Commit | Line | Data |
---|---|---|
4ed46869 | 1 | /* Header for coding system handler. |
75c8c592 | 2 | Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN. |
8f924df7 KH |
3 | Licensed to the Free Software Foundation. |
4 | Copyright (C) 2003 | |
df7492f9 KH |
5 | National Institute of Advanced Industrial Science and Technology (AIST) |
6 | Registration Number H13PRO009 | |
4ed46869 | 7 | |
369314dc KH |
8 | This file is part of GNU Emacs. |
9 | ||
10 | GNU Emacs is free software; you can redistribute it and/or modify | |
11 | it under the terms of the GNU General Public License as published by | |
12 | the Free Software Foundation; either version 2, or (at your option) | |
13 | any later version. | |
4ed46869 | 14 | |
369314dc KH |
15 | GNU Emacs is distributed in the hope that it will be useful, |
16 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
17 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
18 | GNU General Public License for more details. | |
4ed46869 | 19 | |
369314dc KH |
20 | You should have received a copy of the GNU General Public License |
21 | along with GNU Emacs; see the file COPYING. If not, write to | |
22 | the Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
23 | Boston, MA 02111-1307, USA. */ | |
4ed46869 | 24 | |
6f776e81 KH |
25 | #ifndef EMACS_CODING_H |
26 | #define EMACS_CODING_H | |
4ed46869 | 27 | |
df7492f9 | 28 | /* Index to arguments of Fdefine_coding_system_internal. */ |
4ed46869 | 29 | |
df7492f9 KH |
30 | enum define_coding_system_arg_index |
31 | { | |
32 | coding_arg_name, | |
33 | coding_arg_mnemonic, | |
34 | coding_arg_coding_type, | |
35 | coding_arg_charset_list, | |
36 | coding_arg_ascii_compatible_p, | |
37 | coding_arg_decode_translation_table, | |
38 | coding_arg_encode_translation_table, | |
39 | coding_arg_post_read_conversion, | |
40 | coding_arg_pre_write_conversion, | |
41 | coding_arg_default_char, | |
8f924df7 | 42 | coding_arg_for_unibyte, |
df7492f9 KH |
43 | coding_arg_plist, |
44 | coding_arg_eol_type, | |
45 | coding_arg_max | |
46 | }; | |
4ed46869 | 47 | |
df7492f9 | 48 | enum define_coding_iso2022_arg_index |
4ed46869 | 49 | { |
df7492f9 KH |
50 | coding_arg_iso2022_initial = coding_arg_max, |
51 | coding_arg_iso2022_reg_usage, | |
52 | coding_arg_iso2022_request, | |
53 | coding_arg_iso2022_flags, | |
54 | coding_arg_iso2022_max | |
4ed46869 KH |
55 | }; |
56 | ||
df7492f9 | 57 | enum define_coding_utf16_arg_index |
4ed46869 | 58 | { |
df7492f9 KH |
59 | coding_arg_utf16_bom = coding_arg_max, |
60 | coding_arg_utf16_endian, | |
61 | coding_arg_utf16_max | |
4ed46869 KH |
62 | }; |
63 | ||
df7492f9 KH |
64 | enum define_coding_ccl_arg_index |
65 | { | |
35d47d18 | 66 | coding_arg_ccl_decoder = coding_arg_max, |
df7492f9 KH |
67 | coding_arg_ccl_encoder, |
68 | coding_arg_ccl_valids, | |
69 | coding_arg_ccl_max | |
70 | }; | |
4ed46869 | 71 | |
df7492f9 | 72 | extern Lisp_Object Vcoding_system_hash_table; |
4ed46869 | 73 | |
df7492f9 | 74 | /* Enumeration of coding system type. */ |
4ed46869 | 75 | |
df7492f9 KH |
76 | enum coding_system_type |
77 | { | |
78 | coding_type_charset, | |
79 | coding_type_utf_8, | |
80 | coding_type_utf_16, | |
81 | coding_type_iso_2022, | |
82 | coding_type_emacs_mule, | |
83 | coding_type_sjis, | |
84 | coding_type_ccl, | |
85 | coding_type_raw_text, | |
86 | coding_type_undecided, | |
87 | coding_type_max | |
88 | }; | |
4ed46869 | 89 | |
4ed46869 | 90 | |
df7492f9 | 91 | /* Enumeration of end-of-line format type. */ |
4ed46869 | 92 | |
df7492f9 KH |
93 | enum end_of_line_type |
94 | { | |
95 | eol_lf, /* Line-feed only, same as Emacs' internal | |
96 | format. */ | |
97 | eol_crlf, /* Sequence of carriage-return and | |
98 | line-feed. */ | |
99 | eol_cr, /* Carriage-return only. */ | |
100 | eol_any, /* Accept any of above. Produce line-feed | |
101 | only. */ | |
102 | eol_undecided, /* This value is used to denote that the | |
103 | eol-type is not yet undecided. */ | |
104 | eol_type_max | |
105 | }; | |
4ed46869 | 106 | |
df7492f9 | 107 | /* Enumeration of index to an attribute vector of a coding system. */ |
4ed46869 | 108 | |
df7492f9 KH |
109 | enum coding_attr_index |
110 | { | |
111 | coding_attr_base_name, | |
112 | coding_attr_docstring, | |
113 | coding_attr_mnemonic, | |
114 | coding_attr_type, | |
115 | coding_attr_charset_list, | |
116 | coding_attr_ascii_compat, | |
117 | coding_attr_decode_tbl, | |
118 | coding_attr_encode_tbl, | |
404202e7 | 119 | coding_attr_trans_tbl, |
df7492f9 KH |
120 | coding_attr_post_read, |
121 | coding_attr_pre_write, | |
122 | coding_attr_default_char, | |
8f924df7 | 123 | coding_attr_for_unibyte, |
df7492f9 KH |
124 | coding_attr_plist, |
125 | ||
126 | coding_attr_category, | |
127 | coding_attr_safe_charsets, | |
128 | ||
129 | /* The followings are extra attributes for each type. */ | |
130 | coding_attr_charset_valids, | |
131 | ||
132 | coding_attr_ccl_decoder, | |
133 | coding_attr_ccl_encoder, | |
134 | coding_attr_ccl_valids, | |
135 | ||
136 | coding_attr_iso_initial, | |
137 | coding_attr_iso_usage, | |
138 | coding_attr_iso_request, | |
139 | coding_attr_iso_flags, | |
140 | ||
141 | coding_attr_utf_16_bom, | |
142 | coding_attr_utf_16_endian, | |
143 | ||
144 | coding_attr_emacs_mule_full, | |
145 | ||
146 | coding_attr_last_index | |
147 | }; | |
4ed46869 | 148 | |
4ed46869 | 149 | |
df7492f9 KH |
150 | #define CODING_ATTR_BASE_NAME(attrs) AREF (attrs, coding_attr_base_name) |
151 | #define CODING_ATTR_TYPE(attrs) AREF (attrs, coding_attr_type) | |
152 | #define CODING_ATTR_CHARSET_LIST(attrs) AREF (attrs, coding_attr_charset_list) | |
153 | #define CODING_ATTR_MNEMONIC(attrs) AREF (attrs, coding_attr_mnemonic) | |
154 | #define CODING_ATTR_DOCSTRING(attrs) AREF (attrs, coding_attr_docstring) | |
155 | #define CODING_ATTR_ASCII_COMPAT(attrs) AREF (attrs, coding_attr_ascii_compat) | |
156 | #define CODING_ATTR_DECODE_TBL(attrs) AREF (attrs, coding_attr_decode_tbl) | |
157 | #define CODING_ATTR_ENCODE_TBL(attrs) AREF (attrs, coding_attr_encode_tbl) | |
404202e7 | 158 | #define CODING_ATTR_TRANS_TBL(attrs) AREF (attrs, coding_attr_trans_tbl) |
df7492f9 KH |
159 | #define CODING_ATTR_POST_READ(attrs) AREF (attrs, coding_attr_post_read) |
160 | #define CODING_ATTR_PRE_WRITE(attrs) AREF (attrs, coding_attr_pre_write) | |
161 | #define CODING_ATTR_DEFAULT_CHAR(attrs) AREF (attrs, coding_attr_default_char) | |
8f924df7 | 162 | #define CODING_ATTR_FOR_UNIBYTE(attrs) AREF (attrs, coding_attr_for_unibyte) |
df7492f9 KH |
163 | #define CODING_ATTR_FLUSHING(attrs) AREF (attrs, coding_attr_flushing) |
164 | #define CODING_ATTR_PLIST(attrs) AREF (attrs, coding_attr_plist) | |
165 | #define CODING_ATTR_CATEGORY(attrs) AREF (attrs, coding_attr_category) | |
166 | #define CODING_ATTR_SAFE_CHARSETS(attrs)AREF (attrs, coding_attr_safe_charsets) | |
8ddb35b2 | 167 | |
8ddb35b2 | 168 | |
df7492f9 KH |
169 | #define CODING_ID_ATTRS(id) \ |
170 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 0)) | |
fbaa2ed9 | 171 | |
df7492f9 KH |
172 | #define CODING_ID_ALIASES(id) \ |
173 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 1)) | |
c0c69d45 | 174 | |
df7492f9 KH |
175 | #define CODING_ID_EOL_TYPE(id) \ |
176 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 2)) | |
658cc252 | 177 | |
df7492f9 KH |
178 | #define CODING_ID_NAME(id) \ |
179 | (HASH_KEY (XHASH_TABLE (Vcoding_system_hash_table), id)) | |
fbaa2ed9 | 180 | |
df7492f9 KH |
181 | #define CODING_SYSTEM_SPEC(coding_system_symbol) \ |
182 | (Fgethash (coding_system_symbol, Vcoding_system_hash_table, Qnil)) | |
4ed46869 | 183 | |
df7492f9 KH |
184 | #define CODING_SYSTEM_ID(coding_system_symbol) \ |
185 | hash_lookup (XHASH_TABLE (Vcoding_system_hash_table), \ | |
186 | coding_system_symbol, NULL) | |
4ed46869 | 187 | |
df7492f9 KH |
188 | #define CODING_SYSTEM_P(coding_system_symbol) \ |
189 | (! NILP (CODING_SYSTEM_SPEC (coding_system_symbol))) | |
4ed46869 | 190 | |
8f924df7 | 191 | #define CHECK_CODING_SYSTEM(x) \ |
df7492f9 KH |
192 | do { \ |
193 | if (!CODING_SYSTEM_P (x)) \ | |
8f924df7 | 194 | wrong_type_argument (Qcoding_system_p, (x)); \ |
df7492f9 | 195 | } while (0) |
658cc252 | 196 | |
4ed46869 | 197 | |
df7492f9 KH |
198 | #define CHECK_CODING_SYSTEM_GET_SPEC(x, spec) \ |
199 | do { \ | |
200 | spec = CODING_SYSTEM_SPEC (x); \ | |
201 | if (NILP (spec)) \ | |
202 | x = wrong_type_argument (Qcoding_system_p, (x)); \ | |
203 | } while (0) | |
e6de76f8 | 204 | |
8ddb35b2 | 205 | |
df7492f9 KH |
206 | #define CHECK_CODING_SYSTEM_GET_ID(x, id) \ |
207 | do \ | |
208 | { \ | |
209 | id = CODING_SYSTEM_ID (x); \ | |
210 | if (id < 0) \ | |
211 | x = wrong_type_argument (Qcoding_system_p, (x)); \ | |
212 | } while (0) | |
4ed46869 | 213 | |
4ed46869 KH |
214 | |
215 | /*** GENERAL section ***/ | |
216 | ||
df7492f9 KH |
217 | /* Enumeration of result code of code conversion. */ |
218 | enum coding_result_code | |
4ed46869 | 219 | { |
df7492f9 KH |
220 | CODING_RESULT_SUCCESS, |
221 | CODING_RESULT_INSUFFICIENT_SRC, | |
222 | CODING_RESULT_INSUFFICIENT_DST, | |
223 | CODING_RESULT_INCONSISTENT_EOL, | |
63e11478 | 224 | CODING_RESULT_INVALID_SRC, |
df7492f9 KH |
225 | CODING_RESULT_INTERRUPT, |
226 | CODING_RESULT_INSUFFICIENT_MEM | |
4ed46869 KH |
227 | }; |
228 | ||
658cc252 | 229 | |
3b2d77fe | 230 | /* Macros used for the member `mode' of the struct coding_system. */ |
658cc252 KH |
231 | |
232 | /* If set, recover the original CR or LF of the already decoded text | |
233 | when the decoding routine encounters an inconsistent eol format. */ | |
234 | #define CODING_MODE_INHIBIT_INCONSISTENT_EOL 0x01 | |
235 | ||
236 | /* If set, the decoding/encoding routines treat the current data as | |
5998373a | 237 | the last block of the whole text to be converted, and do the |
55496054 | 238 | appropriate finishing job. */ |
658cc252 KH |
239 | #define CODING_MODE_LAST_BLOCK 0x02 |
240 | ||
241 | /* If set, it means that the current source text is in a buffer which | |
242 | enables selective display. */ | |
243 | #define CODING_MODE_SELECTIVE_DISPLAY 0x04 | |
244 | ||
245 | /* This flag is used by the decoding/encoding routines on the fly. If | |
246 | set, it means that right-to-left text is being processed. */ | |
247 | #define CODING_MODE_DIRECTION 0x08 | |
248 | ||
df7492f9 KH |
249 | #define CODING_MODE_FIXED_DESTINATION 0x10 |
250 | ||
251 | #define CODING_MODE_SAFE_ENCODING 0x20 | |
252 | ||
253 | /* Structure of the field `spec.iso_2022' in the structure | |
254 | `coding_system'. */ | |
255 | struct iso_2022_spec | |
4ed46869 | 256 | { |
2ec49574 | 257 | /* Bit-wise-or of CODING_ISO_FLAG_XXX. */ |
df7492f9 | 258 | unsigned flags; |
4ed46869 | 259 | |
df7492f9 KH |
260 | /* The current graphic register invoked to each graphic plane. */ |
261 | int current_invocation[2]; | |
658cc252 | 262 | |
df7492f9 KH |
263 | /* The current charset designated to each graphic register. The |
264 | value -1 means that not charset is designated, -2 means that | |
265 | there was an invalid designation previously. */ | |
266 | int current_designation[4]; | |
4ed46869 | 267 | |
df7492f9 KH |
268 | /* Set to 1 temporarily only when graphic register 2 or 3 is invoked |
269 | by single-shift while encoding. */ | |
270 | int single_shifting; | |
4ed46869 | 271 | |
df7492f9 KH |
272 | /* Set to 1 temporarily only when processing at beginning of line. */ |
273 | int bol; | |
274 | }; | |
4ed46869 | 275 | |
df7492f9 | 276 | struct ccl_spec; |
4ed46869 | 277 | |
df7492f9 KH |
278 | enum utf_16_bom_type |
279 | { | |
280 | utf_16_detect_bom, | |
281 | utf_16_without_bom, | |
282 | utf_16_with_bom | |
283 | }; | |
279d9f7b | 284 | |
df7492f9 KH |
285 | enum utf_16_endian_type |
286 | { | |
287 | utf_16_big_endian, | |
288 | utf_16_little_endian | |
289 | }; | |
279d9f7b | 290 | |
df7492f9 KH |
291 | struct utf_16_spec |
292 | { | |
293 | enum utf_16_bom_type bom; | |
294 | enum utf_16_endian_type endian; | |
295 | int surrogate; | |
296 | }; | |
279d9f7b | 297 | |
4fecac5c KH |
298 | struct coding_detection_info |
299 | { | |
300 | /* Values of these members are bitwise-OR of CATEGORY_MASK_XXXs. */ | |
301 | /* Which categories are already checked. */ | |
302 | int checked; | |
303 | /* Which categories are strongly found. */ | |
304 | int found; | |
305 | /* Which categories are rejected. */ | |
306 | int rejected; | |
307 | }; | |
279d9f7b | 308 | |
279d9f7b | 309 | |
df7492f9 KH |
310 | struct coding_system |
311 | { | |
312 | /* ID number of the coding system. This is an index to | |
313 | Vcoding_system_hash_table. This value is set by | |
314 | setup_coding_system. At the early stage of building time, this | |
315 | value is -1 in the array coding_categories to indicate that no | |
316 | coding-system of that category is yet defined. */ | |
317 | int id; | |
318 | ||
319 | /* Flag bits of the coding system. The meaning of each bit is common | |
320 | to all types of coding systems. */ | |
321 | int common_flags; | |
322 | ||
323 | /* Mode bits of the coding system. See the comments of the macros | |
324 | CODING_MODE_XXX. */ | |
325 | unsigned int mode; | |
450c60a5 | 326 | |
4ed46869 | 327 | /* Detailed information specific to each type of coding system. */ |
df7492f9 | 328 | union |
4ed46869 | 329 | { |
df7492f9 KH |
330 | struct iso_2022_spec iso_2022; |
331 | struct ccl_spec *ccl; /* Defined in ccl.h. */ | |
332 | struct utf_16_spec utf_16; | |
333 | int emacs_mule_full_support; | |
4ed46869 KH |
334 | } spec; |
335 | ||
df7492f9 KH |
336 | int max_charset_id; |
337 | char *safe_charsets; | |
658cc252 | 338 | |
df7492f9 KH |
339 | /* The following two members specify how binary 8-bit code 128..255 |
340 | are represented in source and destination text respectively. 1 | |
341 | means they are represented by 2-byte sequence, 0 means they are | |
342 | represented by 1-byte as is (see the comment in character.h). */ | |
811ea086 KH |
343 | unsigned src_multibyte : 1; |
344 | unsigned dst_multibyte : 1; | |
345 | ||
a137bb00 KH |
346 | /* How may heading bytes we can skip for decoding. This is set to |
347 | -1 in setup_coding_system, and updated by detect_coding. So, | |
348 | when this is equal to the byte length of the text being | |
349 | converted, we can skip the actual conversion process. */ | |
df7492f9 | 350 | int head_ascii; |
658cc252 KH |
351 | |
352 | /* The following members are set by encoding/decoding routine. */ | |
df7492f9 | 353 | EMACS_INT produced, produced_char, consumed, consumed_char; |
658cc252 | 354 | |
811ea086 KH |
355 | /* Number of error source data found in a decoding routine. */ |
356 | int errors; | |
357 | ||
df7492f9 KH |
358 | /* Store the positions of error source data. */ |
359 | EMACS_INT *error_positions; | |
e6a9a0bc | 360 | |
df7492f9 KH |
361 | /* Finish status of code conversion. */ |
362 | enum coding_result_code result; | |
6041c9ce | 363 | |
658cc252 KH |
364 | /* The following members are all Lisp symbols. We don't have to |
365 | protect them from GC because the current garbage collection | |
366 | doesn't relocate Lisp symbols. But, when it is changed, we must | |
367 | find a way to protect them. */ | |
368 | ||
df7492f9 KH |
369 | EMACS_INT src_pos, src_pos_byte, src_chars, src_bytes; |
370 | Lisp_Object src_object; | |
8f924df7 | 371 | const unsigned char *source; |
4ed46869 | 372 | |
df7492f9 KH |
373 | EMACS_INT dst_pos, dst_pos_byte, dst_bytes; |
374 | Lisp_Object dst_object; | |
375 | unsigned char *destination; | |
4ed46869 | 376 | |
df7492f9 | 377 | int chars_at_source; |
4ed46869 | 378 | |
df7492f9 KH |
379 | /* If an element is non-negative, it is a character code. |
380 | ||
381 | If it is in the range -128..-1, it is a 8-bit character code | |
382 | minus 256. | |
383 | ||
384 | If it is less than -128, it specifies the start of an annotation | |
385 | chunk. The length of the chunk is -128 minus the value of the | |
386 | element. The following elements are OFFSET, ANNOTATION-TYPE, and | |
387 | a sequence of actual data for the annotation. OFFSET is a | |
388 | character position offset from dst_pos or src_pos, | |
389 | ANNOTATION-TYPE specfies the meaning of the annotation and how to | |
390 | handle the following data.. */ | |
391 | int *charbuf; | |
392 | int charbuf_size, charbuf_used; | |
393 | ||
394 | /* Set to 1 if charbuf contains an annotation. */ | |
395 | int annotated; | |
4ed46869 | 396 | |
df7492f9 KH |
397 | unsigned char carryover[64]; |
398 | int carryover_bytes; | |
a5ee738b | 399 | |
df7492f9 KH |
400 | int default_char; |
401 | ||
4fecac5c KH |
402 | int (*detector) P_ ((struct coding_system *, |
403 | struct coding_detection_info *)); | |
df7492f9 KH |
404 | void (*decoder) P_ ((struct coding_system *)); |
405 | int (*encoder) P_ ((struct coding_system *)); | |
406 | }; | |
407 | ||
408 | /* Meanings of bits in the member `common_flags' of the structure | |
409 | coding_system. The lowest 8 bits are reserved for various kind of | |
410 | annotations (currently two of them are used). */ | |
411 | #define CODING_ANNOTATION_MASK 0x00FF | |
412 | #define CODING_ANNOTATE_COMPOSITION_MASK 0x0001 | |
413 | #define CODING_ANNOTATE_DIRECTION_MASK 0x0002 | |
4fecac5c | 414 | #define CODING_ANNOTATE_CHARSET_MASK 0x0003 |
df7492f9 KH |
415 | #define CODING_FOR_UNIBYTE_MASK 0x0100 |
416 | #define CODING_REQUIRE_FLUSHING_MASK 0x0200 | |
417 | #define CODING_REQUIRE_DECODING_MASK 0x0400 | |
418 | #define CODING_REQUIRE_ENCODING_MASK 0x0800 | |
419 | #define CODING_REQUIRE_DETECTION_MASK 0x1000 | |
420 | #define CODING_RESET_AT_BOL_MASK 0x2000 | |
421 | ||
422 | /* Return 1 if the coding context CODING requires annotaion | |
423 | handling. */ | |
424 | #define CODING_REQUIRE_ANNOTATION(coding) \ | |
425 | ((coding)->common_flags & CODING_ANNOTATION_MASK) | |
426 | ||
427 | /* Return 1 if the coding context CODING prefers decoding into unibyte. */ | |
428 | #define CODING_FOR_UNIBYTE(coding) \ | |
429 | ((coding)->common_flags & CODING_FOR_UNIBYTE_MASK) | |
430 | ||
431 | /* Return 1 if the coding context CODING requires specific code to be | |
a5ee738b KH |
432 | attached at the tail of converted text. */ |
433 | #define CODING_REQUIRE_FLUSHING(coding) \ | |
434 | ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK) | |
435 | ||
df7492f9 | 436 | /* Return 1 if the coding context CODING requires code conversion on |
a5ee738b KH |
437 | decoding. */ |
438 | #define CODING_REQUIRE_DECODING(coding) \ | |
811ea086 KH |
439 | ((coding)->dst_multibyte \ |
440 | || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK) | |
a5ee738b | 441 | |
df7492f9 KH |
442 | |
443 | /* Return 1 if the coding context CODING requires code conversion on | |
a5ee738b | 444 | encoding. */ |
df7492f9 KH |
445 | #define CODING_REQUIRE_ENCODING(coding) \ |
446 | ((coding)->src_multibyte \ | |
447 | || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK \ | |
448 | || (coding)->mode & CODING_MODE_SELECTIVE_DISPLAY) | |
449 | ||
a5ee738b | 450 | |
df7492f9 | 451 | /* Return 1 if the coding context CODING requires some kind of code |
a5ee738b KH |
452 | detection. */ |
453 | #define CODING_REQUIRE_DETECTION(coding) \ | |
454 | ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK) | |
455 | ||
df7492f9 | 456 | /* Return 1 if the coding context CODING requires code conversion on |
811ea086 | 457 | decoding or some kind of code detection. */ |
658cc252 | 458 | #define CODING_MAY_REQUIRE_DECODING(coding) \ |
811ea086 KH |
459 | (CODING_REQUIRE_DECODING (coding) \ |
460 | || CODING_REQUIRE_DETECTION (coding)) | |
4ed46869 | 461 | |
4ed46869 KH |
462 | /* Macros to decode or encode a character of JISX0208 in SJIS. S1 and |
463 | S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding | |
464 | system. C1 and C2 are the 1st and 2nd position codes of Emacs' | |
465 | internal format. */ | |
466 | ||
df7492f9 KH |
467 | #define SJIS_TO_JIS(code) \ |
468 | do { \ | |
469 | int s1, s2, j1, j2; \ | |
470 | \ | |
471 | s1 = (code) >> 8, s2 = (code) & 0xFF; \ | |
472 | \ | |
473 | if (s2 >= 0x9F) \ | |
474 | (j1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0), \ | |
475 | j2 = s2 - 0x7E); \ | |
476 | else \ | |
477 | (j1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1), \ | |
478 | j2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F)); \ | |
479 | (code) = (j1 << 8) | j2; \ | |
4ed46869 KH |
480 | } while (0) |
481 | ||
6e58724e KH |
482 | #define SJIS_TO_JIS2(code) \ |
483 | do { \ | |
484 | int s1, s2, j1, j2; \ | |
485 | \ | |
486 | s1 = (code) >> 8, s2 = (code) & 0xFF; \ | |
487 | \ | |
488 | if (s2 >= 0x9F) \ | |
489 | { \ | |
490 | j1 = (s1 == 0xF0 ? 0x28 \ | |
491 | : s1 == 0xF1 ? 0x24 \ | |
492 | : s1 == 0xF2 ? 0x2C \ | |
493 | : s1 == 0xF3 ? 0x2E \ | |
494 | : 0x6E + (s1 - 0xF4) * 2); \ | |
495 | j2 = s2 - 0x7E; \ | |
496 | } \ | |
497 | else \ | |
498 | { \ | |
499 | j1 = (s1 <= 0xF2 ? 0x21 + (s1 - 0xF0) * 2 \ | |
500 | : s1 <= 0xF4 ? 0x2D + (s1 - 0xF3) * 2 \ | |
501 | : 0x6F + (s1 - 0xF5) * 2); \ | |
502 | j2 = s2 - ((s2 >= 0x7F ? 0x20 : 0x1F)); \ | |
503 | } \ | |
504 | (code) = (j1 << 8) | j2; \ | |
505 | } while (0) | |
506 | ||
df7492f9 KH |
507 | |
508 | #define JIS_TO_SJIS(code) \ | |
4ed46869 | 509 | do { \ |
df7492f9 KH |
510 | int s1, s2, j1, j2; \ |
511 | \ | |
512 | j1 = (code) >> 8, j2 = (code) & 0xFF; \ | |
513 | if (j1 & 1) \ | |
514 | (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x71 : 0xB1), \ | |
515 | s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F)); \ | |
4ed46869 | 516 | else \ |
df7492f9 KH |
517 | (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x70 : 0xB0), \ |
518 | s2 = j2 + 0x7E); \ | |
5afaefc1 | 519 | (code) = (s1 << 8) | s2; \ |
4ed46869 KH |
520 | } while (0) |
521 | ||
6e58724e KH |
522 | #define JIS_TO_SJIS2(code) \ |
523 | do { \ | |
524 | int s1, s2, j1, j2; \ | |
525 | \ | |
526 | j1 = (code) >> 8, j2 = (code) & 0xFF; \ | |
527 | if (j1 & 1) \ | |
528 | { \ | |
529 | s1 = (j1 <= 0x25 ? 0xF0 + (j1 - 0x21) / 2 \ | |
530 | : j1 <= 0x27 ? 0xF3 + (j1 - 0x2D) / 2 \ | |
531 | : 0xF5 + (j1 - 0x6F) / 2); \ | |
532 | s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F); \ | |
533 | } \ | |
534 | else \ | |
535 | { \ | |
536 | s1 = (j1 == 0x28 ? 0xF0 \ | |
537 | : j1 == 0x24 ? 0xF1 \ | |
538 | : j1 == 0x2C ? 0xF2 \ | |
539 | : j1 == 0x2E ? 0xF3 \ | |
540 | : 0xF4 + (j1 - 0x6E) / 2); \ | |
541 | s2 = j2 + 0x7E; \ | |
542 | } \ | |
543 | (code) = (s1 << 8) | s2; \ | |
544 | } while (0) | |
df7492f9 | 545 | |
290591c8 KH |
546 | /* Encode the file name NAME using the specified coding system |
547 | for file names, if any. */ | |
548 | #define ENCODE_FILE(name) \ | |
549 | (! NILP (Vfile_name_coding_system) \ | |
550 | && XFASTINT (Vfile_name_coding_system) != 0 \ | |
afee9150 | 551 | ? code_convert_string_norecord (name, Vfile_name_coding_system, 1) \ |
290591c8 | 552 | : (! NILP (Vdefault_file_name_coding_system) \ |
afee9150 KH |
553 | && XFASTINT (Vdefault_file_name_coding_system) != 0 \ |
554 | ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \ | |
290591c8 KH |
555 | : name)) |
556 | ||
df7492f9 | 557 | |
290591c8 KH |
558 | /* Decode the file name NAME using the specified coding system |
559 | for file names, if any. */ | |
560 | #define DECODE_FILE(name) \ | |
561 | (! NILP (Vfile_name_coding_system) \ | |
562 | && XFASTINT (Vfile_name_coding_system) != 0 \ | |
afee9150 | 563 | ? code_convert_string_norecord (name, Vfile_name_coding_system, 0) \ |
290591c8 | 564 | : (! NILP (Vdefault_file_name_coding_system) \ |
afee9150 KH |
565 | && XFASTINT (Vdefault_file_name_coding_system) != 0 \ |
566 | ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \ | |
290591c8 KH |
567 | : name)) |
568 | ||
df7492f9 | 569 | |
2dfda962 JR |
570 | #ifdef WINDOWSNT |
571 | /* Encode the string STR using the specified coding system | |
572 | for w32 system functions, if any. */ | |
df7492f9 KH |
573 | #define ENCODE_SYSTEM(str) \ |
574 | (! NILP (Vlocale_coding_system) \ | |
575 | && XFASTINT (Vlocale_coding_system) != 0 \ | |
576 | ? code_convert_string_norecord (str, Vlocale_coding_system, 1) \ | |
2dfda962 JR |
577 | : str) |
578 | ||
579 | /* Decode the string STR using the specified coding system | |
580 | for w32 system functions, if any. */ | |
df7492f9 KH |
581 | #define DECODE_SYSTEM(name) \ |
582 | (! NILP (Vlocale_coding_system) \ | |
583 | && XFASTINT (Vlocale_coding_system) != 0 \ | |
584 | ? code_convert_string_norecord (str, Vlocale_coding_system, 0) \ | |
2dfda962 | 585 | : str) |
cf29bf99 GM |
586 | |
587 | #else /* WINDOWSNT */ | |
588 | ||
589 | #define ENCODE_SYSTEM(str) string_make_unibyte(str) | |
590 | #define DECODE_SYSTEM(name) name | |
591 | ||
592 | #endif /* !WINDOWSNT */ | |
2dfda962 | 593 | |
b3a208b0 DL |
594 | /* Used by the gtk menu code. Note that this encodes utf-8, not |
595 | utf-8-emacs, so it's not a no-op. */ | |
596 | #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1) | |
597 | ||
4ed46869 | 598 | /* Extern declarations. */ |
16869d7e | 599 | extern Lisp_Object code_conversion_save P_ ((int, int)); |
c04809fb AS |
600 | extern int decoding_buffer_size P_ ((struct coding_system *, int)); |
601 | extern int encoding_buffer_size P_ ((struct coding_system *, int)); | |
df7492f9 KH |
602 | extern void setup_coding_system P_ ((Lisp_Object, struct coding_system *)); |
603 | extern void detect_coding P_ ((struct coding_system *)); | |
511dd8e7 | 604 | extern Lisp_Object code_convert_region P_ ((Lisp_Object, Lisp_Object, |
df7492f9 KH |
605 | Lisp_Object, Lisp_Object, |
606 | int, int)); | |
607 | extern Lisp_Object code_convert_string P_ ((Lisp_Object, Lisp_Object, | |
608 | Lisp_Object, int, int, int)); | |
eb545596 DL |
609 | extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object, |
610 | int)); | |
df7492f9 KH |
611 | extern Lisp_Object raw_text_coding_system P_ ((Lisp_Object)); |
612 | extern Lisp_Object coding_inherit_eol_type P_ ((Lisp_Object, Lisp_Object)); | |
613 | ||
614 | extern int decode_coding_gap P_ ((struct coding_system *, | |
615 | EMACS_INT, EMACS_INT)); | |
616 | extern int encode_coding_gap P_ ((struct coding_system *, | |
617 | EMACS_INT, EMACS_INT)); | |
618 | extern void decode_coding_object P_ ((struct coding_system *, | |
619 | Lisp_Object, EMACS_INT, EMACS_INT, | |
620 | EMACS_INT, EMACS_INT, Lisp_Object)); | |
621 | extern void encode_coding_object P_ ((struct coding_system *, | |
622 | Lisp_Object, EMACS_INT, EMACS_INT, | |
623 | EMACS_INT, EMACS_INT, Lisp_Object)); | |
624 | ||
625 | #define decode_coding_region(coding, from, to) \ | |
626 | decode_coding_object (coding, Fcurrent_buffer (), \ | |
627 | from, CHAR_TO_BYTE (from), \ | |
628 | to, CHAR_TO_BYTE (to), Fcurrent_buffer ()) | |
629 | ||
630 | ||
631 | #define encode_coding_region(coding, from, to) \ | |
632 | encode_coding_object (coding, Fcurrent_buffer (), \ | |
633 | from, CHAR_TO_BYTE (from), \ | |
634 | to, CHAR_TO_BYTE (to), Fcurrent_buffer ()) | |
635 | ||
636 | ||
637 | #define decode_coding_string(coding, string, nocopy) \ | |
638 | decode_coding_object (coding, string, 0, 0, XSTRING (string)->size, \ | |
639 | STRING_BYTES (XSTRING (string)), Qt) | |
640 | ||
641 | #define encode_coding_string(coding, string, nocopy) \ | |
642 | (encode_coding_object (coding, string, 0, 0, XSTRING (string)->size, \ | |
643 | STRING_BYTES (XSTRING (string)), Qt), \ | |
644 | (coding)->dst_object) | |
645 | ||
646 | ||
647 | #define decode_coding_c_string(coding, src, bytes, dst_object) \ | |
648 | do { \ | |
649 | (coding)->source = (src); \ | |
650 | (coding)->src_chars = (coding)->src_bytes = (bytes); \ | |
651 | decode_coding_object ((coding), Qnil, 0, 0, (bytes), (bytes), \ | |
652 | (dst_object)); \ | |
653 | } while (0) | |
654 | ||
655 | ||
656 | extern Lisp_Object preferred_coding_system P_ (()); | |
657 | ||
658 | ||
8f924df7 KH |
659 | extern Lisp_Object Qutf_8, Qutf_8_emacs; |
660 | ||
4ed46869 | 661 | extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index; |
df7492f9 KH |
662 | extern Lisp_Object Qcoding_system_p; |
663 | extern Lisp_Object Qraw_text, Qemacs_mule, Qno_conversion, Qundecided; | |
664 | extern Lisp_Object Qiso_2022; | |
4ed46869 | 665 | extern Lisp_Object Qbuffer_file_coding_system; |
df7492f9 KH |
666 | |
667 | extern Lisp_Object Qunix, Qdos, Qmac; | |
4ed46869 | 668 | |
f967223b KH |
669 | extern Lisp_Object Qtranslation_table; |
670 | extern Lisp_Object Qtranslation_table_id; | |
ab45712c | 671 | |
20ee919e EZ |
672 | /* Mnemonic strings to indicate each type of end-of-line. */ |
673 | extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac; | |
674 | /* Mnemonic string to indicate type of end-of-line is not yet decided. */ | |
675 | extern Lisp_Object eol_mnemonic_undecided; | |
4ed46869 | 676 | |
4ed46869 KH |
677 | #ifdef emacs |
678 | extern Lisp_Object Qfile_coding_system; | |
679 | extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument; | |
680 | extern Lisp_Object Qstart_process, Qopen_network_stream; | |
d008a7cc | 681 | extern Lisp_Object Qwrite_region; |
4ed46869 | 682 | |
68c45bf0 PE |
683 | extern char *emacs_strerror P_ ((int)); |
684 | ||
4ed46869 KH |
685 | /* Coding-system for reading files and receiving data from process. */ |
686 | extern Lisp_Object Vcoding_system_for_read; | |
687 | /* Coding-system for writing files and sending data to process. */ | |
688 | extern Lisp_Object Vcoding_system_for_write; | |
689 | /* Coding-system actually used in the latest I/O. */ | |
690 | extern Lisp_Object Vlast_coding_system_used; | |
68c45bf0 PE |
691 | /* Coding-system to use with system messages (e.g. strerror). */ |
692 | extern Lisp_Object Vlocale_coding_system; | |
4ed46869 | 693 | |
77a9bc9a EZ |
694 | /* If non-zero, process buffer inherits the coding system used to decode |
695 | the subprocess output. */ | |
696 | extern int inherit_process_coding_system; | |
697 | ||
4ed46869 KH |
698 | /* Coding-system to be used for encoding terminal output. This |
699 | structure contains information of a coding-system specified by the | |
700 | function `set-terminal-coding-system'. */ | |
701 | extern struct coding_system terminal_coding; | |
702 | ||
fbaa2ed9 KH |
703 | /* Coding system to be used to encode text for terminal display when |
704 | terminal coding system is nil. */ | |
705 | extern struct coding_system safe_terminal_coding; | |
706 | ||
4ed46869 KH |
707 | /* Coding-system of what is sent from terminal keyboard. This |
708 | structure contains information of a coding-system specified by the | |
709 | function `set-keyboard-coding-system'. */ | |
710 | extern struct coding_system keyboard_coding; | |
711 | ||
a5825a24 KH |
712 | /* Default coding systems used for process I/O. */ |
713 | extern Lisp_Object Vdefault_process_coding_system; | |
4ed46869 | 714 | |
df7492f9 | 715 | /* Function to call to force a user to force select a propert coding |
658cc252 KH |
716 | system. */ |
717 | extern Lisp_Object Vselect_safe_coding_system_function; | |
718 | ||
6926d591 KH |
719 | /* If nonzero, on writing a file, Vselect_safe_coding_system_function |
720 | is called even if Vcoding_system_for_write is non-nil. */ | |
721 | extern int coding_system_require_warning; | |
722 | ||
31406df1 RS |
723 | /* Coding system for file names, or nil if none. */ |
724 | extern Lisp_Object Vfile_name_coding_system; | |
725 | ||
726 | /* Coding system for file names used only when | |
727 | Vfile_name_coding_system is nil. */ | |
728 | extern Lisp_Object Vdefault_file_name_coding_system; | |
2dfda962 | 729 | |
4ed46869 KH |
730 | #endif |
731 | ||
d008a7cc GM |
732 | /* Error signaled when there's a problem with detecting coding system */ |
733 | extern Lisp_Object Qcoding_system_error; | |
734 | ||
df7492f9 KH |
735 | extern char emacs_mule_bytes[256]; |
736 | extern int emacs_mule_string_char P_ ((unsigned char *)); | |
737 | ||
6f776e81 | 738 | #endif /* EMACS_CODING_H */ |