHARSET_ON_OUTPUT in CODESYS. */
2557 unsigned char force_charset_on_output[4];
2559 /* Other state variables that need to be preserved across
2561 Lisp_Object current_charset;
2563 int current_char_boundary;
2566 /* Additional information (the state of the running CCL program)
2567 used by the CCL encoder. */
2568 struct ccl_program ccl;
2572 static ssize_t encoding_reader (Lstream *stream, unsigned char *data, size_t size);
2573 static ssize_t encoding_writer (Lstream *stream, const unsigned char *data,
2575 static int encoding_rewinder (Lstream *stream);
2576 static int encoding_seekable_p (Lstream *stream);
2577 static int encoding_flusher (Lstream *stream);
2578 static int encoding_closer (Lstream *stream);
2580 static Lisp_Object encoding_marker (Lisp_Object stream);
2582 DEFINE_LSTREAM_IMPLEMENTATION ("encoding", lstream_encoding,
2583 sizeof (struct encoding_stream));
2586 encoding_marker (Lisp_Object stream)
2588 Lstream *str = ENCODING_STREAM_DATA (XLSTREAM (stream))->other_end;
2589 Lisp_Object str_obj;
2591 /* We do not need to mark the coding systems or charsets stored
2592 within the stream because they are stored in a global list
2593 and automatically marked. */
2595 XSETLSTREAM (str_obj, str);
2596 mark_object (str_obj);
2597 if (str->imp->marker)
2598 return (str->imp->marker) (str_obj);
2603 /* Read SIZE bytes of data and store it into DATA. We are a encoding stream
2604 so we read data from the other end, encode it, and store it into DATA. */
2607 encoding_reader (Lstream *stream, unsigned char *data, size_t size)
2609 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2610 unsigned char *orig_data = data;
2612 int error_occurred = 0;
2614 /* We need to interface to mule_encode(), which expects to take some
2615 amount of data and store the result into a Dynarr. We have
2616 mule_encode() store into str->runoff, and take data from there
2619 /* We loop until we have enough data, reading chunks from the other
2620 end and encoding it. */
2623 /* Take data from the runoff if we can. Make sure to take at
2624 most SIZE bytes, and delete the data from the runoff. */
2625 if (Dynarr_length (str->runoff) > 0)
2627 int chunk = min ((int) size, Dynarr_length (str->runoff));
2628 memcpy (data, Dynarr_atp (str->runoff, 0), chunk);
2629 Dynarr_delete_many (str->runoff, 0, chunk);
2635 break; /* No more room for data */
2637 if (str->flags & CODING_STATE_END)
2638 /* This means that on the previous iteration, we hit the EOF on
2639 the other end. We loop once more so that mule_encode() can
2640 output any final stuff it may be holding, or any "go back
2641 to a sane state" escape sequences. (This latter makes sense
2642 during encoding.) */
2645 /* Exhausted the runoff, so get some more. DATA at least SIZE bytes
2646 left of storage in it, so it's OK to read directly into it.
2647 (We'll be overwriting above, after we've encoded it into the
2649 read_size = Lstream_read (str->other_end, data, size);
2656 /* There might be some more end data produced in the translation.
2657 See the comment above. */
2658 str->flags |= CODING_STATE_END;
2659 mule_encode (stream, data, str->runoff, read_size);
2662 if (data == orig_data)
2663 return error_occurred ? -1 : 0;
2665 return data - orig_data;
2669 encoding_writer (Lstream *stream, const unsigned char *data, size_t size)
2671 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2674 /* Encode all our data into the runoff, and then attempt to write
2675 it all out to the other end. Remove whatever chunk we succeeded
2677 mule_encode (stream, data, str->runoff, size);
2678 retval = Lstream_write (str->other_end, Dynarr_atp (str->runoff, 0),
2679 Dynarr_length (str->runoff));
2681 Dynarr_delete_many (str->runoff, 0, retval);
2682 /* Do NOT return retval. The return value indicates how much
2683 of the incoming data was written, not how many bytes were
2689 reset_encoding_stream (struct encoding_stream *str)
2692 switch (CODING_SYSTEM_TYPE (str->codesys))
2694 case CODESYS_ISO2022:
2698 for (i = 0; i < 4; i++)
2700 str->iso2022.charset[i] =
2701 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, i);
2702 str->iso2022.force_charset_on_output[i] =
2703 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (str->codesys, i);
2705 str->iso2022.register_left = 0;
2706 str->iso2022.register_right = 1;
2707 str->iso2022.current_charset = Qnil;
2708 str->iso2022.current_half = 0;
2709 str->iso2022.current_char_boundary = 1;
2713 setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_ENCODE (str->codesys));
2720 str->flags = str->ch = 0;
2724 encoding_rewinder (Lstream *stream)
2726 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2727 reset_encoding_stream (str);
2728 Dynarr_reset (str->runoff);
2729 return Lstream_rewind (str->other_end);
2733 encoding_seekable_p (Lstream *stream)
2735 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2736 return Lstream_seekable_p (str->other_end);
2740 encoding_flusher (Lstream *stream)
2742 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2743 return Lstream_flush (str->other_end);
2747 encoding_closer (Lstream *stream)
2749 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2750 if (stream->flags & LSTREAM_FL_WRITE)
2752 str->flags |= CODING_STATE_END;
2753 encoding_writer (stream, 0, 0);
2755 Dynarr_free (str->runoff);
2756 return Lstream_close (str->other_end);
2760 encoding_stream_coding_system (Lstream *stream)
2762 Lisp_Object coding_system;
2763 struct encoding_stream *str = ENCODING_STREAM_DATA (stream);
2765 XSETCODING_SYSTEM (coding_system, str->codesys);
2766 return coding_system;
2770 set_encoding_stream_coding_system (Lstream *lstr, Lisp_Object codesys)
2772 Lisp_Coding_System *cs = XCODING_SYSTEM (codesys);
2773 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
2775 reset_encoding_stream (str);
2779 make_encoding_stream_1 (Lstream *stream, Lisp_Object codesys,
2782 Lstream *lstr = Lstream_new (lstream_encoding, mode);
2783 struct encoding_stream *str = ENCODING_STREAM_DATA (lstr);
2787 str->runoff = Dynarr_new (unsigned_char);
2788 str->other_end = stream;
2789 set_encoding_stream_coding_system (lstr, codesys);
2790 XSETLSTREAM (obj, lstr);
2795 make_encoding_input_stream (Lstream *stream, Lisp_Object codesys)
2797 return make_encoding_stream_1 (stream, codesys, "r");
2801 make_encoding_output_stream (Lstream *stream, Lisp_Object codesys)
2803 return make_encoding_stream_1 (stream, codesys, "w");
2806 /* Convert N bytes of internally-formatted data stored in SRC to an
2807 external format, according to the encoding stream ENCODING.
2808 Store the encoded data into DST. */
2811 mule_encode (Lstream *encoding, const unsigned char *src,
2812 unsigned_char_dynarr *dst, unsigned int n)
2814 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
2816 switch (CODING_SYSTEM_TYPE (str->codesys))
2819 case CODESYS_INTERNAL:
2820 Dynarr_add_many (dst, src, n);
2823 case CODESYS_AUTODETECT:
2824 /* If we got this far and still haven't decided on the coding
2825 system, then do no conversion. */
2826 case CODESYS_NO_CONVERSION:
2827 encode_coding_no_conversion (encoding, src, dst, n);
2830 case CODESYS_SHIFT_JIS:
2831 encode_coding_sjis (encoding, src, dst, n);
2834 encode_coding_big5 (encoding, src, dst, n);
2837 encode_coding_ucs4 (encoding, src, dst, n);
2840 encode_coding_utf8 (encoding, src, dst, n);
2843 str->ccl.last_block = str->flags & CODING_STATE_END;
2844 ccl_driver (&str->ccl, src, dst, n, 0, CCL_MODE_ENCODING);
2846 case CODESYS_ISO2022:
2847 encode_coding_iso2022 (encoding, src, dst, n);
2855 DEFUN ("encode-coding-region", Fencode_coding_region, 3, 4, 0, /*
2856 Encode the text between START and END using CODING-SYSTEM.
2857 This will, for example, convert Japanese characters into stuff such as
2858 "^[$B!<!+^[(B" if you use the JIS encoding. Return length of encoded
2859 text. BUFFER defaults to the current buffer if unspecified.
2861 (start, end, coding_system, buffer))
2864 struct buffer *buf = decode_buffer (buffer, 0);
2865 Lisp_Object instream, lb_outstream, de_outstream, outstream;
2866 Lstream *istr, *ostr;
2867 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4;
2869 get_buffer_range_char (buf, start, end, &b, &e, 0);
2871 barf_if_buffer_read_only (buf, b, e);
2873 coding_system = Fget_coding_system (coding_system);
2874 instream = make_lisp_buffer_input_stream (buf, b, e, 0);
2875 lb_outstream = make_lisp_buffer_output_stream (buf, b, 0);
2876 de_outstream = make_decoding_output_stream (XLSTREAM (lb_outstream),
2877 Fget_coding_system (Qbinary));
2878 outstream = make_encoding_output_stream (XLSTREAM (de_outstream),
2880 istr = XLSTREAM (instream);
2881 ostr = XLSTREAM (outstream);
2882 GCPRO4 (instream, outstream, de_outstream, lb_outstream);
2883 /* The chain of streams looks like this:
2885 [BUFFER] <----- send through
2886 ------> [ENCODE AS SPECIFIED]
2887 ------> [DECODE AS BINARY]
2892 char tempbuf[1024]; /* some random amount */
2893 Bufpos newpos, even_newer_pos;
2894 Bufpos oldpos = lisp_buffer_stream_startpos (istr);
2895 ssize_t size_in_bytes = Lstream_read (istr, tempbuf, sizeof (tempbuf));
2899 newpos = lisp_buffer_stream_startpos (istr);
2900 Lstream_write (ostr, tempbuf, size_in_bytes);
2901 even_newer_pos = lisp_buffer_stream_startpos (istr);
2902 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos),
2908 lisp_buffer_stream_startpos (XLSTREAM (instream)) - b;
2909 Lstream_close (istr);
2910 Lstream_close (ostr);
2912 Lstream_delete (istr);
2913 Lstream_delete (ostr);
2914 Lstream_delete (XLSTREAM (de_outstream));
2915 Lstream_delete (XLSTREAM (lb_outstream));
2916 return make_int (retlen);
2922 /************************************************************************/
2923 /* Shift-JIS methods */
2924 /************************************************************************/
2926 /* Shift-JIS is a coding system encoding three character sets: ASCII, right
2927 half of JISX0201-Kana, and JISX0208. An ASCII character is encoded
2928 as is. A character of JISX0201-Kana (DIMENSION1_CHARS94 character set) is
2929 encoded by "position-code + 0x80". A character of JISX0208
2930 (DIMENSION2_CHARS94 character set) is encoded in 2-byte but two
2931 position-codes are divided and shifted so that it fit in the range
2934 --- CODE RANGE of Shift-JIS ---
2935 (character set) (range)
2937 JISX0201-Kana 0xA0 .. 0xDF
2938 JISX0208 (1st byte) 0x80 .. 0x9F and 0xE0 .. 0xEF
2939 (2nd byte) 0x40 .. 0x7E and 0x80 .. 0xFC
2940 -------------------------------
2944 /* Is this the first byte of a Shift-JIS two-byte char? */
2946 #define BYTE_SJIS_TWO_BYTE_1_P(c) \
2947 (((c) >= 0x81 && (c) <= 0x9F) || ((c) >= 0xE0 && (c) <= 0xEF))
2949 /* Is this the second byte of a Shift-JIS two-byte char? */
2951 #define BYTE_SJIS_TWO_BYTE_2_P(c) \
2952 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0x80 && (c) <= 0xFC))
2954 #define BYTE_SJIS_KATAKANA_P(c) \
2955 ((c) >= 0xA1 && (c) <= 0xDF)
2958 detect_coding_sjis (struct detection_state *st, const unsigned char *src,
2966 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
2968 if (st->shift_jis.in_second_byte)
2970 st->shift_jis.in_second_byte = 0;
2974 else if ((c >= 0x80 && c < 0xA0) || c >= 0xE0)
2975 st->shift_jis.in_second_byte = 1;
2977 return CODING_CATEGORY_SHIFT_JIS_MASK;
2980 /* Convert Shift-JIS data to internal format. */
2983 decode_coding_sjis (Lstream *decoding, const unsigned char *src,
2984 unsigned_char_dynarr *dst, unsigned int n)
2987 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
2988 unsigned int flags = str->flags;
2989 unsigned int ch = str->ch;
2990 eol_type_t eol_type = str->eol_type;
2998 /* Previous character was first byte of Shift-JIS Kanji char. */
2999 if (BYTE_SJIS_TWO_BYTE_2_P (c))
3001 unsigned char e1, e2;
3003 Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208);
3004 DECODE_SJIS (ch, c, e1, e2);
3005 Dynarr_add (dst, e1);
3006 Dynarr_add (dst, e2);
3010 DECODE_ADD_BINARY_CHAR (ch, dst);
3011 DECODE_ADD_BINARY_CHAR (c, dst);
3017 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3018 if (BYTE_SJIS_TWO_BYTE_1_P (c))
3020 else if (BYTE_SJIS_KATAKANA_P (c))
3022 Dynarr_add (dst, LEADING_BYTE_KATAKANA_JISX0201);
3023 Dynarr_add (dst, c);
3026 DECODE_ADD_BINARY_CHAR (c, dst);
3028 label_continue_loop:;
3031 DECODE_HANDLE_END_OF_CONVERSION (flags, ch, dst);
3037 /* Convert internally-formatted data to Shift-JIS. */
3040 encode_coding_sjis (Lstream *encoding, const unsigned char *src,
3041 unsigned_char_dynarr *dst, unsigned int n)
3044 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3045 unsigned int flags = str->flags;
3046 unsigned int ch = str->ch;
3047 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3054 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3055 Dynarr_add (dst, '\r');
3056 if (eol_type != EOL_CR)
3057 Dynarr_add (dst, '\n');
3060 else if (BYTE_ASCII_P (c))
3062 Dynarr_add (dst, c);
3065 else if (BUFBYTE_LEADING_BYTE_P (c))
3066 ch = (c == LEADING_BYTE_KATAKANA_JISX0201 ||
3067 c == LEADING_BYTE_JAPANESE_JISX0208_1978 ||
3068 c == LEADING_BYTE_JAPANESE_JISX0208) ? c : 0;
3071 if (ch == LEADING_BYTE_KATAKANA_JISX0201)
3073 Dynarr_add (dst, c);
3076 else if (ch == LEADING_BYTE_JAPANESE_JISX0208_1978 ||
3077 ch == LEADING_BYTE_JAPANESE_JISX0208)
3081 unsigned char j1, j2;
3082 ENCODE_SJIS (ch, c, j1, j2);
3083 Dynarr_add (dst, j1);
3084 Dynarr_add (dst, j2);
3094 DEFUN ("decode-shift-jis-char", Fdecode_shift_jis_char, 1, 1, 0, /*
3095 Decode a JISX0208 character of Shift-JIS coding-system.
3096 CODE is the character code in Shift-JIS as a cons of type bytes.
3097 Return the corresponding character.
3101 unsigned char c1, c2, s1, s2;
3104 CHECK_INT (XCAR (code));
3105 CHECK_INT (XCDR (code));
3106 s1 = XINT (XCAR (code));
3107 s2 = XINT (XCDR (code));
3108 if (BYTE_SJIS_TWO_BYTE_1_P (s1) &&
3109 BYTE_SJIS_TWO_BYTE_2_P (s2))
3111 DECODE_SJIS (s1, s2, c1, c2);
3112 return make_char (MAKE_CHAR (Vcharset_japanese_jisx0208,
3113 c1 & 0x7F, c2 & 0x7F));
3119 DEFUN ("encode-shift-jis-char", Fencode_shift_jis_char, 1, 1, 0, /*
3120 Encode a JISX0208 character CHAR to SHIFT-JIS coding-system.
3121 Return the corresponding character code in SHIFT-JIS as a cons of two bytes.
3125 Lisp_Object charset;
3128 CHECK_CHAR_COERCE_INT (ch);
3129 BREAKUP_CHAR (XCHAR (ch), charset, c1, c2);
3130 if (EQ (charset, Vcharset_japanese_jisx0208))
3132 ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
3133 return Fcons (make_int (s1), make_int (s2));
3140 /************************************************************************/
3142 /************************************************************************/
3144 /* BIG5 is a coding system encoding two character sets: ASCII and
3145 Big5. An ASCII character is encoded as is. Big5 is a two-byte
3146 character set and is encoded in two-byte.
3148 --- CODE RANGE of BIG5 ---
3149 (character set) (range)
3151 Big5 (1st byte) 0xA1 .. 0xFE
3152 (2nd byte) 0x40 .. 0x7E and 0xA1 .. 0xFE
3153 --------------------------
3155 Since the number of characters in Big5 is larger than maximum
3156 characters in Emacs' charset (96x96), it can't be handled as one
3157 charset. So, in Emacs, Big5 is divided into two: `charset-big5-1'
3158 and `charset-big5-2'. Both <type>s are DIMENSION2_CHARS94. The former
3159 contains frequently used characters and the latter contains less
3160 frequently used characters. */
3162 #define BYTE_BIG5_TWO_BYTE_1_P(c) \
3163 ((c) >= 0xA1 && (c) <= 0xFE)
3165 /* Is this the second byte of a Shift-JIS two-byte char? */
3167 #define BYTE_BIG5_TWO_BYTE_2_P(c) \
3168 (((c) >= 0x40 && (c) <= 0x7E) || ((c) >= 0xA1 && (c) <= 0xFE))
3170 /* Number of Big5 characters which have the same code in 1st byte. */
3172 #define BIG5_SAME_ROW (0xFF - 0xA1 + 0x7F - 0x40)
3174 /* Code conversion macros. These are macros because they are used in
3175 inner loops during code conversion.
3177 Note that temporary variables in macros introduce the classic
3178 dynamic-scoping problems with variable names. We use capital-
3179 lettered variables in the assumption that XEmacs does not use
3180 capital letters in variables except in a very formalized way
3183 /* Convert Big5 code (b1, b2) into its internal string representation
3186 /* There is a much simpler way to split the Big5 charset into two.
3187 For the moment I'm going to leave the algorithm as-is because it
3188 claims to separate out the most-used characters into a single
3189 charset, which perhaps will lead to optimizations in various
3192 The way the algorithm works is something like this:
3194 Big5 can be viewed as a 94x157 charset, where the row is
3195 encoded into the bytes 0xA1 .. 0xFE and the column is encoded
3196 into the bytes 0x40 .. 0x7E and 0xA1 .. 0xFE. As for frequency,
3197 the split between low and high column numbers is apparently
3198 meaningless; ascending rows produce less and less frequent chars.
3199 Therefore, we assign the lower half of rows (0xA1 .. 0xC8) to
3200 the first charset, and the upper half (0xC9 .. 0xFE) to the
3201 second. To do the conversion, we convert the character into
3202 a single number where 0 .. 156 is the first row, 157 .. 313
3203 is the second, etc. That way, the characters are ordered by
3204 decreasing frequency. Then we just chop the space in two
3205 and coerce the result into a 94x94 space.
3208 #define DECODE_BIG5(b1, b2, lb, c1, c2) do \
3210 int B1 = b1, B2 = b2; \
3212 = (B1 - 0xA1) * BIG5_SAME_ROW + B2 - (B2 < 0x7F ? 0x40 : 0x62); \
3216 lb = LEADING_BYTE_CHINESE_BIG5_1; \
3220 lb = LEADING_BYTE_CHINESE_BIG5_2; \
3221 I -= (BIG5_SAME_ROW) * (0xC9 - 0xA1); \
3223 c1 = I / (0xFF - 0xA1) + 0xA1; \
3224 c2 = I % (0xFF - 0xA1) + 0xA1; \
3227 /* Convert the internal string representation of a Big5 character
3228 (lb, c1, c2) into Big5 code (b1, b2). */
3230 #define ENCODE_BIG5(lb, c1, c2, b1, b2) do \
3232 unsigned int I = ((c1) - 0xA1) * (0xFF - 0xA1) + ((c2) - 0xA1); \
3234 if (lb == LEADING_BYTE_CHINESE_BIG5_2) \
3236 I += BIG5_SAME_ROW * (0xC9 - 0xA1); \
3238 b1 = I / BIG5_SAME_ROW + 0xA1; \
3239 b2 = I % BIG5_SAME_ROW; \
3240 b2 += b2 < 0x3F ? 0x40 : 0x62; \
3244 detect_coding_big5 (struct detection_state *st, const unsigned char *src,
3252 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO ||
3253 (c >= 0x80 && c <= 0xA0))
3255 if (st->big5.in_second_byte)
3257 st->big5.in_second_byte = 0;
3258 if (c < 0x40 || (c >= 0x80 && c <= 0xA0))
3262 st->big5.in_second_byte = 1;
3264 return CODING_CATEGORY_BIG5_MASK;
3267 /* Convert Big5 data to internal format. */
3270 decode_coding_big5 (Lstream *decoding, const unsigned char *src,
3271 unsigned_char_dynarr *dst, unsigned int n)
3274 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3275 unsigned int flags = str->flags;
3276 unsigned int ch = str->ch;
3277 eol_type_t eol_type = str->eol_type;
3284 /* Previous character was first byte of Big5 char. */
3285 if (BYTE_BIG5_TWO_BYTE_2_P (c))
3287 unsigned char b1, b2, b3;
3288 DECODE_BIG5 (ch, c, b1, b2, b3);
3289 Dynarr_add (dst, b1);
3290 Dynarr_add (dst, b2);
3291 Dynarr_add (dst, b3);
3295 DECODE_ADD_BINARY_CHAR (ch, dst);
3296 DECODE_ADD_BINARY_CHAR (c, dst);
3302 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3303 if (BYTE_BIG5_TWO_BYTE_1_P (c))
3306 DECODE_ADD_BINARY_CHAR (c, dst);
3308 label_continue_loop:;
3311 DECODE_HANDLE_END_OF_CONVERSION (flags, ch, dst);
3317 /* Convert internally-formatted data to Big5. */
3320 encode_coding_big5 (Lstream *encoding, const unsigned char *src,
3321 unsigned_char_dynarr *dst, unsigned int n)
3324 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3325 unsigned int flags = str->flags;
3326 unsigned int ch = str->ch;
3327 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3334 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3335 Dynarr_add (dst, '\r');
3336 if (eol_type != EOL_CR)
3337 Dynarr_add (dst, '\n');
3339 else if (BYTE_ASCII_P (c))
3342 Dynarr_add (dst, c);
3344 else if (BUFBYTE_LEADING_BYTE_P (c))
3346 if (c == LEADING_BYTE_CHINESE_BIG5_1 ||
3347 c == LEADING_BYTE_CHINESE_BIG5_2)
3349 /* A recognized leading byte. */
3351 continue; /* not done with this character. */
3353 /* otherwise just ignore this character. */
3355 else if (ch == LEADING_BYTE_CHINESE_BIG5_1 ||
3356 ch == LEADING_BYTE_CHINESE_BIG5_2)
3358 /* Previous char was a recognized leading byte. */
3360 continue; /* not done with this character. */
3364 /* Encountering second byte of a Big5 character. */
3365 unsigned char b1, b2;
3367 ENCODE_BIG5 (ch >> 8, ch & 0xFF, c, b1, b2);
3368 Dynarr_add (dst, b1);
3369 Dynarr_add (dst, b2);
3380 DEFUN ("decode-big5-char", Fdecode_big5_char, 1, 1, 0, /*
3381 Decode a Big5 character CODE of BIG5 coding-system.
3382 CODE is the character code in BIG5, a cons of two integers.
3383 Return the corresponding character.
3387 unsigned char c1, c2, b1, b2;
3390 CHECK_INT (XCAR (code));
3391 CHECK_INT (XCDR (code));
3392 b1 = XINT (XCAR (code));
3393 b2 = XINT (XCDR (code));
3394 if (BYTE_BIG5_TWO_BYTE_1_P (b1) &&
3395 BYTE_BIG5_TWO_BYTE_2_P (b2))
3398 Lisp_Object charset;
3399 DECODE_BIG5 (b1, b2, leading_byte, c1, c2);
3400 charset = CHARSET_BY_LEADING_BYTE (leading_byte);
3401 return make_char (MAKE_CHAR (charset, c1 & 0x7F, c2 & 0x7F));
3407 DEFUN ("encode-big5-char", Fencode_big5_char, 1, 1, 0, /*
3408 Encode the Big5 character CH to BIG5 coding-system.
3409 Return the corresponding character code in Big5.
3413 Lisp_Object charset;
3416 CHECK_CHAR_COERCE_INT (ch);
3417 BREAKUP_CHAR (XCHAR (ch), charset, c1, c2);
3418 if (EQ (charset, Vcharset_chinese_big5_1) ||
3419 EQ (charset, Vcharset_chinese_big5_2))
3421 ENCODE_BIG5 (XCHARSET_LEADING_BYTE (charset), c1 | 0x80, c2 | 0x80,
3423 return Fcons (make_int (b1), make_int (b2));
3430 /************************************************************************/
3433 /* UCS-4 character codes are implemented as nonnegative integers. */
3435 /************************************************************************/
3438 DEFUN ("set-ucs-char", Fset_ucs_char, 2, 2, 0, /*
3439 Map UCS-4 code CODE to Mule character CHARACTER.
3441 Return T on success, NIL on failure.
3447 CHECK_CHAR (character);
3451 if (c < sizeof (fcd->ucs_to_mule_table))
3453 fcd->ucs_to_mule_table[c] = character;
3461 ucs_to_char (unsigned long code)
3463 if (code < sizeof (fcd->ucs_to_mule_table))
3465 return fcd->ucs_to_mule_table[code];
3467 else if ((0xe00000 <= code) && (code <= 0xe00000 + 94 * 94 * 14))
3472 c = code % (94 * 94);
3474 (MAKE_CHAR (CHARSET_BY_ATTRIBUTES
3475 (CHARSET_TYPE_94X94, code / (94 * 94) + '@',
3476 CHARSET_LEFT_TO_RIGHT),
3477 c / 94 + 33, c % 94 + 33));
3483 DEFUN ("ucs-char", Fucs_char, 1, 1, 0, /*
3484 Return Mule character corresponding to UCS code CODE (a positive integer).
3488 CHECK_NATNUM (code);
3489 return ucs_to_char (XINT (code));
3492 DEFUN ("set-char-ucs", Fset_char_ucs, 2, 2, 0, /*
3493 Map Mule character CHARACTER to UCS code CODE (a positive integer).
3497 /* #### Isn't this gilding the lily? Fput_char_table checks its args.
3498 Fset_char_ucs is more restrictive on index arg, but should
3499 check code arg in a char_table method. */
3500 CHECK_CHAR (character);
3501 CHECK_NATNUM (code);
3502 return Fput_char_table (character, code, mule_to_ucs_table);
3505 DEFUN ("char-ucs", Fchar_ucs, 1, 1, 0, /*
3506 Return the UCS code (a positive integer) corresponding to CHARACTER.
3510 return Fget_char_table (character, mule_to_ucs_table);
3513 /* Decode a UCS-4 character into a buffer. If the lookup fails, use
3514 <GETA MARK> (U+3013) of JIS X 0208, which means correct character
3515 is not found, instead.
3516 #### do something more appropriate (use blob?)
3517 Danger, Will Robinson! Data loss. Should we signal user? */
3519 decode_ucs4 (unsigned long ch, unsigned_char_dynarr *dst)
3521 Lisp_Object chr = ucs_to_char (ch);
3525 Bufbyte work[MAX_EMCHAR_LEN];
3530 simple_set_charptr_emchar (work, ch) :
3531 non_ascii_set_charptr_emchar (work, ch);
3532 Dynarr_add_many (dst, work, len);
3536 Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208);
3537 Dynarr_add (dst, 34 + 128);
3538 Dynarr_add (dst, 46 + 128);
3542 static unsigned long
3543 mule_char_to_ucs4 (Lisp_Object charset,
3544 unsigned char h, unsigned char l)
3547 = Fget_char_table (make_char (MAKE_CHAR (charset, h & 127, l & 127)),
3554 else if ( (XCHARSET_DIMENSION (charset) == 2) &&
3555 (XCHARSET_CHARS (charset) == 94) )
3557 unsigned char final = XCHARSET_FINAL (charset);
3559 if ( ('@' <= final) && (final < 0x7f) )
3561 return 0xe00000 + (final - '@') * 94 * 94
3562 + ((h & 127) - 33) * 94 + (l & 127) - 33;
3576 encode_ucs4 (Lisp_Object charset,
3577 unsigned char h, unsigned char l, unsigned_char_dynarr *dst)
3579 unsigned long code = mule_char_to_ucs4 (charset, h, l);
3580 Dynarr_add (dst, code >> 24);
3581 Dynarr_add (dst, (code >> 16) & 255);
3582 Dynarr_add (dst, (code >> 8) & 255);
3583 Dynarr_add (dst, code & 255);
3587 detect_coding_ucs4 (struct detection_state *st, const unsigned char *src,
3593 switch (st->ucs4.in_byte)
3602 st->ucs4.in_byte = 0;
3608 return CODING_CATEGORY_UCS4_MASK;
3612 decode_coding_ucs4 (Lstream *decoding, const unsigned char *src,
3613 unsigned_char_dynarr *dst, unsigned int n)
3615 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3616 unsigned int flags = str->flags;
3617 unsigned int ch = str->ch;
3618 unsigned char counter = str->counter;
3622 unsigned char c = *src++;
3630 decode_ucs4 ( ( ch << 8 ) | c, dst);
3635 ch = ( ch << 8 ) | c;
3639 if (counter & CODING_STATE_END)
3640 DECODE_OUTPUT_PARTIAL_CHAR (ch);
3644 str->counter = counter;
3648 encode_coding_ucs4 (Lstream *encoding, const unsigned char *src,
3649 unsigned_char_dynarr *dst, unsigned int n)
3651 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3652 unsigned int flags = str->flags;
3653 unsigned int ch = str->ch;
3654 unsigned char char_boundary = str->iso2022.current_char_boundary;
3655 Lisp_Object charset = str->iso2022.current_charset;
3657 #ifdef ENABLE_COMPOSITE_CHARS
3658 /* flags for handling composite chars. We do a little switcharoo
3659 on the source while we're outputting the composite char. */
3660 unsigned int saved_n = 0;
3661 const unsigned char *saved_src = NULL;
3662 int in_composite = 0;
3669 unsigned char c = *src++;
3671 if (BYTE_ASCII_P (c))
3672 { /* Processing ASCII character */
3674 encode_ucs4 (Vcharset_ascii, c, 0, dst);
3677 else if (BUFBYTE_LEADING_BYTE_P (c) || BUFBYTE_LEADING_BYTE_P (ch))
3678 { /* Processing Leading Byte */
3680 charset = CHARSET_BY_LEADING_BYTE (c);
3681 if (LEADING_BYTE_PREFIX_P(c))
3686 { /* Processing Non-ASCII character */
3688 if (EQ (charset, Vcharset_control_1))
3690 encode_ucs4 (Vcharset_control_1, c, 0, dst);
3694 switch (XCHARSET_REP_BYTES (charset))
3697 encode_ucs4 (charset, c, 0, dst);
3700 if (XCHARSET_PRIVATE_P (charset))
3702 encode_ucs4 (charset, c, 0, dst);
3707 #ifdef ENABLE_COMPOSITE_CHARS
3708 if (EQ (charset, Vcharset_composite))
3712 /* #### Bother! We don't know how to
3714 Dynarr_add (dst, 0);
3715 Dynarr_add (dst, 0);
3716 Dynarr_add (dst, 0);
3717 Dynarr_add (dst, '~');
3721 Emchar emch = MAKE_CHAR (Vcharset_composite,
3722 ch & 0x7F, c & 0x7F);
3723 Lisp_Object lstr = composite_char_string (emch);
3727 src = XSTRING_DATA (lstr);
3728 n = XSTRING_LENGTH (lstr);
3732 #endif /* ENABLE_COMPOSITE_CHARS */
3734 encode_ucs4(charset, ch, c, dst);
3747 encode_ucs4 (charset, ch, c, dst);
3763 #ifdef ENABLE_COMPOSITE_CHARS
3769 goto back_to_square_n; /* Wheeeeeeeee ..... */
3771 #endif /* ENABLE_COMPOSITE_CHARS */
3775 str->iso2022.current_char_boundary = char_boundary;
3776 str->iso2022.current_charset = charset;
3778 /* Verbum caro factum est! */
3782 /************************************************************************/
3784 /************************************************************************/
3787 detect_coding_utf8 (struct detection_state *st, const unsigned char *src,
3792 unsigned char c = *src++;
3793 switch (st->utf8.in_byte)
3796 if (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
3799 st->utf8.in_byte = 5;
3801 st->utf8.in_byte = 4;
3803 st->utf8.in_byte = 3;
3805 st->utf8.in_byte = 2;
3807 st->utf8.in_byte = 1;
3812 if ((c & 0xc0) != 0x80)
3818 return CODING_CATEGORY_UTF8_MASK;
3822 decode_coding_utf8 (Lstream *decoding, const unsigned char *src,
3823 unsigned_char_dynarr *dst, unsigned int n)
3825 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
3826 unsigned int flags = str->flags;
3827 unsigned int ch = str->ch;
3828 eol_type_t eol_type = str->eol_type;
3829 unsigned char counter = str->counter;
3833 unsigned char c = *src++;
3842 else if ( c >= 0xf8 )
3847 else if ( c >= 0xf0 )
3852 else if ( c >= 0xe0 )
3857 else if ( c >= 0xc0 )
3864 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
3865 decode_ucs4 (c, dst);
3869 ch = ( ch << 6 ) | ( c & 0x3f );
3870 decode_ucs4 (ch, dst);
3875 ch = ( ch << 6 ) | ( c & 0x3f );
3878 label_continue_loop:;
3881 if (flags & CODING_STATE_END)
3882 DECODE_OUTPUT_PARTIAL_CHAR (ch);
3886 str->counter = counter;
3890 encode_utf8 (Lisp_Object charset,
3891 unsigned char h, unsigned char l, unsigned_char_dynarr *dst)
3893 unsigned long code = mule_char_to_ucs4 (charset, h, l);
3896 Dynarr_add (dst, code);
3898 else if ( code <= 0x7ff )
3900 Dynarr_add (dst, (code >> 6) | 0xc0);
3901 Dynarr_add (dst, (code & 0x3f) | 0x80);
3903 else if ( code <= 0xffff )
3905 Dynarr_add (dst, (code >> 12) | 0xe0);
3906 Dynarr_add (dst, ((code >> 6) & 0x3f) | 0x80);
3907 Dynarr_add (dst, (code & 0x3f) | 0x80);
3909 else if ( code <= 0x1fffff )
3911 Dynarr_add (dst, (code >> 18) | 0xf0);
3912 Dynarr_add (dst, ((code >> 12) & 0x3f) | 0x80);
3913 Dynarr_add (dst, ((code >> 6) & 0x3f) | 0x80);
3914 Dynarr_add (dst, (code & 0x3f) | 0x80);
3916 else if ( code <= 0x3ffffff )
3918 Dynarr_add (dst, (code >> 24) | 0xf8);
3919 Dynarr_add (dst, ((code >> 18) & 0x3f) | 0x80);
3920 Dynarr_add (dst, ((code >> 12) & 0x3f) | 0x80);
3921 Dynarr_add (dst, ((code >> 6) & 0x3f) | 0x80);
3922 Dynarr_add (dst, (code & 0x3f) | 0x80);
3926 Dynarr_add (dst, (code >> 30) | 0xfc);
3927 Dynarr_add (dst, ((code >> 24) & 0x3f) | 0x80);
3928 Dynarr_add (dst, ((code >> 18) & 0x3f) | 0x80);
3929 Dynarr_add (dst, ((code >> 12) & 0x3f) | 0x80);
3930 Dynarr_add (dst, ((code >> 6) & 0x3f) | 0x80);
3931 Dynarr_add (dst, (code & 0x3f) | 0x80);
3936 encode_coding_utf8 (Lstream *encoding, const unsigned char *src,
3937 unsigned_char_dynarr *dst, unsigned int n)
3939 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
3940 unsigned int flags = str->flags;
3941 unsigned int ch = str->ch;
3942 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
3943 unsigned char char_boundary = str->iso2022.current_char_boundary;
3944 Lisp_Object charset = str->iso2022.current_charset;
3946 #ifdef ENABLE_COMPOSITE_CHARS
3947 /* flags for handling composite chars. We do a little switcharoo
3948 on the source while we're outputting the composite char. */
3949 unsigned int saved_n = 0;
3950 const unsigned char *saved_src = NULL;
3951 int in_composite = 0;
3954 #endif /* ENABLE_COMPOSITE_CHARS */
3958 unsigned char c = *src++;
3960 if (BYTE_ASCII_P (c))
3961 { /* Processing ASCII character */
3965 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
3966 Dynarr_add (dst, '\r');
3967 if (eol_type != EOL_CR)
3968 Dynarr_add (dst, c);
3971 encode_utf8 (Vcharset_ascii, c, 0, dst);
3974 else if (BUFBYTE_LEADING_BYTE_P (c) || BUFBYTE_LEADING_BYTE_P (ch))
3975 { /* Processing Leading Byte */
3977 charset = CHARSET_BY_LEADING_BYTE (c);
3978 if (LEADING_BYTE_PREFIX_P(c))
3983 { /* Processing Non-ASCII character */
3985 if (EQ (charset, Vcharset_control_1))
3987 encode_utf8 (Vcharset_control_1, c, 0, dst);
3991 switch (XCHARSET_REP_BYTES (charset))
3994 encode_utf8 (charset, c, 0, dst);
3997 if (XCHARSET_PRIVATE_P (charset))
3999 encode_utf8 (charset, c, 0, dst);
4004 #ifdef ENABLE_COMPOSITE_CHARS
4005 if (EQ (charset, Vcharset_composite))
4009 /* #### Bother! We don't know how to
4011 encode_utf8 (Vcharset_ascii, '~', 0, dst);
4015 Emchar emch = MAKE_CHAR (Vcharset_composite,
4016 ch & 0x7F, c & 0x7F);
4017 Lisp_Object lstr = composite_char_string (emch);
4021 src = XSTRING_DATA (lstr);
4022 n = XSTRING_LENGTH (lstr);
4026 #endif /* ENABLE_COMPOSITE_CHARS */
4028 encode_utf8 (charset, ch, c, dst);
4041 encode_utf8 (charset, ch, c, dst);
4057 #ifdef ENABLE_COMPOSITE_CHARS
4063 goto back_to_square_n; /* Wheeeeeeeee ..... */
4069 str->iso2022.current_char_boundary = char_boundary;
4070 str->iso2022.current_charset = charset;
4072 /* Verbum caro factum est! */
4076 /************************************************************************/
4077 /* ISO2022 methods */
4078 /************************************************************************/
4080 /* The following note describes the coding system ISO2022 briefly.
4081 Since the intention of this note is to help understand the
4082 functions in this file, some parts are NOT ACCURATE or OVERLY
4083 SIMPLIFIED. For thorough understanding, please refer to the
4084 original document of ISO2022.
4086 ISO2022 provides many mechanisms to encode several character sets
4087 in 7-bit and 8-bit environments. For 7-bit environments, all text
4088 is encoded using bytes less than 128. This may make the encoded
4089 text a little bit longer, but the text passes more easily through
4090 several gateways, some of which strip off MSB (Most Signigant Bit).
4092 There are two kinds of character sets: control character set and
4093 graphic character set. The former contains control characters such
4094 as `newline' and `escape' to provide control functions (control
4095 functions are also provided by escape sequences). The latter
4096 contains graphic characters such as 'A' and '-'. Emacs recognizes
4097 two control character sets and many graphic character sets.
4099 Graphic character sets are classified into one of the following
4100 four classes, according to the number of bytes (DIMENSION) and
4101 number of characters in one dimension (CHARS) of the set:
4102 - DIMENSION1_CHARS94
4103 - DIMENSION1_CHARS96
4104 - DIMENSION2_CHARS94
4105 - DIMENSION2_CHARS96
4107 In addition, each character set is assigned an identification tag,
4108 unique for each set, called "final character" (denoted as <F>
4109 hereafter). The <F> of each character set is decided by ECMA(*)
4110 when it is registered in ISO. The code range of <F> is 0x30..0x7F
4111 (0x30..0x3F are for private use only).
4113 Note (*): ECMA = European Computer Manufacturers Association
4115 Here are examples of graphic character set [NAME(<F>)]:
4116 o DIMENSION1_CHARS94 -- ASCII('B'), right-half-of-JISX0201('I'), ...
4117 o DIMENSION1_CHARS96 -- right-half-of-ISO8859-1('A'), ...
4118 o DIMENSION2_CHARS94 -- GB2312('A'), JISX0208('B'), ...
4119 o DIMENSION2_CHARS96 -- none for the moment
4121 A code area (1 byte = 8 bits) is divided into 4 areas, C0, GL, C1, and GR.
4122 C0 [0x00..0x1F] -- control character plane 0
4123 GL [0x20..0x7F] -- graphic character plane 0
4124 C1 [0x80..0x9F] -- control character plane 1
4125 GR [0xA0..0xFF] -- graphic character plane 1
4127 A control character set is directly designated and invoked to C0 or
4128 C1 by an escape sequence. The most common case is that:
4129 - ISO646's control character set is designated/invoked to C0, and
4130 - ISO6429's control character set is designated/invoked to C1,
4131 and usually these designations/invocations are omitted in encoded
4132 text. In a 7-bit environment, only C0 can be used, and a control
4133 character for C1 is encoded by an appropriate escape sequence to
4134 fit into the environment. All control characters for C1 are
4135 defined to have corresponding escape sequences.
4137 A graphic character set is at first designated to one of four
4138 graphic registers (G0 through G3), then these graphic registers are
4139 invoked to GL or GR. These designations and invocations can be
4140 done independently. The most common case is that G0 is invoked to
4141 GL, G1 is invoked to GR, and ASCII is designated to G0. Usually
4142 these invocations and designations are omitted in encoded text.
4143 In a 7-bit environment, only GL can be used.
4145 When a graphic character set of CHARS94 is invoked to GL, codes
4146 0x20 and 0x7F of the GL area work as control characters SPACE and
4147 DEL respectively, and codes 0xA0 and 0xFF of the GR area should not
4150 There are two ways of invocation: locking-shift and single-shift.
4151 With locking-shift, the invocation lasts until the next different
4152 invocation, whereas with single-shift, the invocation affects the
4153 following character only and doesn't affect the locking-shift
4154 state. Invocations are done by the following control characters or
4157 ----------------------------------------------------------------------
4158 abbrev function cntrl escape seq description
4159 ----------------------------------------------------------------------
4160 SI/LS0 (shift-in) 0x0F none invoke G0 into GL
4161 SO/LS1 (shift-out) 0x0E none invoke G1 into GL
4162 LS2 (locking-shift-2) none ESC 'n' invoke G2 into GL
4163 LS3 (locking-shift-3) none ESC 'o' invoke G3 into GL
4164 LS1R (locking-shift-1 right) none ESC '~' invoke G1 into GR (*)
4165 LS2R (locking-shift-2 right) none ESC '}' invoke G2 into GR (*)
4166 LS3R (locking-shift 3 right) none ESC '|' invoke G3 into GR (*)
4167 SS2 (single-shift-2) 0x8E ESC 'N' invoke G2 for one char
4168 SS3 (single-shift-3) 0x8F ESC 'O' invoke G3 for one char
4169 ----------------------------------------------------------------------
4170 (*) These are not used by any known coding system.
4172 Control characters for these functions are defined by macros
4173 ISO_CODE_XXX in `coding.h'.
4175 Designations are done by the following escape sequences:
4176 ----------------------------------------------------------------------
4177 escape sequence description
4178 ----------------------------------------------------------------------
4179 ESC '(' <F> designate DIMENSION1_CHARS94<F> to G0
4180 ESC ')' <F> designate DIMENSION1_CHARS94<F> to G1
4181 ESC '*' <F> designate DIMENSION1_CHARS94<F> to G2
4182 ESC '+' <F> designate DIMENSION1_CHARS94<F> to G3
4183 ESC ',' <F> designate DIMENSION1_CHARS96<F> to G0 (*)
4184 ESC '-' <F> designate DIMENSION1_CHARS96<F> to G1
4185 ESC '.' <F> designate DIMENSION1_CHARS96<F> to G2
4186 ESC '/' <F> designate DIMENSION1_CHARS96<F> to G3
4187 ESC '$' '(' <F> designate DIMENSION2_CHARS94<F> to G0 (**)
4188 ESC '$' ')' <F> designate DIMENSION2_CHARS94<F> to G1
4189 ESC '$' '*' <F> designate DIMENSION2_CHARS94<F> to G2
4190 ESC '$' '+' <F> designate DIMENSION2_CHARS94<F> to G3
4191 ESC '$' ',' <F> designate DIMENSION2_CHARS96<F> to G0 (*)
4192 ESC '$' '-' <F> designate DIMENSION2_CHARS96<F> to G1
4193 ESC '$' '.' <F> designate DIMENSION2_CHARS96<F> to G2
4194 ESC '$' '/' <F> designate DIMENSION2_CHARS96<F> to G3
4195 ----------------------------------------------------------------------
4197 In this list, "DIMENSION1_CHARS94<F>" means a graphic character set
4198 of dimension 1, chars 94, and final character <F>, etc...
4200 Note (*): Although these designations are not allowed in ISO2022,
4201 Emacs accepts them on decoding, and produces them on encoding
4202 CHARS96 character sets in a coding system which is characterized as
4203 7-bit environment, non-locking-shift, and non-single-shift.
4205 Note (**): If <F> is '@', 'A', or 'B', the intermediate character
4206 '(' can be omitted. We refer to this as "short-form" hereafter.
4208 Now you may notice that there are a lot of ways for encoding the
4209 same multilingual text in ISO2022. Actually, there exist many
4210 coding systems such as Compound Text (used in X11's inter client
4211 communication, ISO-2022-JP (used in Japanese internet), ISO-2022-KR
4212 (used in Korean internet), EUC (Extended UNIX Code, used in Asian
4213 localized platforms), and all of these are variants of ISO2022.
4215 In addition to the above, Emacs handles two more kinds of escape
4216 sequences: ISO6429's direction specification and Emacs' private
4217 sequence for specifying character composition.
4219 ISO6429's direction specification takes the following form:
4220 o CSI ']' -- end of the current direction
4221 o CSI '0' ']' -- end of the current direction
4222 o CSI '1' ']' -- start of left-to-right text
4223 o CSI '2' ']' -- start of right-to-left text
4224 The control character CSI (0x9B: control sequence introducer) is
4225 abbreviated to the escape sequence ESC '[' in a 7-bit environment.
4227 Character composition specification takes the following form:
4228 o ESC '0' -- start character composition
4229 o ESC '1' -- end character composition
4230 Since these are not standard escape sequences of any ISO standard,
4231 their use with these meanings is restricted to Emacs only. */
4234 reset_iso2022 (Lisp_Object coding_system, struct iso2022_decoder *iso)
4238 for (i = 0; i < 4; i++)
4240 if (!NILP (coding_system))
4242 XCODING_SYSTEM_ISO2022_INITIAL_CHARSET (coding_system, i);
4244 iso->charset[i] = Qt;
4245 iso->invalid_designated[i] = 0;
4247 iso->esc = ISO_ESC_NOTHING;
4248 iso->esc_bytes_index = 0;
4249 iso->register_left = 0;
4250 iso->register_right = 1;
4251 iso->switched_dir_and_no_valid_charset_yet = 0;
4252 iso->invalid_switch_dir = 0;
4253 iso->output_direction_sequence = 0;
4254 iso->output_literally = 0;
4255 #ifdef ENABLE_COMPOSITE_CHARS
4256 if (iso->composite_chars)
4257 Dynarr_reset (iso->composite_chars);
4262 fit_to_be_escape_quoted (unsigned char c)
4279 /* Parse one byte of an ISO2022 escape sequence.
4280 If the result is an invalid escape sequence, return 0 and
4281 do not change anything in STR. Otherwise, if the result is
4282 an incomplete escape sequence, update ISO2022.ESC and
4283 ISO2022.ESC_BYTES and return -1. Otherwise, update
4284 all the state variables (but not ISO2022.ESC_BYTES) and
4287 If CHECK_INVALID_CHARSETS is non-zero, check for designation
4288 or invocation of an invalid character set and treat that as
4289 an unrecognized escape sequence. */
4292 parse_iso2022_esc (Lisp_Object codesys, struct iso2022_decoder *iso,
4293 unsigned char c, unsigned int *flags,
4294 int check_invalid_charsets)
4296 /* (1) If we're at the end of a designation sequence, CS is the
4297 charset being designated and REG is the register to designate
4300 (2) If we're at the end of a locking-shift sequence, REG is
4301 the register to invoke and HALF (0 == left, 1 == right) is
4302 the half to invoke it into.
4304 (3) If we're at the end of a single-shift sequence, REG is
4305 the register to invoke. */
4306 Lisp_Object cs = Qnil;
4309 /* NOTE: This code does goto's all over the fucking place.
4310 The reason for this is that we're basically implementing
4311 a state machine here, and hierarchical languages like C
4312 don't really provide a clean way of doing this. */
4314 if (! (*flags & CODING_STATE_ESCAPE))
4315 /* At beginning of escape sequence; we need to reset our
4316 escape-state variables. */
4317 iso->esc = ISO_ESC_NOTHING;
4319 iso->output_literally = 0;
4320 iso->output_direction_sequence = 0;
4324 case ISO_ESC_NOTHING:
4325 iso->esc_bytes_index = 0;
4328 case ISO_CODE_ESC: /* Start escape sequence */
4329 *flags |= CODING_STATE_ESCAPE;
4333 case ISO_CODE_CSI: /* ISO6429 (specifying directionality) */
4334 *flags |= CODING_STATE_ESCAPE;
4335 iso->esc = ISO_ESC_5_11;
4338 case ISO_CODE_SO: /* locking shift 1 */
4341 case ISO_CODE_SI: /* locking shift 0 */
4345 case ISO_CODE_SS2: /* single shift */
4348 case ISO_CODE_SS3: /* single shift */
4352 default: /* Other control characters */
4359 /**** single shift ****/
4361 case 'N': /* single shift 2 */
4364 case 'O': /* single shift 3 */
4368 /**** locking shift ****/
4370 case '~': /* locking shift 1 right */
4373 case 'n': /* locking shift 2 */
4376 case '}': /* locking shift 2 right */
4379 case 'o': /* locking shift 3 */
4382 case '|': /* locking shift 3 right */
4386 #ifdef ENABLE_COMPOSITE_CHARS
4387 /**** composite ****/
4390 iso->esc = ISO_ESC_START_COMPOSITE;
4391 *flags = (*flags & CODING_STATE_ISO2022_LOCK) |
4392 CODING_STATE_COMPOSITE;
4396 iso->esc = ISO_ESC_END_COMPOSITE;
4397 *flags = (*flags & CODING_STATE_ISO2022_LOCK) &
4398 ~CODING_STATE_COMPOSITE;
4400 #endif /* ENABLE_COMPOSITE_CHARS */
4402 /**** directionality ****/
4405 iso->esc = ISO_ESC_5_11;
4408 /**** designation ****/
4410 case '$': /* multibyte charset prefix */
4411 iso->esc = ISO_ESC_2_4;
4415 if (0x28 <= c && c <= 0x2F)
4417 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_8);
4421 /* This function is called with CODESYS equal to nil when
4422 doing coding-system detection. */
4424 && XCODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
4425 && fit_to_be_escape_quoted (c))
4427 iso->esc = ISO_ESC_LITERAL;
4428 *flags &= CODING_STATE_ISO2022_LOCK;
4438 /**** directionality ****/
4440 case ISO_ESC_5_11: /* ISO6429 direction control */
4443 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4444 goto directionality;
4446 if (c == '0') iso->esc = ISO_ESC_5_11_0;
4447 else if (c == '1') iso->esc = ISO_ESC_5_11_1;
4448 else if (c == '2') iso->esc = ISO_ESC_5_11_2;
4452 case ISO_ESC_5_11_0:
4455 *flags &= (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4456 goto directionality;
4460 case ISO_ESC_5_11_1:
4463 *flags = (CODING_STATE_ISO2022_LOCK & ~CODING_STATE_R2L);
4464 goto directionality;
4468 case ISO_ESC_5_11_2:
4471 *flags = (*flags & CODING_STATE_ISO2022_LOCK) | CODING_STATE_R2L;
4472 goto directionality;
4477 iso->esc = ISO_ESC_DIRECTIONALITY;
4478 /* Various junk here to attempt to preserve the direction sequences
4479 literally in the text if they would otherwise be swallowed due
4480 to invalid designations that don't show up as actual charset
4481 changes in the text. */
4482 if (iso->invalid_switch_dir)
4484 /* We already inserted a direction switch literally into the
4485 text. We assume (#### this may not be right) that the
4486 next direction switch is the one going the other way,
4487 and we need to output that literally as well. */
4488 iso->output_literally = 1;
4489 iso->invalid_switch_dir = 0;
4495 /* If we are in the thrall of an invalid designation,
4496 then stick the directionality sequence literally into the
4497 output stream so it ends up in the original text again. */
4498 for (jj = 0; jj < 4; jj++)
4499 if (iso->invalid_designated[jj])
4503 iso->output_literally = 1;
4504 iso->invalid_switch_dir = 1;
4507 /* Indicate that we haven't yet seen a valid designation,
4508 so that if a switch-dir is directly followed by an
4509 invalid designation, both get inserted literally. */
4510 iso->switched_dir_and_no_valid_charset_yet = 1;
4515 /**** designation ****/
4518 if (0x28 <= c && c <= 0x2F)
4520 iso->esc = (enum iso_esc_flag) (c - 0x28 + ISO_ESC_2_4_8);
4523 if (0x40 <= c && c <= 0x42)
4525 cs = CHARSET_BY_ATTRIBUTES (CHARSET_TYPE_94X94, c,
4526 *flags & CODING_STATE_R2L ?
4527 CHARSET_RIGHT_TO_LEFT :
4528 CHARSET_LEFT_TO_RIGHT);
4538 if (c < '0' || c > '~')
4539 return 0; /* bad final byte */
4541 if (iso->esc >= ISO_ESC_2_8 &&
4542 iso->esc <= ISO_ESC_2_15)
4544 type = ((iso->esc >= ISO_ESC_2_12) ?
4545 CHARSET_TYPE_96 : CHARSET_TYPE_94);
4546 reg = (iso->esc - ISO_ESC_2_8) & 3;
4548 else if (iso->esc >= ISO_ESC_2_4_8 &&
4549 iso->esc <= ISO_ESC_2_4_15)
4551 type = ((iso->esc >= ISO_ESC_2_4_12) ?
4552 CHARSET_TYPE_96X96 : CHARSET_TYPE_94X94);
4553 reg = (iso->esc - ISO_ESC_2_4_8) & 3;
4557 /* Can this ever be reached? -slb */
4561 cs = CHARSET_BY_ATTRIBUTES (type, c,
4562 *flags & CODING_STATE_R2L ?
4563 CHARSET_RIGHT_TO_LEFT :
4564 CHARSET_LEFT_TO_RIGHT);
4570 iso->esc_bytes[iso->esc_bytes_index++] = (unsigned char) c;
4574 if (check_invalid_charsets && !CHARSETP (iso->charset[reg]))
4575 /* can't invoke something that ain't there. */
4577 iso->esc = ISO_ESC_SINGLE_SHIFT;
4578 *flags &= CODING_STATE_ISO2022_LOCK;
4580 *flags |= CODING_STATE_SS2;
4582 *flags |= CODING_STATE_SS3;
4586 if (check_invalid_charsets &&
4587 !CHARSETP (iso->charset[reg]))
4588 /* can't invoke something that ain't there. */
4591 iso->register_right = reg;
4593 iso->register_left = reg;
4594 *flags &= CODING_STATE_ISO2022_LOCK;
4595 iso->esc = ISO_ESC_LOCKING_SHIFT;
4599 if (NILP (cs) && check_invalid_charsets)
4601 iso->invalid_designated[reg] = 1;
4602 iso->charset[reg] = Vcharset_ascii;
4603 iso->esc = ISO_ESC_DESIGNATE;
4604 *flags &= CODING_STATE_ISO2022_LOCK;
4605 iso->output_literally = 1;
4606 if (iso->switched_dir_and_no_valid_charset_yet)
4608 /* We encountered a switch-direction followed by an
4609 invalid designation. Ensure that the switch-direction
4610 gets outputted; otherwise it will probably get eaten
4611 when the text is written out again. */
4612 iso->switched_dir_and_no_valid_charset_yet = 0;
4613 iso->output_direction_sequence = 1;
4614 /* And make sure that the switch-dir going the other
4615 way gets outputted, as well. */
4616 iso->invalid_switch_dir = 1;
4620 /* This function is called with CODESYS equal to nil when
4621 doing coding-system detection. */
4622 if (!NILP (codesys))
4624 charset_conversion_spec_dynarr *dyn =
4625 XCODING_SYSTEM (codesys)->iso2022.input_conv;
4631 for (i = 0; i < Dynarr_length (dyn); i++)
4633 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
4634 if (EQ (cs, spec->from_charset))
4635 cs = spec->to_charset;
4640 iso->charset[reg] = cs;
4641 iso->esc = ISO_ESC_DESIGNATE;
4642 *flags &= CODING_STATE_ISO2022_LOCK;
4643 if (iso->invalid_designated[reg])
4645 iso->invalid_designated[reg] = 0;
4646 iso->output_literally = 1;
4648 if (iso->switched_dir_and_no_valid_charset_yet)
4649 iso->switched_dir_and_no_valid_charset_yet = 0;
4654 detect_coding_iso2022 (struct detection_state *st, const unsigned char *src,
4659 /* #### There are serious deficiencies in the recognition mechanism
4660 here. This needs to be much smarter if it's going to cut it.
4661 The sequence "\xff\x0f" is currently detected as LOCK_SHIFT while
4662 it should be detected as Latin-1.
4663 All the ISO2022 stuff in this file should be synced up with the
4664 code from FSF Emacs-20.4, in which Mule should be more or less stable.
4665 Perhaps we should wait till R2L works in FSF Emacs? */
4667 if (!st->iso2022.initted)
4669 reset_iso2022 (Qnil, &st->iso2022.iso);
4670 st->iso2022.mask = (CODING_CATEGORY_ISO_7_MASK |
4671 CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
4672 CODING_CATEGORY_ISO_8_1_MASK |
4673 CODING_CATEGORY_ISO_8_2_MASK |
4674 CODING_CATEGORY_ISO_LOCK_SHIFT_MASK);
4675 st->iso2022.flags = 0;
4676 st->iso2022.high_byte_count = 0;
4677 st->iso2022.saw_single_shift = 0;
4678 st->iso2022.initted = 1;
4681 mask = st->iso2022.mask;
4688 mask &= ~CODING_CATEGORY_ISO_7_MASK;
4689 st->iso2022.high_byte_count++;
4693 if (st->iso2022.high_byte_count && !st->iso2022.saw_single_shift)
4695 if (st->iso2022.high_byte_count & 1)
4696 /* odd number of high bytes; assume not iso-8-2 */
4697 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
4699 st->iso2022.high_byte_count = 0;
4700 st->iso2022.saw_single_shift = 0;
4702 mask &= ~CODING_CATEGORY_ISO_7_MASK;
4704 if (!(st->iso2022.flags & CODING_STATE_ESCAPE)
4705 && (BYTE_C0_P (c) || BYTE_C1_P (c)))
4706 { /* control chars */
4709 /* Allow and ignore control characters that you might
4710 reasonably see in a text file */
4715 case 8: /* backspace */
4716 case 11: /* vertical tab */
4717 case 12: /* form feed */
4718 case 26: /* MS-DOS C-z junk */
4719 case 31: /* '^_' -- for info */
4720 goto label_continue_loop;
4727 if ((st->iso2022.flags & CODING_STATE_ESCAPE) || BYTE_C0_P (c)
4730 if (parse_iso2022_esc (Qnil, &st->iso2022.iso, c,
4731 &st->iso2022.flags, 0))
4733 switch (st->iso2022.iso.esc)
4735 case ISO_ESC_DESIGNATE:
4736 mask &= ~CODING_CATEGORY_ISO_8_1_MASK;
4737 mask &= ~CODING_CATEGORY_ISO_8_2_MASK;
4739 case ISO_ESC_LOCKING_SHIFT:
4740 mask = CODING_CATEGORY_ISO_LOCK_SHIFT_MASK;
4741 goto ran_out_of_chars;
4742 case ISO_ESC_SINGLE_SHIFT:
4743 mask &= ~CODING_CATEGORY_ISO_8_DESIGNATE_MASK;
4744 st->iso2022.saw_single_shift = 1;
4753 goto ran_out_of_chars;
4756 label_continue_loop:;
4765 postprocess_iso2022_mask (int mask)
4767 /* #### kind of cheesy */
4768 /* If seven-bit ISO is allowed, then assume that the encoding is
4769 entirely seven-bit and turn off the eight-bit ones. */
4770 if (mask & CODING_CATEGORY_ISO_7_MASK)
4771 mask &= ~ (CODING_CATEGORY_ISO_8_DESIGNATE_MASK |
4772 CODING_CATEGORY_ISO_8_1_MASK |
4773 CODING_CATEGORY_ISO_8_2_MASK);
4777 /* If FLAGS is a null pointer or specifies right-to-left motion,
4778 output a switch-dir-to-left-to-right sequence to DST.
4779 Also update FLAGS if it is not a null pointer.
4780 If INTERNAL_P is set, we are outputting in internal format and
4781 need to handle the CSI differently. */
4784 restore_left_to_right_direction (Lisp_Coding_System *codesys,
4785 unsigned_char_dynarr *dst,
4786 unsigned int *flags,
4789 if (!flags || (*flags & CODING_STATE_R2L))
4791 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
4793 Dynarr_add (dst, ISO_CODE_ESC);
4794 Dynarr_add (dst, '[');
4796 else if (internal_p)
4797 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
4799 Dynarr_add (dst, ISO_CODE_CSI);
4800 Dynarr_add (dst, '0');
4801 Dynarr_add (dst, ']');
4803 *flags &= ~CODING_STATE_R2L;
4807 /* If FLAGS is a null pointer or specifies a direction different from
4808 DIRECTION (which should be either CHARSET_RIGHT_TO_LEFT or
4809 CHARSET_LEFT_TO_RIGHT), output the appropriate switch-dir escape
4810 sequence to DST. Also update FLAGS if it is not a null pointer.
4811 If INTERNAL_P is set, we are outputting in internal format and
4812 need to handle the CSI differently. */
4815 ensure_correct_direction (int direction, Lisp_Coding_System *codesys,
4816 unsigned_char_dynarr *dst, unsigned int *flags,
4819 if ((!flags || (*flags & CODING_STATE_R2L)) &&
4820 direction == CHARSET_LEFT_TO_RIGHT)
4821 restore_left_to_right_direction (codesys, dst, flags, internal_p);
4822 else if (!CODING_SYSTEM_ISO2022_NO_ISO6429 (codesys)
4823 && (!flags || !(*flags & CODING_STATE_R2L)) &&
4824 direction == CHARSET_RIGHT_TO_LEFT)
4826 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
4828 Dynarr_add (dst, ISO_CODE_ESC);
4829 Dynarr_add (dst, '[');
4831 else if (internal_p)
4832 DECODE_ADD_BINARY_CHAR (ISO_CODE_CSI, dst);
4834 Dynarr_add (dst, ISO_CODE_CSI);
4835 Dynarr_add (dst, '2');
4836 Dynarr_add (dst, ']');
4838 *flags |= CODING_STATE_R2L;
4842 /* Convert ISO2022-format data to internal format. */
4845 decode_coding_iso2022 (Lstream *decoding, const unsigned char *src,
4846 unsigned_char_dynarr *dst, unsigned int n)
4848 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
4849 unsigned int flags = str->flags;
4850 unsigned int ch = str->ch;
4851 eol_type_t eol_type = str->eol_type;
4852 #ifdef ENABLE_COMPOSITE_CHARS
4853 unsigned_char_dynarr *real_dst = dst;
4855 Lisp_Object coding_system;
4857 XSETCODING_SYSTEM (coding_system, str->codesys);
4859 #ifdef ENABLE_COMPOSITE_CHARS
4860 if (flags & CODING_STATE_COMPOSITE)
4861 dst = str->iso2022.composite_chars;
4862 #endif /* ENABLE_COMPOSITE_CHARS */
4866 unsigned char c = *src++;
4867 if (flags & CODING_STATE_ESCAPE)
4868 { /* Within ESC sequence */
4869 int retval = parse_iso2022_esc (coding_system, &str->iso2022,
4874 switch (str->iso2022.esc)
4876 #ifdef ENABLE_COMPOSITE_CHARS
4877 case ISO_ESC_START_COMPOSITE:
4878 if (str->iso2022.composite_chars)
4879 Dynarr_reset (str->iso2022.composite_chars);
4881 str->iso2022.composite_chars = Dynarr_new (unsigned_char);
4882 dst = str->iso2022.composite_chars;
4884 case ISO_ESC_END_COMPOSITE:
4886 Bufbyte comstr[MAX_EMCHAR_LEN];
4888 Emchar emch = lookup_composite_char (Dynarr_atp (dst, 0),
4889 Dynarr_length (dst));
4891 len = set_charptr_emchar (comstr, emch);
4892 Dynarr_add_many (dst, comstr, len);
4895 #endif /* ENABLE_COMPOSITE_CHARS */
4897 case ISO_ESC_LITERAL:
4898 DECODE_ADD_BINARY_CHAR (c, dst);
4902 /* Everything else handled already */
4907 /* Attempted error recovery. */
4908 if (str->iso2022.output_direction_sequence)
4909 ensure_correct_direction (flags & CODING_STATE_R2L ?
4910 CHARSET_RIGHT_TO_LEFT :
4911 CHARSET_LEFT_TO_RIGHT,
4912 str->codesys, dst, 0, 1);
4913 /* More error recovery. */
4914 if (!retval || str->iso2022.output_literally)
4916 /* Output the (possibly invalid) sequence */
4918 for (i = 0; i < str->iso2022.esc_bytes_index; i++)
4919 DECODE_ADD_BINARY_CHAR (str->iso2022.esc_bytes[i], dst);
4920 flags &= CODING_STATE_ISO2022_LOCK;
4922 n++, src--;/* Repeat the loop with the same character. */
4925 /* No sense in reprocessing the final byte of the
4926 escape sequence; it could mess things up anyway.
4928 DECODE_ADD_BINARY_CHAR (c, dst);
4933 else if (BYTE_C0_P (c) || BYTE_C1_P (c))
4934 { /* Control characters */
4936 /***** Error-handling *****/
4938 /* If we were in the middle of a character, dump out the
4939 partial character. */
4940 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4942 /* If we just saw a single-shift character, dump it out.
4943 This may dump out the wrong sort of single-shift character,
4944 but least it will give an indication that something went
4946 if (flags & CODING_STATE_SS2)
4948 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS2, dst);
4949 flags &= ~CODING_STATE_SS2;
4951 if (flags & CODING_STATE_SS3)
4953 DECODE_ADD_BINARY_CHAR (ISO_CODE_SS3, dst);
4954 flags &= ~CODING_STATE_SS3;
4957 /***** Now handle the control characters. *****/
4960 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
4962 flags &= CODING_STATE_ISO2022_LOCK;
4964 if (!parse_iso2022_esc (coding_system, &str->iso2022, c, &flags, 1))
4965 DECODE_ADD_BINARY_CHAR (c, dst);
4968 { /* Graphic characters */
4969 Lisp_Object charset;
4973 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
4975 /* Now determine the charset. */
4976 reg = ((flags & CODING_STATE_SS2) ? 2
4977 : (flags & CODING_STATE_SS3) ? 3
4978 : !BYTE_ASCII_P (c) ? str->iso2022.register_right
4979 : str->iso2022.register_left);
4980 charset = str->iso2022.charset[reg];
4982 /* Error checking: */
4983 if (! CHARSETP (charset)
4984 || str->iso2022.invalid_designated[reg]
4985 || (((c & 0x7F) == ' ' || (c & 0x7F) == ISO_CODE_DEL)
4986 && XCHARSET_CHARS (charset) == 94))
4987 /* Mrmph. We are trying to invoke a register that has no
4988 or an invalid charset in it, or trying to add a character
4989 outside the range of the charset. Insert that char literally
4990 to preserve it for the output. */
4992 DECODE_OUTPUT_PARTIAL_CHAR (ch);
4993 DECODE_ADD_BINARY_CHAR (c, dst);
4998 /* Things are probably hunky-dorey. */
5000 /* Fetch reverse charset, maybe. */
5001 if (((flags & CODING_STATE_R2L) &&
5002 XCHARSET_DIRECTION (charset) == CHARSET_LEFT_TO_RIGHT)
5004 (!(flags & CODING_STATE_R2L) &&
5005 XCHARSET_DIRECTION (charset) == CHARSET_RIGHT_TO_LEFT))
5007 Lisp_Object new_charset =
5008 XCHARSET_REVERSE_DIRECTION_CHARSET (charset);
5009 if (!NILP (new_charset))
5010 charset = new_charset;
5013 lb = XCHARSET_LEADING_BYTE (charset);
5014 switch (XCHARSET_REP_BYTES (charset))
5017 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5018 Dynarr_add (dst, c & 0x7F);
5021 case 2: /* one-byte official */
5022 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5023 Dynarr_add (dst, lb);
5024 Dynarr_add (dst, c | 0x80);
5027 case 3: /* one-byte private or two-byte official */
5028 if (XCHARSET_PRIVATE_P (charset))
5030 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5031 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_1);
5032 Dynarr_add (dst, lb);
5033 Dynarr_add (dst, c | 0x80);
5039 Dynarr_add (dst, lb);
5040 Dynarr_add (dst, ch | 0x80);
5041 Dynarr_add (dst, c | 0x80);
5049 default: /* two-byte private */
5052 Dynarr_add (dst, PRE_LEADING_BYTE_PRIVATE_2);
5053 Dynarr_add (dst, lb);
5054 Dynarr_add (dst, ch | 0x80);
5055 Dynarr_add (dst, c | 0x80);
5064 flags &= CODING_STATE_ISO2022_LOCK;
5067 label_continue_loop:;
5070 if (flags & CODING_STATE_END)
5071 DECODE_OUTPUT_PARTIAL_CHAR (ch);
5078 /***** ISO2022 encoder *****/
5080 /* Designate CHARSET into register REG. */
5083 iso2022_designate (Lisp_Object charset, unsigned char reg,
5084 struct encoding_stream *str, unsigned_char_dynarr *dst)
5086 static const char inter94[] = "()*+";
5087 static const char inter96[] = ",-./";
5089 unsigned char final;
5090 Lisp_Object old_charset = str->iso2022.charset[reg];
5092 str->iso2022.charset[reg] = charset;
5093 if (!CHARSETP (charset))
5094 /* charset might be an initial nil or t. */
5096 type = XCHARSET_TYPE (charset);
5097 final = XCHARSET_FINAL (charset);
5098 if (!str->iso2022.force_charset_on_output[reg] &&
5099 CHARSETP (old_charset) &&
5100 XCHARSET_TYPE (old_charset) == type &&
5101 XCHARSET_FINAL (old_charset) == final)
5104 str->iso2022.force_charset_on_output[reg] = 0;
5107 charset_conversion_spec_dynarr *dyn =
5108 str->codesys->iso2022.output_conv;
5114 for (i = 0; i < Dynarr_length (dyn); i++)
5116 struct charset_conversion_spec *spec = Dynarr_atp (dyn, i);
5117 if (EQ (charset, spec->from_charset))
5118 charset = spec->to_charset;
5123 Dynarr_add (dst, ISO_CODE_ESC);
5126 case CHARSET_TYPE_94:
5127 Dynarr_add (dst, inter94[reg]);
5129 case CHARSET_TYPE_96:
5130 Dynarr_add (dst, inter96[reg]);
5132 case CHARSET_TYPE_94X94:
5133 Dynarr_add (dst, '$');
5135 || !(CODING_SYSTEM_ISO2022_SHORT (str->codesys))
5138 Dynarr_add (dst, inter94[reg]);
5140 case CHARSET_TYPE_96X96:
5141 Dynarr_add (dst, '$');
5142 Dynarr_add (dst, inter96[reg]);
5145 Dynarr_add (dst, final);
5149 ensure_normal_shift (struct encoding_stream *str, unsigned_char_dynarr *dst)
5151 if (str->iso2022.register_left != 0)
5153 Dynarr_add (dst, ISO_CODE_SI);
5154 str->iso2022.register_left = 0;
5159 ensure_shift_out (struct encoding_stream *str, unsigned_char_dynarr *dst)
5161 if (str->iso2022.register_left != 1)
5163 Dynarr_add (dst, ISO_CODE_SO);
5164 str->iso2022.register_left = 1;
5168 /* Convert internally-formatted data to ISO2022 format. */
5171 encode_coding_iso2022 (Lstream *encoding, const unsigned char *src,
5172 unsigned_char_dynarr *dst, unsigned int n)
5174 unsigned char charmask, c;
5175 unsigned char char_boundary;
5176 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
5177 unsigned int flags = str->flags;
5178 unsigned int ch = str->ch;
5179 Lisp_Coding_System *codesys = str->codesys;
5180 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
5182 Lisp_Object charset;
5185 #ifdef ENABLE_COMPOSITE_CHARS
5186 /* flags for handling composite chars. We do a little switcharoo
5187 on the source while we're outputting the composite char. */
5188 unsigned int saved_n = 0;
5189 const unsigned char *saved_src = NULL;
5190 int in_composite = 0;
5191 #endif /* ENABLE_COMPOSITE_CHARS */
5193 char_boundary = str->iso2022.current_char_boundary;
5194 charset = str->iso2022.current_charset;
5195 half = str->iso2022.current_half;
5197 #ifdef ENABLE_COMPOSITE_CHARS
5204 if (BYTE_ASCII_P (c))
5205 { /* Processing ASCII character */
5208 restore_left_to_right_direction (codesys, dst, &flags, 0);
5210 /* Make sure G0 contains ASCII */
5211 if ((c > ' ' && c < ISO_CODE_DEL) ||
5212 !CODING_SYSTEM_ISO2022_NO_ASCII_CNTL (codesys))
5214 ensure_normal_shift (str, dst);
5215 iso2022_designate (Vcharset_ascii, 0, str, dst);
5218 /* If necessary, restore everything to the default state
5221 !(CODING_SYSTEM_ISO2022_NO_ASCII_EOL (codesys)))
5223 restore_left_to_right_direction (codesys, dst, &flags, 0);
5225 ensure_normal_shift (str, dst);
5227 for (i = 0; i < 4; i++)
5229 Lisp_Object initial_charset =
5230 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
5231 iso2022_designate (initial_charset, i, str, dst);
5236 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5237 Dynarr_add (dst, '\r');
5238 if (eol_type != EOL_CR)
5239 Dynarr_add (dst, c);
5243 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
5244 && fit_to_be_escape_quoted (c))
5245 Dynarr_add (dst, ISO_CODE_ESC);
5246 Dynarr_add (dst, c);
5251 else if (BUFBYTE_LEADING_BYTE_P (c) || BUFBYTE_LEADING_BYTE_P (ch))
5252 { /* Processing Leading Byte */
5254 charset = CHARSET_BY_LEADING_BYTE (c);
5255 if (LEADING_BYTE_PREFIX_P(c))
5257 else if (!EQ (charset, Vcharset_control_1)
5258 #ifdef ENABLE_COMPOSITE_CHARS
5259 && !EQ (charset, Vcharset_composite)
5265 ensure_correct_direction (XCHARSET_DIRECTION (charset),
5266 codesys, dst, &flags, 0);
5268 /* Now determine which register to use. */
5270 for (i = 0; i < 4; i++)
5272 if (EQ (charset, str->iso2022.charset[i]) ||
5274 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i)))
5283 if (XCHARSET_GRAPHIC (charset) != 0)
5285 if (!NILP (str->iso2022.charset[1]) &&
5286 (!CODING_SYSTEM_ISO2022_SEVEN (codesys) ||
5287 CODING_SYSTEM_ISO2022_LOCK_SHIFT (codesys)))
5289 else if (!NILP (str->iso2022.charset[2]))
5291 else if (!NILP (str->iso2022.charset[3]))
5300 iso2022_designate (charset, reg, str, dst);
5302 /* Now invoke that register. */
5306 ensure_normal_shift (str, dst);
5311 if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
5313 ensure_shift_out (str, dst);
5321 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
5323 Dynarr_add (dst, ISO_CODE_ESC);
5324 Dynarr_add (dst, 'N');
5329 Dynarr_add (dst, ISO_CODE_SS2);
5335 if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
5337 Dynarr_add (dst, ISO_CODE_ESC);
5338 Dynarr_add (dst, 'O');
5343 Dynarr_add (dst, ISO_CODE_SS3);
5355 { /* Processing Non-ASCII character */
5356 charmask = (half == 0 ? 0x7F : 0xFF);
5358 if (EQ (charset, Vcharset_control_1))
5360 if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
5361 && fit_to_be_escape_quoted (c))
5362 Dynarr_add (dst, ISO_CODE_ESC);
5363 /* you asked for it ... */
5364 Dynarr_add (dst, c - 0x20);
5368 switch (XCHARSET_REP_BYTES (charset))
5371 Dynarr_add (dst, c & charmask);
5374 if (XCHARSET_PRIVATE_P (charset))
5376 Dynarr_add (dst, c & charmask);
5381 #ifdef ENABLE_COMPOSITE_CHARS
5382 if (EQ (charset, Vcharset_composite))
5386 /* #### Bother! We don't know how to
5388 Dynarr_add (dst, '~');
5392 Emchar emch = MAKE_CHAR (Vcharset_composite,
5393 ch & 0x7F, c & 0x7F);
5394 Lisp_Object lstr = composite_char_string (emch);
5398 src = XSTRING_DATA (lstr);
5399 n = XSTRING_LENGTH (lstr);
5400 Dynarr_add (dst, ISO_CODE_ESC);
5401 Dynarr_add (dst, '0'); /* start composing */
5405 #endif /* ENABLE_COMPOSITE_CHARS */
5407 Dynarr_add (dst, ch & charmask);
5408 Dynarr_add (dst, c & charmask);
5421 Dynarr_add (dst, ch & charmask);
5422 Dynarr_add (dst, c & charmask);
5438 #ifdef ENABLE_COMPOSITE_CHARS
5444 Dynarr_add (dst, ISO_CODE_ESC);
5445 Dynarr_add (dst, '1'); /* end composing */
5446 goto back_to_square_n; /* Wheeeeeeeee ..... */
5448 #endif /* ENABLE_COMPOSITE_CHARS */
5450 if (char_boundary && flags & CODING_STATE_END)
5452 restore_left_to_right_direction (codesys, dst, &flags, 0);
5453 ensure_normal_shift (str, dst);
5454 for (i = 0; i < 4; i++)
5456 Lisp_Object initial_charset =
5457 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
5458 iso2022_designate (initial_charset, i, str, dst);
5464 str->iso2022.current_char_boundary = char_boundary;
5465 str->iso2022.current_charset = charset;
5466 str->iso2022.current_half = half;
5468 /* Verbum caro factum est! */
5472 /************************************************************************/
5473 /* No-conversion methods */
5474 /************************************************************************/
5476 /* This is used when reading in "binary" files -- i.e. files that may
5477 contain all 256 possible byte values and that are not to be
5478 interpreted as being in any particular decoding. */
5480 decode_coding_no_conversion (Lstream *decoding, const unsigned char *src,
5481 unsigned_char_dynarr *dst, unsigned int n)
5484 struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
5485 unsigned int flags = str->flags;
5486 unsigned int ch = str->ch;
5487 eol_type_t eol_type = str->eol_type;
5493 DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
5494 DECODE_ADD_BINARY_CHAR (c, dst);
5495 label_continue_loop:;
5498 DECODE_HANDLE_END_OF_CONVERSION (flags, ch, dst);
5505 encode_coding_no_conversion (Lstream *encoding, const unsigned char *src,
5506 unsigned_char_dynarr *dst, unsigned int n)
5509 struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
5510 unsigned int flags = str->flags;
5511 unsigned int ch = str->ch;
5512 eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
5519 if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
5520 Dynarr_add (dst, '\r');
5521 if (eol_type != EOL_CR)
5522 Dynarr_add (dst, '\n');
5525 else if (BYTE_ASCII_P (c))
5528 Dynarr_add (dst, c);
5530 else if (BUFBYTE_LEADING_BYTE_P (c))
5533 if (c == LEADING_BYTE_LATIN_ISO8859_1 ||
5534 c == LEADING_BYTE_CONTROL_1)
5537 Dynarr_add (dst, '~'); /* untranslatable character */
5541 if (ch == LEADING_BYTE_LATIN_ISO8859_1)
5542 Dynarr_add (dst, c);
5543 else if (ch == LEADING_BYTE_CONTROL_1)
5546 Dynarr_add (dst, c - 0x20);
5548 /* else it should be the second or third byte of an
5549 untranslatable character, so ignore it */
5560 /************************************************************************/
5561 /* Initialization */
5562 /************************************************************************/
5565 syms_of_file_coding (void)
5567 INIT_LRECORD_IMPLEMENTATION (coding_system);
5569 deferror (&Qcoding_system_error, "coding-system-error",
5570 "Coding-system error", Qio_error);
5572 DEFSUBR (Fcoding_system_p);
5573 DEFSUBR (Ffind_coding_system);
5574 DEFSUBR (Fget_coding_system);
5575 DEFSUBR (Fcoding_system_list);
5576 DEFSUBR (Fcoding_system_name);
5577 DEFSUBR (Fmake_coding_system);
5578 DEFSUBR (Fcopy_coding_system);
5579 DEFSUBR (Fcoding_system_canonical_name_p);
5580 DEFSUBR (Fcoding_system_alias_p);
5581 DEFSUBR (Fcoding_system_aliasee);
5582 DEFSUBR (Fdefine_coding_system_alias);
5583 DEFSUBR (Fsubsidiary_coding_system);
5585 DEFSUBR (Fcoding_system_type);
5586 DEFSUBR (Fcoding_system_doc_string);
5588 DEFSUBR (Fcoding_system_charset);
5590 DEFSUBR (Fcoding_system_property);
5592 DEFSUBR (Fcoding_category_list);
5593 DEFSUBR (Fset_coding_priority_list);
5594 DEFSUBR (Fcoding_priority_list);
5595 DEFSUBR (Fset_coding_category_system);
5596 DEFSUBR (Fcoding_category_system);
5598 DEFSUBR (Fdetect_coding_region);
5599 DEFSUBR (Fdecode_coding_region);
5600 DEFSUBR (Fencode_coding_region);
5602 DEFSUBR (Fdecode_shift_jis_char);
5603 DEFSUBR (Fencode_shift_jis_char);
5604 DEFSUBR (Fdecode_big5_char);
5605 DEFSUBR (Fencode_big5_char);
5606 DEFSUBR (Fset_ucs_char);
5607 DEFSUBR (Fucs_char);
5608 DEFSUBR (Fset_char_ucs);
5609 DEFSUBR (Fchar_ucs);
5611 defsymbol (&Qcoding_systemp, "coding-system-p");
5612 defsymbol (&Qno_conversion, "no-conversion");
5613 defsymbol (&Qraw_text, "raw-text");
5615 defsymbol (&Qbig5, "big5");
5616 defsymbol (&Qshift_jis, "shift-jis");
5617 defsymbol (&Qucs4, "ucs-4");
5618 defsymbol (&Qutf8, "utf-8");
5619 defsymbol (&Qccl, "ccl");
5620 defsymbol (&Qiso2022, "iso2022");
5622 defsymbol (&Qmnemonic, "mnemonic");
5623 defsymbol (&Qeol_type, "eol-type");
5624 defsymbol (&Qpost_read_conversion, "post-read-conversion");
5625 defsymbol (&Qpre_write_conversion, "pre-write-conversion");
5627 defsymbol (&Qcr, "cr");
5628 defsymbol (&Qlf, "lf");
5629 defsymbol (&Qcrlf, "crlf");
5630 defsymbol (&Qeol_cr, "eol-cr");
5631 defsymbol (&Qeol_lf, "eol-lf");
5632 defsymbol (&Qeol_crlf, "eol-crlf");
5634 defsymbol (&Qcharset_g0, "charset-g0");
5635 defsymbol (&Qcharset_g1, "charset-g1");
5636 defsymbol (&Qcharset_g2, "charset-g2");
5637 defsymbol (&Qcharset_g3, "charset-g3");
5638 defsymbol (&Qforce_g0_on_output, "force-g0-on-output");
5639 defsymbol (&Qforce_g1_on_output, "force-g1-on-output");
5640 defsymbol (&Qforce_g2_on_output, "force-g2-on-output");
5641 defsymbol (&Qforce_g3_on_output, "force-g3-on-output");
5642 defsymbol (&Qno_iso6429, "no-iso6429");
5643 defsymbol (&Qinput_charset_conversion, "input-charset-conversion");
5644 defsymbol (&Qoutput_charset_conversion, "output-charset-conversion");
5646 defsymbol (&Qshort, "short");
5647 defsymbol (&Qno_ascii_eol, "no-ascii-eol");
5648 defsymbol (&Qno_ascii_cntl, "no-ascii-cntl");
5649 defsymbol (&Qseven, "seven");
5650 defsymbol (&Qlock_shift, "lock-shift");
5651 defsymbol (&Qescape_quoted, "escape-quoted");
5653 defsymbol (&Qencode, "encode");
5654 defsymbol (&Qdecode, "decode");
5657 defsymbol (&coding_category_symbol[CODING_CATEGORY_SHIFT_JIS],
5659 defsymbol (&coding_category_symbol[CODING_CATEGORY_BIG5],
5661 defsymbol (&coding_category_symbol[CODING_CATEGORY_UCS4],
5663 defsymbol (&coding_category_symbol[CODING_CATEGORY_UTF8],
5665 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_7],
5667 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_DESIGNATE],
5669 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_1],
5671 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_8_2],
5673 defsymbol (&coding_category_symbol[CODING_CATEGORY_ISO_LOCK_SHIFT],
5676 defsymbol (&coding_category_symbol[CODING_CATEGORY_NO_CONVERSION],
5681 lstream_type_create_file_coding (void)
5683 LSTREAM_HAS_METHOD (decoding, reader);
5684 LSTREAM_HAS_METHOD (decoding, writer);
5685 LSTREAM_HAS_METHOD (decoding, rewinder);
5686 LSTREAM_HAS_METHOD (decoding, seekable_p);
5687 LSTREAM_HAS_METHOD (decoding, flusher);
5688 LSTREAM_HAS_METHOD (decoding, closer);
5689 LSTREAM_HAS_METHOD (decoding, marker);
5691 LSTREAM_HAS_METHOD (encoding, reader);
5692 LSTREAM_HAS_METHOD (encoding, writer);
5693 LSTREAM_HAS_METHOD (encoding, rewinder);
5694 LSTREAM_HAS_METHOD (encoding, seekable_p);
5695 LSTREAM_HAS_METHOD (encoding, flusher);
5696 LSTREAM_HAS_METHOD (encoding, closer);
5697 LSTREAM_HAS_METHOD (encoding, marker);
5701 vars_of_file_coding (void)
5705 fcd = xnew (struct file_coding_dump);
5706 dumpstruct (&fcd, &fcd_description);
5708 /* Initialize to something reasonable ... */
5709 for (i = 0; i <= CODING_CATEGORY_LAST; i++)
5711 fcd->coding_category_system[i] = Qnil;
5712 fcd->coding_category_by_priority[i] = i;
5715 Fprovide (intern ("file-coding"));
5717 DEFVAR_LISP ("keyboard-coding-system", &Vkeyboard_coding_system /*
5718 Coding system used for TTY keyboard input.
5719 Not used under a windowing system.
5721 Vkeyboard_coding_system = Qnil;
5723 DEFVAR_LISP ("terminal-coding-system", &Vterminal_coding_system /*
5724 Coding system used for TTY display output.
5725 Not used under a windowing system.
5727 Vterminal_coding_system = Qnil;
5729 DEFVAR_LISP ("coding-system-for-read", &Vcoding_system_for_read /*
5730 Overriding coding system used when reading from a file or process.
5731 You should bind this variable with `let', but do not set it globally.
5732 If this is non-nil, it specifies the coding system that will be used
5733 to decode input on read operations, such as from a file or process.
5734 It overrides `buffer-file-coding-system-for-read',
5735 `insert-file-contents-pre-hook', etc. Use those variables instead of
5736 this one for permanent changes to the environment. */ );
5737 Vcoding_system_for_read = Qnil;
5739 DEFVAR_LISP ("coding-system-for-write",
5740 &Vcoding_system_for_write /*
5741 Overriding coding system used when writing to a file or process.
5742 You should bind this variable with `let', but do not set it globally.
5743 If this is non-nil, it specifies the coding system that will be used
5744 to encode output for write operations, such as to a file or process.
5745 It overrides `buffer-file-coding-system', `write-region-pre-hook', etc.
5746 Use those variables instead of this one for permanent changes to the
5748 Vcoding_system_for_write = Qnil;
5750 DEFVAR_LISP ("file-name-coding-system", &Vfile_name_coding_system /*
5751 Coding system used to convert pathnames when accessing files.
5753 Vfile_name_coding_system = Qnil;
5755 DEFVAR_BOOL ("enable-multibyte-characters", &enable_multibyte_characters /*
5756 Non-nil means the buffer contents are regarded as multi-byte form
5757 of characters, not a binary code. This affects the display, file I/O,
5758 and behaviors of various editing commands.
5760 Setting this to nil does not do anything.
5762 enable_multibyte_characters = 1;
5766 complex_vars_of_file_coding (void)
5768 staticpro (&Vcoding_system_hash_table);
5769 Vcoding_system_hash_table =
5770 make_lisp_hash_table (50, HASH_TABLE_NON_WEAK, HASH_TABLE_EQ);
5772 the_codesys_prop_dynarr = Dynarr_new (codesys_prop);
5773 dumpstruct (&the_codesys_prop_dynarr, &codesys_prop_dynarr_description);
5775 #define DEFINE_CODESYS_PROP(Prop_Type, Sym) do \
5777 struct codesys_prop csp; \
5779 csp.prop_type = (Prop_Type); \
5780 Dynarr_add (the_codesys_prop_dynarr, csp); \
5783 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qmnemonic);
5784 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_type);
5785 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_cr);
5786 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_crlf);
5787 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qeol_lf);
5788 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpost_read_conversion);
5789 DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK, Qpre_write_conversion);
5791 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g0);
5792 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g1);
5793 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g2);
5794 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qcharset_g3);
5795 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g0_on_output);
5796 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g1_on_output);
5797 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g2_on_output);
5798 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qforce_g3_on_output);
5799 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qshort);
5800 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_eol);
5801 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_ascii_cntl);
5802 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qseven);
5803 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qlock_shift);
5804 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qno_iso6429);
5805 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qescape_quoted);
5806 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qinput_charset_conversion);
5807 DEFINE_CODESYS_PROP (CODESYS_PROP_ISO2022, Qoutput_charset_conversion);
5809 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qencode);
5810 DEFINE_CODESYS_PROP (CODESYS_PROP_CCL, Qdecode);
5812 /* Need to create this here or we're really screwed. */
5814 (Qraw_text, Qno_conversion,
5815 build_string ("Raw text, which means it converts only line-break-codes."),
5816 list2 (Qmnemonic, build_string ("Raw")));
5819 (Qbinary, Qno_conversion,
5820 build_string ("Binary, which means it does not convert anything."),
5821 list4 (Qeol_type, Qlf,
5822 Qmnemonic, build_string ("Binary")));
5824 Fdefine_coding_system_alias (Qno_conversion, Qraw_text);
5826 Fdefine_coding_system_alias (Qfile_name, Qbinary);
5828 Fdefine_coding_system_alias (Qterminal, Qbinary);
5829 Fdefine_coding_system_alias (Qkeyboard, Qbinary);
5831 /* Need this for bootstrapping */
5832 fcd->coding_category_system[CODING_CATEGORY_NO_CONVERSION] =
5833 Fget_coding_system (Qraw_text);
5839 for (i = 0; i < 65536; i++)
5840 fcd->ucs_to_mule_table[i] = Qnil;
5842 staticpro (&mule_to_ucs_table);
5843 mule_to_ucs_table = Fmake_char_table(Qgeneric);