Commit | Line | Data |
---|---|---|
4ed46869 | 1 | /* Header for coding system handler. |
429ab54e | 2 | Copyright (C) 2001, 2002, 2003, 2004, 2005, |
76b6f707 | 3 | 2006, 2007, 2008, 2009 Free Software Foundation, Inc. |
7976eda0 | 4 | Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, |
76b6f707 | 5 | 2005, 2006, 2007, 2008, 2009 |
ce03bf76 KH |
6 | National Institute of Advanced Industrial Science and Technology (AIST) |
7 | Registration Number H14PRO021 | |
8f924df7 | 8 | Copyright (C) 2003 |
df7492f9 KH |
9 | National Institute of Advanced Industrial Science and Technology (AIST) |
10 | Registration Number H13PRO009 | |
4ed46869 | 11 | |
369314dc KH |
12 | This file is part of GNU Emacs. |
13 | ||
b9b1cc14 | 14 | GNU Emacs is free software: you can redistribute it and/or modify |
369314dc | 15 | it under the terms of the GNU General Public License as published by |
b9b1cc14 GM |
16 | the Free Software Foundation, either version 3 of the License, or |
17 | (at your option) any later version. | |
4ed46869 | 18 | |
369314dc KH |
19 | GNU Emacs is distributed in the hope that it will be useful, |
20 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
21 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
22 | GNU General Public License for more details. | |
4ed46869 | 23 | |
369314dc | 24 | You should have received a copy of the GNU General Public License |
b9b1cc14 | 25 | along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */ |
4ed46869 | 26 | |
6f776e81 KH |
27 | #ifndef EMACS_CODING_H |
28 | #define EMACS_CODING_H | |
4ed46869 | 29 | |
df7492f9 | 30 | /* Index to arguments of Fdefine_coding_system_internal. */ |
4ed46869 | 31 | |
df7492f9 KH |
32 | enum define_coding_system_arg_index |
33 | { | |
34 | coding_arg_name, | |
35 | coding_arg_mnemonic, | |
36 | coding_arg_coding_type, | |
37 | coding_arg_charset_list, | |
38 | coding_arg_ascii_compatible_p, | |
39 | coding_arg_decode_translation_table, | |
40 | coding_arg_encode_translation_table, | |
41 | coding_arg_post_read_conversion, | |
42 | coding_arg_pre_write_conversion, | |
43 | coding_arg_default_char, | |
8f924df7 | 44 | coding_arg_for_unibyte, |
df7492f9 KH |
45 | coding_arg_plist, |
46 | coding_arg_eol_type, | |
47 | coding_arg_max | |
48 | }; | |
4ed46869 | 49 | |
df7492f9 | 50 | enum define_coding_iso2022_arg_index |
4ed46869 | 51 | { |
df7492f9 KH |
52 | coding_arg_iso2022_initial = coding_arg_max, |
53 | coding_arg_iso2022_reg_usage, | |
54 | coding_arg_iso2022_request, | |
55 | coding_arg_iso2022_flags, | |
56 | coding_arg_iso2022_max | |
4ed46869 KH |
57 | }; |
58 | ||
e4215ddd KH |
59 | enum define_coding_utf8_arg_index |
60 | { | |
61 | coding_arg_utf8_bom = coding_arg_max, | |
62 | coding_arg_utf8_max | |
63 | }; | |
64 | ||
df7492f9 | 65 | enum define_coding_utf16_arg_index |
4ed46869 | 66 | { |
df7492f9 KH |
67 | coding_arg_utf16_bom = coding_arg_max, |
68 | coding_arg_utf16_endian, | |
69 | coding_arg_utf16_max | |
4ed46869 KH |
70 | }; |
71 | ||
df7492f9 KH |
72 | enum define_coding_ccl_arg_index |
73 | { | |
35d47d18 | 74 | coding_arg_ccl_decoder = coding_arg_max, |
df7492f9 KH |
75 | coding_arg_ccl_encoder, |
76 | coding_arg_ccl_valids, | |
77 | coding_arg_ccl_max | |
78 | }; | |
4ed46869 | 79 | |
933373ed KH |
80 | /* Hash table for all coding systems. Keys are coding system symbols |
81 | and values are spec vectors of the corresponding coding system. A | |
82 | spec vector has the form [ ATTRS ALIASES EOL-TYPE ]. ATTRS is a | |
83 | vector of attribute of the coding system. ALIASES is a list of | |
84 | aliases (symbols) of the coding system. EOL-TYPE is `unix', `dos', | |
85 | `mac' or a vector of coding systems (symbols). */ | |
86 | ||
df7492f9 | 87 | extern Lisp_Object Vcoding_system_hash_table; |
4ed46869 | 88 | |
933373ed | 89 | |
df7492f9 | 90 | /* Enumeration of coding system type. */ |
4ed46869 | 91 | |
df7492f9 KH |
92 | enum coding_system_type |
93 | { | |
94 | coding_type_charset, | |
95 | coding_type_utf_8, | |
96 | coding_type_utf_16, | |
97 | coding_type_iso_2022, | |
98 | coding_type_emacs_mule, | |
99 | coding_type_sjis, | |
100 | coding_type_ccl, | |
101 | coding_type_raw_text, | |
102 | coding_type_undecided, | |
103 | coding_type_max | |
104 | }; | |
4ed46869 | 105 | |
4ed46869 | 106 | |
df7492f9 | 107 | /* Enumeration of end-of-line format type. */ |
4ed46869 | 108 | |
df7492f9 KH |
109 | enum end_of_line_type |
110 | { | |
111 | eol_lf, /* Line-feed only, same as Emacs' internal | |
112 | format. */ | |
113 | eol_crlf, /* Sequence of carriage-return and | |
114 | line-feed. */ | |
115 | eol_cr, /* Carriage-return only. */ | |
116 | eol_any, /* Accept any of above. Produce line-feed | |
117 | only. */ | |
118 | eol_undecided, /* This value is used to denote that the | |
119 | eol-type is not yet undecided. */ | |
120 | eol_type_max | |
121 | }; | |
4ed46869 | 122 | |
df7492f9 | 123 | /* Enumeration of index to an attribute vector of a coding system. */ |
4ed46869 | 124 | |
df7492f9 KH |
125 | enum coding_attr_index |
126 | { | |
127 | coding_attr_base_name, | |
128 | coding_attr_docstring, | |
129 | coding_attr_mnemonic, | |
130 | coding_attr_type, | |
131 | coding_attr_charset_list, | |
132 | coding_attr_ascii_compat, | |
133 | coding_attr_decode_tbl, | |
134 | coding_attr_encode_tbl, | |
404202e7 | 135 | coding_attr_trans_tbl, |
df7492f9 KH |
136 | coding_attr_post_read, |
137 | coding_attr_pre_write, | |
138 | coding_attr_default_char, | |
8f924df7 | 139 | coding_attr_for_unibyte, |
df7492f9 KH |
140 | coding_attr_plist, |
141 | ||
142 | coding_attr_category, | |
143 | coding_attr_safe_charsets, | |
144 | ||
145 | /* The followings are extra attributes for each type. */ | |
146 | coding_attr_charset_valids, | |
147 | ||
148 | coding_attr_ccl_decoder, | |
149 | coding_attr_ccl_encoder, | |
150 | coding_attr_ccl_valids, | |
151 | ||
152 | coding_attr_iso_initial, | |
153 | coding_attr_iso_usage, | |
154 | coding_attr_iso_request, | |
155 | coding_attr_iso_flags, | |
156 | ||
e4215ddd | 157 | coding_attr_utf_bom, |
df7492f9 KH |
158 | coding_attr_utf_16_endian, |
159 | ||
160 | coding_attr_emacs_mule_full, | |
161 | ||
162 | coding_attr_last_index | |
163 | }; | |
4ed46869 | 164 | |
4ed46869 | 165 | |
933373ed KH |
166 | /* Macros to access an element of an attribute vector. */ |
167 | ||
04e05596 JB |
168 | #define CODING_ATTR_BASE_NAME(attrs) AREF (attrs, coding_attr_base_name) |
169 | #define CODING_ATTR_TYPE(attrs) AREF (attrs, coding_attr_type) | |
170 | #define CODING_ATTR_CHARSET_LIST(attrs) AREF (attrs, coding_attr_charset_list) | |
171 | #define CODING_ATTR_MNEMONIC(attrs) AREF (attrs, coding_attr_mnemonic) | |
172 | #define CODING_ATTR_DOCSTRING(attrs) AREF (attrs, coding_attr_docstring) | |
173 | #define CODING_ATTR_ASCII_COMPAT(attrs) AREF (attrs, coding_attr_ascii_compat) | |
174 | #define CODING_ATTR_DECODE_TBL(attrs) AREF (attrs, coding_attr_decode_tbl) | |
175 | #define CODING_ATTR_ENCODE_TBL(attrs) AREF (attrs, coding_attr_encode_tbl) | |
176 | #define CODING_ATTR_TRANS_TBL(attrs) AREF (attrs, coding_attr_trans_tbl) | |
177 | #define CODING_ATTR_POST_READ(attrs) AREF (attrs, coding_attr_post_read) | |
178 | #define CODING_ATTR_PRE_WRITE(attrs) AREF (attrs, coding_attr_pre_write) | |
179 | #define CODING_ATTR_DEFAULT_CHAR(attrs) AREF (attrs, coding_attr_default_char) | |
180 | #define CODING_ATTR_FOR_UNIBYTE(attrs) AREF (attrs, coding_attr_for_unibyte) | |
181 | #define CODING_ATTR_FLUSHING(attrs) AREF (attrs, coding_attr_flushing) | |
182 | #define CODING_ATTR_PLIST(attrs) AREF (attrs, coding_attr_plist) | |
183 | #define CODING_ATTR_CATEGORY(attrs) AREF (attrs, coding_attr_category) | |
184 | #define CODING_ATTR_SAFE_CHARSETS(attrs)AREF (attrs, coding_attr_safe_charsets) | |
8ddb35b2 | 185 | |
8ddb35b2 | 186 | |
933373ed KH |
187 | /* Return the name of a coding system specified by ID. */ |
188 | #define CODING_ID_NAME(id) \ | |
189 | (HASH_KEY (XHASH_TABLE (Vcoding_system_hash_table), id)) | |
190 | ||
191 | /* Return the attribute vector of a coding system specified by ID. */ | |
192 | ||
df7492f9 KH |
193 | #define CODING_ID_ATTRS(id) \ |
194 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 0)) | |
fbaa2ed9 | 195 | |
933373ed KH |
196 | /* Return the list of aliases of a coding system specified by ID. */ |
197 | ||
df7492f9 KH |
198 | #define CODING_ID_ALIASES(id) \ |
199 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 1)) | |
c0c69d45 | 200 | |
933373ed KH |
201 | /* Return the eol-type of a coding system specified by ID. */ |
202 | ||
df7492f9 KH |
203 | #define CODING_ID_EOL_TYPE(id) \ |
204 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 2)) | |
658cc252 | 205 | |
933373ed KH |
206 | |
207 | /* Return the spec vector of CODING_SYSTEM_SYMBOL. */ | |
fbaa2ed9 | 208 | |
df7492f9 KH |
209 | #define CODING_SYSTEM_SPEC(coding_system_symbol) \ |
210 | (Fgethash (coding_system_symbol, Vcoding_system_hash_table, Qnil)) | |
4ed46869 | 211 | |
933373ed KH |
212 | |
213 | /* Return the ID of CODING_SYSTEM_SYMBOL. */ | |
214 | ||
df7492f9 KH |
215 | #define CODING_SYSTEM_ID(coding_system_symbol) \ |
216 | hash_lookup (XHASH_TABLE (Vcoding_system_hash_table), \ | |
217 | coding_system_symbol, NULL) | |
4ed46869 | 218 | |
b2e6b10f | 219 | /* Return 1 if CODING_SYSTEM_SYMBOL is a coding system. */ |
933373ed | 220 | |
7c00e33d KH |
221 | #define CODING_SYSTEM_P(coding_system_symbol) \ |
222 | (CODING_SYSTEM_ID (coding_system_symbol) >= 0 \ | |
223 | || (! NILP (coding_system_symbol) \ | |
224 | && ! NILP (Fcoding_system_p (coding_system_symbol)))) | |
4ed46869 | 225 | |
933373ed KH |
226 | /* Check if X is a coding system or not. */ |
227 | ||
8f924df7 | 228 | #define CHECK_CODING_SYSTEM(x) \ |
df7492f9 | 229 | do { \ |
7c00e33d KH |
230 | if (CODING_SYSTEM_ID (x) < 0 \ |
231 | && NILP (Fcheck_coding_system (x))) \ | |
8f924df7 | 232 | wrong_type_argument (Qcoding_system_p, (x)); \ |
df7492f9 | 233 | } while (0) |
658cc252 | 234 | |
4ed46869 | 235 | |
933373ed KH |
236 | /* Check if X is a coding system or not. If it is, set SEPC to the |
237 | spec vector of the coding system. */ | |
238 | ||
df7492f9 KH |
239 | #define CHECK_CODING_SYSTEM_GET_SPEC(x, spec) \ |
240 | do { \ | |
241 | spec = CODING_SYSTEM_SPEC (x); \ | |
7c00e33d KH |
242 | if (NILP (spec)) \ |
243 | { \ | |
244 | Fcheck_coding_system (x); \ | |
245 | spec = CODING_SYSTEM_SPEC (x); \ | |
246 | } \ | |
df7492f9 | 247 | if (NILP (spec)) \ |
02dfeba8 | 248 | wrong_type_argument (Qcoding_system_p, (x)); \ |
df7492f9 | 249 | } while (0) |
e6de76f8 | 250 | |
8ddb35b2 | 251 | |
933373ed KH |
252 | /* Check if X is a coding system or not. If it is, set ID to the |
253 | ID of the coding system. */ | |
254 | ||
df7492f9 KH |
255 | #define CHECK_CODING_SYSTEM_GET_ID(x, id) \ |
256 | do \ | |
257 | { \ | |
258 | id = CODING_SYSTEM_ID (x); \ | |
7c00e33d KH |
259 | if (id < 0) \ |
260 | { \ | |
261 | Fcheck_coding_system (x); \ | |
262 | id = CODING_SYSTEM_ID (x); \ | |
263 | } \ | |
df7492f9 | 264 | if (id < 0) \ |
02dfeba8 | 265 | wrong_type_argument (Qcoding_system_p, (x)); \ |
df7492f9 | 266 | } while (0) |
4ed46869 | 267 | |
4ed46869 KH |
268 | |
269 | /*** GENERAL section ***/ | |
270 | ||
df7492f9 KH |
271 | /* Enumeration of result code of code conversion. */ |
272 | enum coding_result_code | |
4ed46869 | 273 | { |
df7492f9 KH |
274 | CODING_RESULT_SUCCESS, |
275 | CODING_RESULT_INSUFFICIENT_SRC, | |
276 | CODING_RESULT_INSUFFICIENT_DST, | |
277 | CODING_RESULT_INCONSISTENT_EOL, | |
63e11478 | 278 | CODING_RESULT_INVALID_SRC, |
df7492f9 KH |
279 | CODING_RESULT_INTERRUPT, |
280 | CODING_RESULT_INSUFFICIENT_MEM | |
4ed46869 KH |
281 | }; |
282 | ||
658cc252 | 283 | |
3b2d77fe | 284 | /* Macros used for the member `mode' of the struct coding_system. */ |
658cc252 KH |
285 | |
286 | /* If set, recover the original CR or LF of the already decoded text | |
287 | when the decoding routine encounters an inconsistent eol format. */ | |
288 | #define CODING_MODE_INHIBIT_INCONSISTENT_EOL 0x01 | |
289 | ||
290 | /* If set, the decoding/encoding routines treat the current data as | |
5998373a | 291 | the last block of the whole text to be converted, and do the |
55496054 | 292 | appropriate finishing job. */ |
658cc252 KH |
293 | #define CODING_MODE_LAST_BLOCK 0x02 |
294 | ||
295 | /* If set, it means that the current source text is in a buffer which | |
296 | enables selective display. */ | |
297 | #define CODING_MODE_SELECTIVE_DISPLAY 0x04 | |
298 | ||
299 | /* This flag is used by the decoding/encoding routines on the fly. If | |
300 | set, it means that right-to-left text is being processed. */ | |
301 | #define CODING_MODE_DIRECTION 0x08 | |
302 | ||
df7492f9 KH |
303 | #define CODING_MODE_FIXED_DESTINATION 0x10 |
304 | ||
933373ed KH |
305 | /* If set, it means that the encoding routines produces some safe |
306 | ASCII characters (usually '?') for unsupported characters. */ | |
df7492f9 KH |
307 | #define CODING_MODE_SAFE_ENCODING 0x20 |
308 | ||
825d0875 KH |
309 | /* For handling composition sequence. */ |
310 | #include "composite.h" | |
311 | ||
312 | enum composition_state | |
313 | { | |
314 | COMPOSING_NO, | |
315 | COMPOSING_CHAR, | |
316 | COMPOSING_RULE, | |
317 | COMPOSING_COMPONENT_CHAR, | |
318 | COMPOSING_COMPONENT_RULE | |
319 | }; | |
320 | ||
321 | /* Structure for the current composition status. */ | |
322 | struct composition_status | |
323 | { | |
324 | enum composition_state state; | |
325 | enum composition_method method; | |
326 | int old_form; /* 0:pre-21 form, 1:post-21 form */ | |
327 | int length; /* number of elements produced in charbuf */ | |
328 | int nchars; /* number of characters composed */ | |
329 | int ncomps; /* number of composition components */ | |
330 | /* Maximum carryover is for the case of COMPOSITION_WITH_RULE_ALTCHARS. | |
331 | See the comment in coding.c. */ | |
332 | int carryover[4 /* annotation header */ | |
333 | + MAX_COMPOSITION_COMPONENTS * 3 - 2 /* ALTs and RULEs */ | |
334 | + 2 /* intermediate -1 -1 */ | |
335 | + MAX_COMPOSITION_COMPONENTS /* CHARs */ | |
336 | ]; | |
337 | }; | |
338 | ||
339 | ||
df7492f9 KH |
340 | /* Structure of the field `spec.iso_2022' in the structure |
341 | `coding_system'. */ | |
342 | struct iso_2022_spec | |
4ed46869 | 343 | { |
2ec49574 | 344 | /* Bit-wise-or of CODING_ISO_FLAG_XXX. */ |
df7492f9 | 345 | unsigned flags; |
4ed46869 | 346 | |
df7492f9 KH |
347 | /* The current graphic register invoked to each graphic plane. */ |
348 | int current_invocation[2]; | |
658cc252 | 349 | |
df7492f9 KH |
350 | /* The current charset designated to each graphic register. The |
351 | value -1 means that not charset is designated, -2 means that | |
352 | there was an invalid designation previously. */ | |
353 | int current_designation[4]; | |
4ed46869 | 354 | |
df7492f9 KH |
355 | /* Set to 1 temporarily only when graphic register 2 or 3 is invoked |
356 | by single-shift while encoding. */ | |
357 | int single_shifting; | |
4ed46869 | 358 | |
df7492f9 KH |
359 | /* Set to 1 temporarily only when processing at beginning of line. */ |
360 | int bol; | |
825d0875 KH |
361 | |
362 | /* If positive, we are now scanning CTEXT extended segment. */ | |
363 | int ctext_extended_segment_len; | |
364 | ||
365 | /* If nonzero, we are now scanning embedded UTF-8 sequence. */ | |
366 | int embedded_utf_8; | |
367 | ||
368 | /* The current composition. */ | |
369 | struct composition_status cmp_status; | |
370 | }; | |
371 | ||
372 | struct emacs_mule_spec | |
373 | { | |
374 | int full_support; | |
375 | struct composition_status cmp_status; | |
df7492f9 | 376 | }; |
4ed46869 | 377 | |
df7492f9 | 378 | struct ccl_spec; |
4ed46869 | 379 | |
e4215ddd | 380 | enum utf_bom_type |
df7492f9 | 381 | { |
e4215ddd KH |
382 | utf_detect_bom, |
383 | utf_without_bom, | |
384 | utf_with_bom | |
df7492f9 | 385 | }; |
279d9f7b | 386 | |
df7492f9 KH |
387 | enum utf_16_endian_type |
388 | { | |
389 | utf_16_big_endian, | |
390 | utf_16_little_endian | |
391 | }; | |
279d9f7b | 392 | |
df7492f9 KH |
393 | struct utf_16_spec |
394 | { | |
e4215ddd | 395 | enum utf_bom_type bom; |
df7492f9 KH |
396 | enum utf_16_endian_type endian; |
397 | int surrogate; | |
398 | }; | |
279d9f7b | 399 | |
4fecac5c KH |
400 | struct coding_detection_info |
401 | { | |
402 | /* Values of these members are bitwise-OR of CATEGORY_MASK_XXXs. */ | |
403 | /* Which categories are already checked. */ | |
404 | int checked; | |
405 | /* Which categories are strongly found. */ | |
406 | int found; | |
407 | /* Which categories are rejected. */ | |
408 | int rejected; | |
409 | }; | |
279d9f7b | 410 | |
279d9f7b | 411 | |
df7492f9 KH |
412 | struct coding_system |
413 | { | |
414 | /* ID number of the coding system. This is an index to | |
415 | Vcoding_system_hash_table. This value is set by | |
416 | setup_coding_system. At the early stage of building time, this | |
417 | value is -1 in the array coding_categories to indicate that no | |
418 | coding-system of that category is yet defined. */ | |
419 | int id; | |
420 | ||
421 | /* Flag bits of the coding system. The meaning of each bit is common | |
422 | to all types of coding systems. */ | |
423 | int common_flags; | |
424 | ||
425 | /* Mode bits of the coding system. See the comments of the macros | |
426 | CODING_MODE_XXX. */ | |
427 | unsigned int mode; | |
450c60a5 | 428 | |
4ed46869 | 429 | /* Detailed information specific to each type of coding system. */ |
df7492f9 | 430 | union |
4ed46869 | 431 | { |
df7492f9 KH |
432 | struct iso_2022_spec iso_2022; |
433 | struct ccl_spec *ccl; /* Defined in ccl.h. */ | |
434 | struct utf_16_spec utf_16; | |
e4215ddd | 435 | enum utf_bom_type utf_8_bom; |
825d0875 | 436 | struct emacs_mule_spec emacs_mule; |
4ed46869 KH |
437 | } spec; |
438 | ||
df7492f9 | 439 | int max_charset_id; |
1b3b981b | 440 | unsigned char *safe_charsets; |
658cc252 | 441 | |
df7492f9 KH |
442 | /* The following two members specify how binary 8-bit code 128..255 |
443 | are represented in source and destination text respectively. 1 | |
444 | means they are represented by 2-byte sequence, 0 means they are | |
445 | represented by 1-byte as is (see the comment in character.h). */ | |
811ea086 KH |
446 | unsigned src_multibyte : 1; |
447 | unsigned dst_multibyte : 1; | |
448 | ||
a137bb00 KH |
449 | /* How may heading bytes we can skip for decoding. This is set to |
450 | -1 in setup_coding_system, and updated by detect_coding. So, | |
451 | when this is equal to the byte length of the text being | |
452 | converted, we can skip the actual conversion process. */ | |
df7492f9 | 453 | int head_ascii; |
658cc252 KH |
454 | |
455 | /* The following members are set by encoding/decoding routine. */ | |
df7492f9 | 456 | EMACS_INT produced, produced_char, consumed, consumed_char; |
658cc252 | 457 | |
811ea086 KH |
458 | /* Number of error source data found in a decoding routine. */ |
459 | int errors; | |
460 | ||
df7492f9 KH |
461 | /* Store the positions of error source data. */ |
462 | EMACS_INT *error_positions; | |
e6a9a0bc | 463 | |
df7492f9 KH |
464 | /* Finish status of code conversion. */ |
465 | enum coding_result_code result; | |
6041c9ce | 466 | |
df7492f9 KH |
467 | EMACS_INT src_pos, src_pos_byte, src_chars, src_bytes; |
468 | Lisp_Object src_object; | |
8f924df7 | 469 | const unsigned char *source; |
4ed46869 | 470 | |
df7492f9 KH |
471 | EMACS_INT dst_pos, dst_pos_byte, dst_bytes; |
472 | Lisp_Object dst_object; | |
473 | unsigned char *destination; | |
4ed46869 | 474 | |
b2e6b10f | 475 | /* Set to 1 if the source of conversion is not in the member |
933373ed | 476 | `charbuf', but at `src_object'. */ |
df7492f9 | 477 | int chars_at_source; |
4ed46869 | 478 | |
df7492f9 KH |
479 | /* If an element is non-negative, it is a character code. |
480 | ||
481 | If it is in the range -128..-1, it is a 8-bit character code | |
482 | minus 256. | |
483 | ||
484 | If it is less than -128, it specifies the start of an annotation | |
485 | chunk. The length of the chunk is -128 minus the value of the | |
486 | element. The following elements are OFFSET, ANNOTATION-TYPE, and | |
487 | a sequence of actual data for the annotation. OFFSET is a | |
488 | character position offset from dst_pos or src_pos, | |
489 | ANNOTATION-TYPE specfies the meaning of the annotation and how to | |
490 | handle the following data.. */ | |
491 | int *charbuf; | |
492 | int charbuf_size, charbuf_used; | |
493 | ||
494 | /* Set to 1 if charbuf contains an annotation. */ | |
495 | int annotated; | |
4ed46869 | 496 | |
df7492f9 KH |
497 | unsigned char carryover[64]; |
498 | int carryover_bytes; | |
a5ee738b | 499 | |
df7492f9 KH |
500 | int default_char; |
501 | ||
4fecac5c KH |
502 | int (*detector) P_ ((struct coding_system *, |
503 | struct coding_detection_info *)); | |
df7492f9 KH |
504 | void (*decoder) P_ ((struct coding_system *)); |
505 | int (*encoder) P_ ((struct coding_system *)); | |
506 | }; | |
507 | ||
508 | /* Meanings of bits in the member `common_flags' of the structure | |
509 | coding_system. The lowest 8 bits are reserved for various kind of | |
510 | annotations (currently two of them are used). */ | |
511 | #define CODING_ANNOTATION_MASK 0x00FF | |
512 | #define CODING_ANNOTATE_COMPOSITION_MASK 0x0001 | |
513 | #define CODING_ANNOTATE_DIRECTION_MASK 0x0002 | |
4fecac5c | 514 | #define CODING_ANNOTATE_CHARSET_MASK 0x0003 |
df7492f9 KH |
515 | #define CODING_FOR_UNIBYTE_MASK 0x0100 |
516 | #define CODING_REQUIRE_FLUSHING_MASK 0x0200 | |
517 | #define CODING_REQUIRE_DECODING_MASK 0x0400 | |
518 | #define CODING_REQUIRE_ENCODING_MASK 0x0800 | |
519 | #define CODING_REQUIRE_DETECTION_MASK 0x1000 | |
520 | #define CODING_RESET_AT_BOL_MASK 0x2000 | |
521 | ||
522 | /* Return 1 if the coding context CODING requires annotaion | |
523 | handling. */ | |
524 | #define CODING_REQUIRE_ANNOTATION(coding) \ | |
525 | ((coding)->common_flags & CODING_ANNOTATION_MASK) | |
526 | ||
527 | /* Return 1 if the coding context CODING prefers decoding into unibyte. */ | |
528 | #define CODING_FOR_UNIBYTE(coding) \ | |
529 | ((coding)->common_flags & CODING_FOR_UNIBYTE_MASK) | |
530 | ||
531 | /* Return 1 if the coding context CODING requires specific code to be | |
a5ee738b KH |
532 | attached at the tail of converted text. */ |
533 | #define CODING_REQUIRE_FLUSHING(coding) \ | |
534 | ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK) | |
535 | ||
df7492f9 | 536 | /* Return 1 if the coding context CODING requires code conversion on |
a5ee738b KH |
537 | decoding. */ |
538 | #define CODING_REQUIRE_DECODING(coding) \ | |
811ea086 KH |
539 | ((coding)->dst_multibyte \ |
540 | || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK) | |
a5ee738b | 541 | |
df7492f9 KH |
542 | |
543 | /* Return 1 if the coding context CODING requires code conversion on | |
c198294f KH |
544 | encoding. |
545 | The non-multibyte part of the condition is to support encoding of | |
546 | unibyte strings/buffers generated by string-as-unibyte or | |
547 | (set-buffer-multibyte nil) from multibyte strings/buffers. */ | |
df7492f9 KH |
548 | #define CODING_REQUIRE_ENCODING(coding) \ |
549 | ((coding)->src_multibyte \ | |
550 | || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK \ | |
551 | || (coding)->mode & CODING_MODE_SELECTIVE_DISPLAY) | |
552 | ||
a5ee738b | 553 | |
df7492f9 | 554 | /* Return 1 if the coding context CODING requires some kind of code |
a5ee738b KH |
555 | detection. */ |
556 | #define CODING_REQUIRE_DETECTION(coding) \ | |
557 | ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK) | |
558 | ||
df7492f9 | 559 | /* Return 1 if the coding context CODING requires code conversion on |
811ea086 | 560 | decoding or some kind of code detection. */ |
658cc252 | 561 | #define CODING_MAY_REQUIRE_DECODING(coding) \ |
811ea086 KH |
562 | (CODING_REQUIRE_DECODING (coding) \ |
563 | || CODING_REQUIRE_DETECTION (coding)) | |
4ed46869 | 564 | |
4ed46869 KH |
565 | /* Macros to decode or encode a character of JISX0208 in SJIS. S1 and |
566 | S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding | |
567 | system. C1 and C2 are the 1st and 2nd position codes of Emacs' | |
568 | internal format. */ | |
569 | ||
df7492f9 KH |
570 | #define SJIS_TO_JIS(code) \ |
571 | do { \ | |
572 | int s1, s2, j1, j2; \ | |
573 | \ | |
574 | s1 = (code) >> 8, s2 = (code) & 0xFF; \ | |
575 | \ | |
576 | if (s2 >= 0x9F) \ | |
577 | (j1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0), \ | |
578 | j2 = s2 - 0x7E); \ | |
579 | else \ | |
580 | (j1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1), \ | |
581 | j2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F)); \ | |
582 | (code) = (j1 << 8) | j2; \ | |
4ed46869 KH |
583 | } while (0) |
584 | ||
6e58724e KH |
585 | #define SJIS_TO_JIS2(code) \ |
586 | do { \ | |
587 | int s1, s2, j1, j2; \ | |
588 | \ | |
589 | s1 = (code) >> 8, s2 = (code) & 0xFF; \ | |
590 | \ | |
591 | if (s2 >= 0x9F) \ | |
592 | { \ | |
593 | j1 = (s1 == 0xF0 ? 0x28 \ | |
594 | : s1 == 0xF1 ? 0x24 \ | |
595 | : s1 == 0xF2 ? 0x2C \ | |
596 | : s1 == 0xF3 ? 0x2E \ | |
597 | : 0x6E + (s1 - 0xF4) * 2); \ | |
598 | j2 = s2 - 0x7E; \ | |
599 | } \ | |
600 | else \ | |
601 | { \ | |
602 | j1 = (s1 <= 0xF2 ? 0x21 + (s1 - 0xF0) * 2 \ | |
603 | : s1 <= 0xF4 ? 0x2D + (s1 - 0xF3) * 2 \ | |
604 | : 0x6F + (s1 - 0xF5) * 2); \ | |
605 | j2 = s2 - ((s2 >= 0x7F ? 0x20 : 0x1F)); \ | |
606 | } \ | |
607 | (code) = (j1 << 8) | j2; \ | |
608 | } while (0) | |
609 | ||
df7492f9 KH |
610 | |
611 | #define JIS_TO_SJIS(code) \ | |
4ed46869 | 612 | do { \ |
df7492f9 KH |
613 | int s1, s2, j1, j2; \ |
614 | \ | |
615 | j1 = (code) >> 8, j2 = (code) & 0xFF; \ | |
616 | if (j1 & 1) \ | |
617 | (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x71 : 0xB1), \ | |
618 | s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F)); \ | |
4ed46869 | 619 | else \ |
df7492f9 KH |
620 | (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x70 : 0xB0), \ |
621 | s2 = j2 + 0x7E); \ | |
5afaefc1 | 622 | (code) = (s1 << 8) | s2; \ |
4ed46869 KH |
623 | } while (0) |
624 | ||
6e58724e KH |
625 | #define JIS_TO_SJIS2(code) \ |
626 | do { \ | |
627 | int s1, s2, j1, j2; \ | |
628 | \ | |
629 | j1 = (code) >> 8, j2 = (code) & 0xFF; \ | |
630 | if (j1 & 1) \ | |
631 | { \ | |
632 | s1 = (j1 <= 0x25 ? 0xF0 + (j1 - 0x21) / 2 \ | |
633 | : j1 <= 0x27 ? 0xF3 + (j1 - 0x2D) / 2 \ | |
634 | : 0xF5 + (j1 - 0x6F) / 2); \ | |
635 | s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F); \ | |
636 | } \ | |
637 | else \ | |
638 | { \ | |
639 | s1 = (j1 == 0x28 ? 0xF0 \ | |
640 | : j1 == 0x24 ? 0xF1 \ | |
641 | : j1 == 0x2C ? 0xF2 \ | |
642 | : j1 == 0x2E ? 0xF3 \ | |
643 | : 0xF4 + (j1 - 0x6E) / 2); \ | |
644 | s2 = j2 + 0x7E; \ | |
645 | } \ | |
646 | (code) = (s1 << 8) | s2; \ | |
647 | } while (0) | |
df7492f9 | 648 | |
290591c8 KH |
649 | /* Encode the file name NAME using the specified coding system |
650 | for file names, if any. */ | |
651 | #define ENCODE_FILE(name) \ | |
652 | (! NILP (Vfile_name_coding_system) \ | |
5e92b1ca | 653 | && !EQ (Vfile_name_coding_system, make_number (0)) \ |
afee9150 | 654 | ? code_convert_string_norecord (name, Vfile_name_coding_system, 1) \ |
290591c8 | 655 | : (! NILP (Vdefault_file_name_coding_system) \ |
5e92b1ca | 656 | && !EQ (Vdefault_file_name_coding_system, make_number (0)) \ |
afee9150 | 657 | ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \ |
290591c8 KH |
658 | : name)) |
659 | ||
df7492f9 | 660 | |
290591c8 KH |
661 | /* Decode the file name NAME using the specified coding system |
662 | for file names, if any. */ | |
663 | #define DECODE_FILE(name) \ | |
664 | (! NILP (Vfile_name_coding_system) \ | |
5e92b1ca | 665 | && !EQ (Vfile_name_coding_system, make_number (0)) \ |
afee9150 | 666 | ? code_convert_string_norecord (name, Vfile_name_coding_system, 0) \ |
290591c8 | 667 | : (! NILP (Vdefault_file_name_coding_system) \ |
5e92b1ca | 668 | && !EQ (Vdefault_file_name_coding_system, make_number (0)) \ |
afee9150 | 669 | ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \ |
290591c8 KH |
670 | : name)) |
671 | ||
df7492f9 | 672 | |
2dfda962 | 673 | /* Encode the string STR using the specified coding system |
53eda481 | 674 | for system functions, if any. */ |
2dfda962 | 675 | #define ENCODE_SYSTEM(str) \ |
9b58c683 | 676 | (! NILP (Vlocale_coding_system) \ |
5e92b1ca | 677 | && !EQ (Vlocale_coding_system, make_number (0)) \ |
9b58c683 | 678 | ? code_convert_string_norecord (str, Vlocale_coding_system, 1) \ |
2dfda962 JR |
679 | : str) |
680 | ||
681 | /* Decode the string STR using the specified coding system | |
53eda481 | 682 | for system functions, if any. */ |
581e7427 | 683 | #define DECODE_SYSTEM(str) \ |
9b58c683 | 684 | (! NILP (Vlocale_coding_system) \ |
5e92b1ca | 685 | && !EQ (Vlocale_coding_system, make_number (0)) \ |
9b58c683 | 686 | ? code_convert_string_norecord (str, Vlocale_coding_system, 0) \ |
2dfda962 | 687 | : str) |
cf29bf99 | 688 | |
b3a208b0 DL |
689 | /* Used by the gtk menu code. Note that this encodes utf-8, not |
690 | utf-8-emacs, so it's not a no-op. */ | |
691 | #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1) | |
692 | ||
4ed46869 | 693 | /* Extern declarations. */ |
16869d7e | 694 | extern Lisp_Object code_conversion_save P_ ((int, int)); |
c04809fb AS |
695 | extern int decoding_buffer_size P_ ((struct coding_system *, int)); |
696 | extern int encoding_buffer_size P_ ((struct coding_system *, int)); | |
df7492f9 | 697 | extern void setup_coding_system P_ ((Lisp_Object, struct coding_system *)); |
76c26397 | 698 | extern Lisp_Object coding_charset_list P_ ((struct coding_system *)); |
cde42f0f | 699 | extern Lisp_Object coding_system_charset_list P_ ((Lisp_Object)); |
df7492f9 | 700 | extern void detect_coding P_ ((struct coding_system *)); |
511dd8e7 | 701 | extern Lisp_Object code_convert_region P_ ((Lisp_Object, Lisp_Object, |
df7492f9 KH |
702 | Lisp_Object, Lisp_Object, |
703 | int, int)); | |
704 | extern Lisp_Object code_convert_string P_ ((Lisp_Object, Lisp_Object, | |
705 | Lisp_Object, int, int, int)); | |
eb545596 DL |
706 | extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object, |
707 | int)); | |
df7492f9 KH |
708 | extern Lisp_Object raw_text_coding_system P_ ((Lisp_Object)); |
709 | extern Lisp_Object coding_inherit_eol_type P_ ((Lisp_Object, Lisp_Object)); | |
710 | ||
711 | extern int decode_coding_gap P_ ((struct coding_system *, | |
712 | EMACS_INT, EMACS_INT)); | |
713 | extern int encode_coding_gap P_ ((struct coding_system *, | |
714 | EMACS_INT, EMACS_INT)); | |
715 | extern void decode_coding_object P_ ((struct coding_system *, | |
716 | Lisp_Object, EMACS_INT, EMACS_INT, | |
717 | EMACS_INT, EMACS_INT, Lisp_Object)); | |
718 | extern void encode_coding_object P_ ((struct coding_system *, | |
719 | Lisp_Object, EMACS_INT, EMACS_INT, | |
720 | EMACS_INT, EMACS_INT, Lisp_Object)); | |
721 | ||
933373ed KH |
722 | /* Macros for backward compatibility. */ |
723 | ||
df7492f9 KH |
724 | #define decode_coding_region(coding, from, to) \ |
725 | decode_coding_object (coding, Fcurrent_buffer (), \ | |
726 | from, CHAR_TO_BYTE (from), \ | |
727 | to, CHAR_TO_BYTE (to), Fcurrent_buffer ()) | |
728 | ||
729 | ||
730 | #define encode_coding_region(coding, from, to) \ | |
731 | encode_coding_object (coding, Fcurrent_buffer (), \ | |
732 | from, CHAR_TO_BYTE (from), \ | |
733 | to, CHAR_TO_BYTE (to), Fcurrent_buffer ()) | |
734 | ||
735 | ||
736 | #define decode_coding_string(coding, string, nocopy) \ | |
737 | decode_coding_object (coding, string, 0, 0, XSTRING (string)->size, \ | |
738 | STRING_BYTES (XSTRING (string)), Qt) | |
739 | ||
740 | #define encode_coding_string(coding, string, nocopy) \ | |
741 | (encode_coding_object (coding, string, 0, 0, XSTRING (string)->size, \ | |
742 | STRING_BYTES (XSTRING (string)), Qt), \ | |
743 | (coding)->dst_object) | |
744 | ||
745 | ||
746 | #define decode_coding_c_string(coding, src, bytes, dst_object) \ | |
747 | do { \ | |
748 | (coding)->source = (src); \ | |
749 | (coding)->src_chars = (coding)->src_bytes = (bytes); \ | |
750 | decode_coding_object ((coding), Qnil, 0, 0, (bytes), (bytes), \ | |
751 | (dst_object)); \ | |
752 | } while (0) | |
753 | ||
754 | ||
755 | extern Lisp_Object preferred_coding_system P_ (()); | |
756 | ||
757 | ||
8f924df7 KH |
758 | extern Lisp_Object Qutf_8, Qutf_8_emacs; |
759 | ||
4ed46869 | 760 | extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index; |
df7492f9 KH |
761 | extern Lisp_Object Qcoding_system_p; |
762 | extern Lisp_Object Qraw_text, Qemacs_mule, Qno_conversion, Qundecided; | |
763 | extern Lisp_Object Qiso_2022; | |
4ed46869 | 764 | extern Lisp_Object Qbuffer_file_coding_system; |
df7492f9 KH |
765 | |
766 | extern Lisp_Object Qunix, Qdos, Qmac; | |
4ed46869 | 767 | |
f967223b KH |
768 | extern Lisp_Object Qtranslation_table; |
769 | extern Lisp_Object Qtranslation_table_id; | |
ab45712c | 770 | |
20ee919e EZ |
771 | /* Mnemonic strings to indicate each type of end-of-line. */ |
772 | extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac; | |
773 | /* Mnemonic string to indicate type of end-of-line is not yet decided. */ | |
774 | extern Lisp_Object eol_mnemonic_undecided; | |
4ed46869 | 775 | |
4ed46869 KH |
776 | #ifdef emacs |
777 | extern Lisp_Object Qfile_coding_system; | |
387f6ba5 | 778 | extern Lisp_Object Qcall_process, Qcall_process_region; |
4ed46869 | 779 | extern Lisp_Object Qstart_process, Qopen_network_stream; |
d008a7cc | 780 | extern Lisp_Object Qwrite_region; |
4ed46869 | 781 | |
68c45bf0 PE |
782 | extern char *emacs_strerror P_ ((int)); |
783 | ||
4ed46869 KH |
784 | /* Coding-system for reading files and receiving data from process. */ |
785 | extern Lisp_Object Vcoding_system_for_read; | |
786 | /* Coding-system for writing files and sending data to process. */ | |
787 | extern Lisp_Object Vcoding_system_for_write; | |
788 | /* Coding-system actually used in the latest I/O. */ | |
789 | extern Lisp_Object Vlast_coding_system_used; | |
68c45bf0 PE |
790 | /* Coding-system to use with system messages (e.g. strerror). */ |
791 | extern Lisp_Object Vlocale_coding_system; | |
4ed46869 | 792 | |
77a9bc9a EZ |
793 | /* If non-zero, process buffer inherits the coding system used to decode |
794 | the subprocess output. */ | |
795 | extern int inherit_process_coding_system; | |
796 | ||
fbaa2ed9 KH |
797 | /* Coding system to be used to encode text for terminal display when |
798 | terminal coding system is nil. */ | |
799 | extern struct coding_system safe_terminal_coding; | |
800 | ||
a5825a24 KH |
801 | /* Default coding systems used for process I/O. */ |
802 | extern Lisp_Object Vdefault_process_coding_system; | |
4ed46869 | 803 | |
df7492f9 | 804 | /* Function to call to force a user to force select a propert coding |
658cc252 KH |
805 | system. */ |
806 | extern Lisp_Object Vselect_safe_coding_system_function; | |
807 | ||
6926d591 KH |
808 | /* If nonzero, on writing a file, Vselect_safe_coding_system_function |
809 | is called even if Vcoding_system_for_write is non-nil. */ | |
810 | extern int coding_system_require_warning; | |
811 | ||
31406df1 RS |
812 | /* Coding system for file names, or nil if none. */ |
813 | extern Lisp_Object Vfile_name_coding_system; | |
814 | ||
815 | /* Coding system for file names used only when | |
816 | Vfile_name_coding_system is nil. */ | |
817 | extern Lisp_Object Vdefault_file_name_coding_system; | |
2dfda962 | 818 | |
4ed46869 KH |
819 | #endif |
820 | ||
d008a7cc GM |
821 | /* Error signaled when there's a problem with detecting coding system */ |
822 | extern Lisp_Object Qcoding_system_error; | |
823 | ||
df7492f9 KH |
824 | extern char emacs_mule_bytes[256]; |
825 | extern int emacs_mule_string_char P_ ((unsigned char *)); | |
826 | ||
6f776e81 | 827 | #endif /* EMACS_CODING_H */ |
cefd8c4f KH |
828 | |
829 | /* arch-tag: 2bc3b4fa-6870-4f64-8135-b962b2d290e4 | |
830 | (do not change this comment) */ |