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 | |
933373ed KH |
72 | /* Hash table for all coding systems. Keys are coding system symbols |
73 | and values are spec vectors of the corresponding coding system. A | |
74 | spec vector has the form [ ATTRS ALIASES EOL-TYPE ]. ATTRS is a | |
75 | vector of attribute of the coding system. ALIASES is a list of | |
76 | aliases (symbols) of the coding system. EOL-TYPE is `unix', `dos', | |
77 | `mac' or a vector of coding systems (symbols). */ | |
78 | ||
df7492f9 | 79 | extern Lisp_Object Vcoding_system_hash_table; |
4ed46869 | 80 | |
933373ed | 81 | |
df7492f9 | 82 | /* Enumeration of coding system type. */ |
4ed46869 | 83 | |
df7492f9 KH |
84 | enum coding_system_type |
85 | { | |
86 | coding_type_charset, | |
87 | coding_type_utf_8, | |
88 | coding_type_utf_16, | |
89 | coding_type_iso_2022, | |
90 | coding_type_emacs_mule, | |
91 | coding_type_sjis, | |
92 | coding_type_ccl, | |
93 | coding_type_raw_text, | |
94 | coding_type_undecided, | |
95 | coding_type_max | |
96 | }; | |
4ed46869 | 97 | |
4ed46869 | 98 | |
df7492f9 | 99 | /* Enumeration of end-of-line format type. */ |
4ed46869 | 100 | |
df7492f9 KH |
101 | enum end_of_line_type |
102 | { | |
103 | eol_lf, /* Line-feed only, same as Emacs' internal | |
104 | format. */ | |
105 | eol_crlf, /* Sequence of carriage-return and | |
106 | line-feed. */ | |
107 | eol_cr, /* Carriage-return only. */ | |
108 | eol_any, /* Accept any of above. Produce line-feed | |
109 | only. */ | |
110 | eol_undecided, /* This value is used to denote that the | |
111 | eol-type is not yet undecided. */ | |
112 | eol_type_max | |
113 | }; | |
4ed46869 | 114 | |
df7492f9 | 115 | /* Enumeration of index to an attribute vector of a coding system. */ |
4ed46869 | 116 | |
df7492f9 KH |
117 | enum coding_attr_index |
118 | { | |
119 | coding_attr_base_name, | |
120 | coding_attr_docstring, | |
121 | coding_attr_mnemonic, | |
122 | coding_attr_type, | |
123 | coding_attr_charset_list, | |
124 | coding_attr_ascii_compat, | |
125 | coding_attr_decode_tbl, | |
126 | coding_attr_encode_tbl, | |
404202e7 | 127 | coding_attr_trans_tbl, |
df7492f9 KH |
128 | coding_attr_post_read, |
129 | coding_attr_pre_write, | |
130 | coding_attr_default_char, | |
8f924df7 | 131 | coding_attr_for_unibyte, |
df7492f9 KH |
132 | coding_attr_plist, |
133 | ||
134 | coding_attr_category, | |
135 | coding_attr_safe_charsets, | |
136 | ||
137 | /* The followings are extra attributes for each type. */ | |
138 | coding_attr_charset_valids, | |
139 | ||
140 | coding_attr_ccl_decoder, | |
141 | coding_attr_ccl_encoder, | |
142 | coding_attr_ccl_valids, | |
143 | ||
144 | coding_attr_iso_initial, | |
145 | coding_attr_iso_usage, | |
146 | coding_attr_iso_request, | |
147 | coding_attr_iso_flags, | |
148 | ||
149 | coding_attr_utf_16_bom, | |
150 | coding_attr_utf_16_endian, | |
151 | ||
152 | coding_attr_emacs_mule_full, | |
153 | ||
154 | coding_attr_last_index | |
155 | }; | |
4ed46869 | 156 | |
4ed46869 | 157 | |
933373ed KH |
158 | /* Macros to access an element of an attribute vector. */ |
159 | ||
df7492f9 KH |
160 | #define CODING_ATTR_BASE_NAME(attrs) AREF (attrs, coding_attr_base_name) |
161 | #define CODING_ATTR_TYPE(attrs) AREF (attrs, coding_attr_type) | |
162 | #define CODING_ATTR_CHARSET_LIST(attrs) AREF (attrs, coding_attr_charset_list) | |
163 | #define CODING_ATTR_MNEMONIC(attrs) AREF (attrs, coding_attr_mnemonic) | |
164 | #define CODING_ATTR_DOCSTRING(attrs) AREF (attrs, coding_attr_docstring) | |
165 | #define CODING_ATTR_ASCII_COMPAT(attrs) AREF (attrs, coding_attr_ascii_compat) | |
166 | #define CODING_ATTR_DECODE_TBL(attrs) AREF (attrs, coding_attr_decode_tbl) | |
167 | #define CODING_ATTR_ENCODE_TBL(attrs) AREF (attrs, coding_attr_encode_tbl) | |
404202e7 | 168 | #define CODING_ATTR_TRANS_TBL(attrs) AREF (attrs, coding_attr_trans_tbl) |
df7492f9 KH |
169 | #define CODING_ATTR_POST_READ(attrs) AREF (attrs, coding_attr_post_read) |
170 | #define CODING_ATTR_PRE_WRITE(attrs) AREF (attrs, coding_attr_pre_write) | |
171 | #define CODING_ATTR_DEFAULT_CHAR(attrs) AREF (attrs, coding_attr_default_char) | |
8f924df7 | 172 | #define CODING_ATTR_FOR_UNIBYTE(attrs) AREF (attrs, coding_attr_for_unibyte) |
df7492f9 KH |
173 | #define CODING_ATTR_FLUSHING(attrs) AREF (attrs, coding_attr_flushing) |
174 | #define CODING_ATTR_PLIST(attrs) AREF (attrs, coding_attr_plist) | |
175 | #define CODING_ATTR_CATEGORY(attrs) AREF (attrs, coding_attr_category) | |
176 | #define CODING_ATTR_SAFE_CHARSETS(attrs)AREF (attrs, coding_attr_safe_charsets) | |
8ddb35b2 | 177 | |
8ddb35b2 | 178 | |
933373ed KH |
179 | /* Return the name of a coding system specified by ID. */ |
180 | #define CODING_ID_NAME(id) \ | |
181 | (HASH_KEY (XHASH_TABLE (Vcoding_system_hash_table), id)) | |
182 | ||
183 | /* Return the attribute vector of a coding system specified by ID. */ | |
184 | ||
df7492f9 KH |
185 | #define CODING_ID_ATTRS(id) \ |
186 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 0)) | |
fbaa2ed9 | 187 | |
933373ed KH |
188 | /* Return the list of aliases of a coding system specified by ID. */ |
189 | ||
df7492f9 KH |
190 | #define CODING_ID_ALIASES(id) \ |
191 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 1)) | |
c0c69d45 | 192 | |
933373ed KH |
193 | /* Return the eol-type of a coding system specified by ID. */ |
194 | ||
df7492f9 KH |
195 | #define CODING_ID_EOL_TYPE(id) \ |
196 | (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 2)) | |
658cc252 | 197 | |
933373ed KH |
198 | |
199 | /* Return the spec vector of CODING_SYSTEM_SYMBOL. */ | |
fbaa2ed9 | 200 | |
df7492f9 KH |
201 | #define CODING_SYSTEM_SPEC(coding_system_symbol) \ |
202 | (Fgethash (coding_system_symbol, Vcoding_system_hash_table, Qnil)) | |
4ed46869 | 203 | |
933373ed KH |
204 | |
205 | /* Return the ID of CODING_SYSTEM_SYMBOL. */ | |
206 | ||
df7492f9 KH |
207 | #define CODING_SYSTEM_ID(coding_system_symbol) \ |
208 | hash_lookup (XHASH_TABLE (Vcoding_system_hash_table), \ | |
209 | coding_system_symbol, NULL) | |
4ed46869 | 210 | |
933373ed KH |
211 | /* Return 1 iff CODING_SYSTEM_SYMBOL is a coding system. */ |
212 | ||
df7492f9 KH |
213 | #define CODING_SYSTEM_P(coding_system_symbol) \ |
214 | (! NILP (CODING_SYSTEM_SPEC (coding_system_symbol))) | |
4ed46869 | 215 | |
933373ed KH |
216 | /* Check if X is a coding system or not. */ |
217 | ||
8f924df7 | 218 | #define CHECK_CODING_SYSTEM(x) \ |
df7492f9 KH |
219 | do { \ |
220 | if (!CODING_SYSTEM_P (x)) \ | |
8f924df7 | 221 | wrong_type_argument (Qcoding_system_p, (x)); \ |
df7492f9 | 222 | } while (0) |
658cc252 | 223 | |
4ed46869 | 224 | |
933373ed KH |
225 | /* Check if X is a coding system or not. If it is, set SEPC to the |
226 | spec vector of the coding system. */ | |
227 | ||
df7492f9 KH |
228 | #define CHECK_CODING_SYSTEM_GET_SPEC(x, spec) \ |
229 | do { \ | |
230 | spec = CODING_SYSTEM_SPEC (x); \ | |
231 | if (NILP (spec)) \ | |
232 | x = wrong_type_argument (Qcoding_system_p, (x)); \ | |
233 | } while (0) | |
e6de76f8 | 234 | |
8ddb35b2 | 235 | |
933373ed KH |
236 | /* Check if X is a coding system or not. If it is, set ID to the |
237 | ID of the coding system. */ | |
238 | ||
df7492f9 KH |
239 | #define CHECK_CODING_SYSTEM_GET_ID(x, id) \ |
240 | do \ | |
241 | { \ | |
242 | id = CODING_SYSTEM_ID (x); \ | |
243 | if (id < 0) \ | |
244 | x = wrong_type_argument (Qcoding_system_p, (x)); \ | |
245 | } while (0) | |
4ed46869 | 246 | |
4ed46869 KH |
247 | |
248 | /*** GENERAL section ***/ | |
249 | ||
df7492f9 KH |
250 | /* Enumeration of result code of code conversion. */ |
251 | enum coding_result_code | |
4ed46869 | 252 | { |
df7492f9 KH |
253 | CODING_RESULT_SUCCESS, |
254 | CODING_RESULT_INSUFFICIENT_SRC, | |
255 | CODING_RESULT_INSUFFICIENT_DST, | |
256 | CODING_RESULT_INCONSISTENT_EOL, | |
63e11478 | 257 | CODING_RESULT_INVALID_SRC, |
df7492f9 KH |
258 | CODING_RESULT_INTERRUPT, |
259 | CODING_RESULT_INSUFFICIENT_MEM | |
4ed46869 KH |
260 | }; |
261 | ||
658cc252 | 262 | |
3b2d77fe | 263 | /* Macros used for the member `mode' of the struct coding_system. */ |
658cc252 KH |
264 | |
265 | /* If set, recover the original CR or LF of the already decoded text | |
266 | when the decoding routine encounters an inconsistent eol format. */ | |
267 | #define CODING_MODE_INHIBIT_INCONSISTENT_EOL 0x01 | |
268 | ||
269 | /* If set, the decoding/encoding routines treat the current data as | |
5998373a | 270 | the last block of the whole text to be converted, and do the |
55496054 | 271 | appropriate finishing job. */ |
658cc252 KH |
272 | #define CODING_MODE_LAST_BLOCK 0x02 |
273 | ||
274 | /* If set, it means that the current source text is in a buffer which | |
275 | enables selective display. */ | |
276 | #define CODING_MODE_SELECTIVE_DISPLAY 0x04 | |
277 | ||
278 | /* This flag is used by the decoding/encoding routines on the fly. If | |
279 | set, it means that right-to-left text is being processed. */ | |
280 | #define CODING_MODE_DIRECTION 0x08 | |
281 | ||
df7492f9 KH |
282 | #define CODING_MODE_FIXED_DESTINATION 0x10 |
283 | ||
933373ed KH |
284 | /* If set, it means that the encoding routines produces some safe |
285 | ASCII characters (usually '?') for unsupported characters. */ | |
df7492f9 KH |
286 | #define CODING_MODE_SAFE_ENCODING 0x20 |
287 | ||
288 | /* Structure of the field `spec.iso_2022' in the structure | |
289 | `coding_system'. */ | |
290 | struct iso_2022_spec | |
4ed46869 | 291 | { |
2ec49574 | 292 | /* Bit-wise-or of CODING_ISO_FLAG_XXX. */ |
df7492f9 | 293 | unsigned flags; |
4ed46869 | 294 | |
df7492f9 KH |
295 | /* The current graphic register invoked to each graphic plane. */ |
296 | int current_invocation[2]; | |
658cc252 | 297 | |
df7492f9 KH |
298 | /* The current charset designated to each graphic register. The |
299 | value -1 means that not charset is designated, -2 means that | |
300 | there was an invalid designation previously. */ | |
301 | int current_designation[4]; | |
4ed46869 | 302 | |
df7492f9 KH |
303 | /* Set to 1 temporarily only when graphic register 2 or 3 is invoked |
304 | by single-shift while encoding. */ | |
305 | int single_shifting; | |
4ed46869 | 306 | |
df7492f9 KH |
307 | /* Set to 1 temporarily only when processing at beginning of line. */ |
308 | int bol; | |
309 | }; | |
4ed46869 | 310 | |
df7492f9 | 311 | struct ccl_spec; |
4ed46869 | 312 | |
df7492f9 KH |
313 | enum utf_16_bom_type |
314 | { | |
315 | utf_16_detect_bom, | |
316 | utf_16_without_bom, | |
317 | utf_16_with_bom | |
318 | }; | |
279d9f7b | 319 | |
df7492f9 KH |
320 | enum utf_16_endian_type |
321 | { | |
322 | utf_16_big_endian, | |
323 | utf_16_little_endian | |
324 | }; | |
279d9f7b | 325 | |
df7492f9 KH |
326 | struct utf_16_spec |
327 | { | |
328 | enum utf_16_bom_type bom; | |
329 | enum utf_16_endian_type endian; | |
330 | int surrogate; | |
331 | }; | |
279d9f7b | 332 | |
4fecac5c KH |
333 | struct coding_detection_info |
334 | { | |
335 | /* Values of these members are bitwise-OR of CATEGORY_MASK_XXXs. */ | |
336 | /* Which categories are already checked. */ | |
337 | int checked; | |
338 | /* Which categories are strongly found. */ | |
339 | int found; | |
340 | /* Which categories are rejected. */ | |
341 | int rejected; | |
342 | }; | |
279d9f7b | 343 | |
279d9f7b | 344 | |
df7492f9 KH |
345 | struct coding_system |
346 | { | |
347 | /* ID number of the coding system. This is an index to | |
348 | Vcoding_system_hash_table. This value is set by | |
349 | setup_coding_system. At the early stage of building time, this | |
350 | value is -1 in the array coding_categories to indicate that no | |
351 | coding-system of that category is yet defined. */ | |
352 | int id; | |
353 | ||
354 | /* Flag bits of the coding system. The meaning of each bit is common | |
355 | to all types of coding systems. */ | |
356 | int common_flags; | |
357 | ||
358 | /* Mode bits of the coding system. See the comments of the macros | |
359 | CODING_MODE_XXX. */ | |
360 | unsigned int mode; | |
450c60a5 | 361 | |
4ed46869 | 362 | /* Detailed information specific to each type of coding system. */ |
df7492f9 | 363 | union |
4ed46869 | 364 | { |
df7492f9 KH |
365 | struct iso_2022_spec iso_2022; |
366 | struct ccl_spec *ccl; /* Defined in ccl.h. */ | |
367 | struct utf_16_spec utf_16; | |
368 | int emacs_mule_full_support; | |
4ed46869 KH |
369 | } spec; |
370 | ||
df7492f9 KH |
371 | int max_charset_id; |
372 | char *safe_charsets; | |
658cc252 | 373 | |
df7492f9 KH |
374 | /* The following two members specify how binary 8-bit code 128..255 |
375 | are represented in source and destination text respectively. 1 | |
376 | means they are represented by 2-byte sequence, 0 means they are | |
377 | represented by 1-byte as is (see the comment in character.h). */ | |
811ea086 KH |
378 | unsigned src_multibyte : 1; |
379 | unsigned dst_multibyte : 1; | |
380 | ||
a137bb00 KH |
381 | /* How may heading bytes we can skip for decoding. This is set to |
382 | -1 in setup_coding_system, and updated by detect_coding. So, | |
383 | when this is equal to the byte length of the text being | |
384 | converted, we can skip the actual conversion process. */ | |
df7492f9 | 385 | int head_ascii; |
658cc252 KH |
386 | |
387 | /* The following members are set by encoding/decoding routine. */ | |
df7492f9 | 388 | EMACS_INT produced, produced_char, consumed, consumed_char; |
658cc252 | 389 | |
811ea086 KH |
390 | /* Number of error source data found in a decoding routine. */ |
391 | int errors; | |
392 | ||
df7492f9 KH |
393 | /* Store the positions of error source data. */ |
394 | EMACS_INT *error_positions; | |
e6a9a0bc | 395 | |
df7492f9 KH |
396 | /* Finish status of code conversion. */ |
397 | enum coding_result_code result; | |
6041c9ce | 398 | |
658cc252 KH |
399 | /* The following members are all Lisp symbols. We don't have to |
400 | protect them from GC because the current garbage collection | |
401 | doesn't relocate Lisp symbols. But, when it is changed, we must | |
402 | find a way to protect them. */ | |
403 | ||
df7492f9 KH |
404 | EMACS_INT src_pos, src_pos_byte, src_chars, src_bytes; |
405 | Lisp_Object src_object; | |
8f924df7 | 406 | const unsigned char *source; |
4ed46869 | 407 | |
df7492f9 KH |
408 | EMACS_INT dst_pos, dst_pos_byte, dst_bytes; |
409 | Lisp_Object dst_object; | |
410 | unsigned char *destination; | |
4ed46869 | 411 | |
933373ed KH |
412 | /* Set to 1 iff the source of conversion is not in the member |
413 | `charbuf', but at `src_object'. */ | |
df7492f9 | 414 | int chars_at_source; |
4ed46869 | 415 | |
df7492f9 KH |
416 | /* If an element is non-negative, it is a character code. |
417 | ||
418 | If it is in the range -128..-1, it is a 8-bit character code | |
419 | minus 256. | |
420 | ||
421 | If it is less than -128, it specifies the start of an annotation | |
422 | chunk. The length of the chunk is -128 minus the value of the | |
423 | element. The following elements are OFFSET, ANNOTATION-TYPE, and | |
424 | a sequence of actual data for the annotation. OFFSET is a | |
425 | character position offset from dst_pos or src_pos, | |
426 | ANNOTATION-TYPE specfies the meaning of the annotation and how to | |
427 | handle the following data.. */ | |
428 | int *charbuf; | |
429 | int charbuf_size, charbuf_used; | |
430 | ||
431 | /* Set to 1 if charbuf contains an annotation. */ | |
432 | int annotated; | |
4ed46869 | 433 | |
df7492f9 KH |
434 | unsigned char carryover[64]; |
435 | int carryover_bytes; | |
a5ee738b | 436 | |
df7492f9 KH |
437 | int default_char; |
438 | ||
4fecac5c KH |
439 | int (*detector) P_ ((struct coding_system *, |
440 | struct coding_detection_info *)); | |
df7492f9 KH |
441 | void (*decoder) P_ ((struct coding_system *)); |
442 | int (*encoder) P_ ((struct coding_system *)); | |
443 | }; | |
444 | ||
445 | /* Meanings of bits in the member `common_flags' of the structure | |
446 | coding_system. The lowest 8 bits are reserved for various kind of | |
447 | annotations (currently two of them are used). */ | |
448 | #define CODING_ANNOTATION_MASK 0x00FF | |
449 | #define CODING_ANNOTATE_COMPOSITION_MASK 0x0001 | |
450 | #define CODING_ANNOTATE_DIRECTION_MASK 0x0002 | |
4fecac5c | 451 | #define CODING_ANNOTATE_CHARSET_MASK 0x0003 |
df7492f9 KH |
452 | #define CODING_FOR_UNIBYTE_MASK 0x0100 |
453 | #define CODING_REQUIRE_FLUSHING_MASK 0x0200 | |
454 | #define CODING_REQUIRE_DECODING_MASK 0x0400 | |
455 | #define CODING_REQUIRE_ENCODING_MASK 0x0800 | |
456 | #define CODING_REQUIRE_DETECTION_MASK 0x1000 | |
457 | #define CODING_RESET_AT_BOL_MASK 0x2000 | |
458 | ||
459 | /* Return 1 if the coding context CODING requires annotaion | |
460 | handling. */ | |
461 | #define CODING_REQUIRE_ANNOTATION(coding) \ | |
462 | ((coding)->common_flags & CODING_ANNOTATION_MASK) | |
463 | ||
464 | /* Return 1 if the coding context CODING prefers decoding into unibyte. */ | |
465 | #define CODING_FOR_UNIBYTE(coding) \ | |
466 | ((coding)->common_flags & CODING_FOR_UNIBYTE_MASK) | |
467 | ||
468 | /* Return 1 if the coding context CODING requires specific code to be | |
a5ee738b KH |
469 | attached at the tail of converted text. */ |
470 | #define CODING_REQUIRE_FLUSHING(coding) \ | |
471 | ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK) | |
472 | ||
df7492f9 | 473 | /* Return 1 if the coding context CODING requires code conversion on |
a5ee738b KH |
474 | decoding. */ |
475 | #define CODING_REQUIRE_DECODING(coding) \ | |
811ea086 KH |
476 | ((coding)->dst_multibyte \ |
477 | || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK) | |
a5ee738b | 478 | |
df7492f9 KH |
479 | |
480 | /* Return 1 if the coding context CODING requires code conversion on | |
a5ee738b | 481 | encoding. */ |
df7492f9 KH |
482 | #define CODING_REQUIRE_ENCODING(coding) \ |
483 | ((coding)->src_multibyte \ | |
484 | || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK \ | |
485 | || (coding)->mode & CODING_MODE_SELECTIVE_DISPLAY) | |
486 | ||
a5ee738b | 487 | |
df7492f9 | 488 | /* Return 1 if the coding context CODING requires some kind of code |
a5ee738b KH |
489 | detection. */ |
490 | #define CODING_REQUIRE_DETECTION(coding) \ | |
491 | ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK) | |
492 | ||
df7492f9 | 493 | /* Return 1 if the coding context CODING requires code conversion on |
811ea086 | 494 | decoding or some kind of code detection. */ |
658cc252 | 495 | #define CODING_MAY_REQUIRE_DECODING(coding) \ |
811ea086 KH |
496 | (CODING_REQUIRE_DECODING (coding) \ |
497 | || CODING_REQUIRE_DETECTION (coding)) | |
4ed46869 | 498 | |
4ed46869 KH |
499 | /* Macros to decode or encode a character of JISX0208 in SJIS. S1 and |
500 | S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding | |
501 | system. C1 and C2 are the 1st and 2nd position codes of Emacs' | |
502 | internal format. */ | |
503 | ||
df7492f9 KH |
504 | #define SJIS_TO_JIS(code) \ |
505 | do { \ | |
506 | int s1, s2, j1, j2; \ | |
507 | \ | |
508 | s1 = (code) >> 8, s2 = (code) & 0xFF; \ | |
509 | \ | |
510 | if (s2 >= 0x9F) \ | |
511 | (j1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0), \ | |
512 | j2 = s2 - 0x7E); \ | |
513 | else \ | |
514 | (j1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1), \ | |
515 | j2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F)); \ | |
516 | (code) = (j1 << 8) | j2; \ | |
4ed46869 KH |
517 | } while (0) |
518 | ||
6e58724e KH |
519 | #define SJIS_TO_JIS2(code) \ |
520 | do { \ | |
521 | int s1, s2, j1, j2; \ | |
522 | \ | |
523 | s1 = (code) >> 8, s2 = (code) & 0xFF; \ | |
524 | \ | |
525 | if (s2 >= 0x9F) \ | |
526 | { \ | |
527 | j1 = (s1 == 0xF0 ? 0x28 \ | |
528 | : s1 == 0xF1 ? 0x24 \ | |
529 | : s1 == 0xF2 ? 0x2C \ | |
530 | : s1 == 0xF3 ? 0x2E \ | |
531 | : 0x6E + (s1 - 0xF4) * 2); \ | |
532 | j2 = s2 - 0x7E; \ | |
533 | } \ | |
534 | else \ | |
535 | { \ | |
536 | j1 = (s1 <= 0xF2 ? 0x21 + (s1 - 0xF0) * 2 \ | |
537 | : s1 <= 0xF4 ? 0x2D + (s1 - 0xF3) * 2 \ | |
538 | : 0x6F + (s1 - 0xF5) * 2); \ | |
539 | j2 = s2 - ((s2 >= 0x7F ? 0x20 : 0x1F)); \ | |
540 | } \ | |
541 | (code) = (j1 << 8) | j2; \ | |
542 | } while (0) | |
543 | ||
df7492f9 KH |
544 | |
545 | #define JIS_TO_SJIS(code) \ | |
4ed46869 | 546 | do { \ |
df7492f9 KH |
547 | int s1, s2, j1, j2; \ |
548 | \ | |
549 | j1 = (code) >> 8, j2 = (code) & 0xFF; \ | |
550 | if (j1 & 1) \ | |
551 | (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x71 : 0xB1), \ | |
552 | s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F)); \ | |
4ed46869 | 553 | else \ |
df7492f9 KH |
554 | (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x70 : 0xB0), \ |
555 | s2 = j2 + 0x7E); \ | |
5afaefc1 | 556 | (code) = (s1 << 8) | s2; \ |
4ed46869 KH |
557 | } while (0) |
558 | ||
6e58724e KH |
559 | #define JIS_TO_SJIS2(code) \ |
560 | do { \ | |
561 | int s1, s2, j1, j2; \ | |
562 | \ | |
563 | j1 = (code) >> 8, j2 = (code) & 0xFF; \ | |
564 | if (j1 & 1) \ | |
565 | { \ | |
566 | s1 = (j1 <= 0x25 ? 0xF0 + (j1 - 0x21) / 2 \ | |
567 | : j1 <= 0x27 ? 0xF3 + (j1 - 0x2D) / 2 \ | |
568 | : 0xF5 + (j1 - 0x6F) / 2); \ | |
569 | s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F); \ | |
570 | } \ | |
571 | else \ | |
572 | { \ | |
573 | s1 = (j1 == 0x28 ? 0xF0 \ | |
574 | : j1 == 0x24 ? 0xF1 \ | |
575 | : j1 == 0x2C ? 0xF2 \ | |
576 | : j1 == 0x2E ? 0xF3 \ | |
577 | : 0xF4 + (j1 - 0x6E) / 2); \ | |
578 | s2 = j2 + 0x7E; \ | |
579 | } \ | |
580 | (code) = (s1 << 8) | s2; \ | |
581 | } while (0) | |
df7492f9 | 582 | |
290591c8 KH |
583 | /* Encode the file name NAME using the specified coding system |
584 | for file names, if any. */ | |
585 | #define ENCODE_FILE(name) \ | |
586 | (! NILP (Vfile_name_coding_system) \ | |
587 | && XFASTINT (Vfile_name_coding_system) != 0 \ | |
afee9150 | 588 | ? code_convert_string_norecord (name, Vfile_name_coding_system, 1) \ |
290591c8 | 589 | : (! NILP (Vdefault_file_name_coding_system) \ |
afee9150 KH |
590 | && XFASTINT (Vdefault_file_name_coding_system) != 0 \ |
591 | ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \ | |
290591c8 KH |
592 | : name)) |
593 | ||
df7492f9 | 594 | |
290591c8 KH |
595 | /* Decode the file name NAME using the specified coding system |
596 | for file names, if any. */ | |
597 | #define DECODE_FILE(name) \ | |
598 | (! NILP (Vfile_name_coding_system) \ | |
599 | && XFASTINT (Vfile_name_coding_system) != 0 \ | |
afee9150 | 600 | ? code_convert_string_norecord (name, Vfile_name_coding_system, 0) \ |
290591c8 | 601 | : (! NILP (Vdefault_file_name_coding_system) \ |
afee9150 KH |
602 | && XFASTINT (Vdefault_file_name_coding_system) != 0 \ |
603 | ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \ | |
290591c8 KH |
604 | : name)) |
605 | ||
df7492f9 | 606 | |
2dfda962 JR |
607 | #ifdef WINDOWSNT |
608 | /* Encode the string STR using the specified coding system | |
609 | for w32 system functions, if any. */ | |
df7492f9 KH |
610 | #define ENCODE_SYSTEM(str) \ |
611 | (! NILP (Vlocale_coding_system) \ | |
612 | && XFASTINT (Vlocale_coding_system) != 0 \ | |
613 | ? code_convert_string_norecord (str, Vlocale_coding_system, 1) \ | |
2dfda962 JR |
614 | : str) |
615 | ||
616 | /* Decode the string STR using the specified coding system | |
617 | for w32 system functions, if any. */ | |
df7492f9 KH |
618 | #define DECODE_SYSTEM(name) \ |
619 | (! NILP (Vlocale_coding_system) \ | |
620 | && XFASTINT (Vlocale_coding_system) != 0 \ | |
621 | ? code_convert_string_norecord (str, Vlocale_coding_system, 0) \ | |
2dfda962 | 622 | : str) |
cf29bf99 GM |
623 | |
624 | #else /* WINDOWSNT */ | |
625 | ||
626 | #define ENCODE_SYSTEM(str) string_make_unibyte(str) | |
627 | #define DECODE_SYSTEM(name) name | |
628 | ||
629 | #endif /* !WINDOWSNT */ | |
2dfda962 | 630 | |
b3a208b0 DL |
631 | /* Used by the gtk menu code. Note that this encodes utf-8, not |
632 | utf-8-emacs, so it's not a no-op. */ | |
633 | #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1) | |
634 | ||
4ed46869 | 635 | /* Extern declarations. */ |
16869d7e | 636 | extern Lisp_Object code_conversion_save P_ ((int, int)); |
c04809fb AS |
637 | extern int decoding_buffer_size P_ ((struct coding_system *, int)); |
638 | extern int encoding_buffer_size P_ ((struct coding_system *, int)); | |
df7492f9 KH |
639 | extern void setup_coding_system P_ ((Lisp_Object, struct coding_system *)); |
640 | extern void detect_coding P_ ((struct coding_system *)); | |
511dd8e7 | 641 | extern Lisp_Object code_convert_region P_ ((Lisp_Object, Lisp_Object, |
df7492f9 KH |
642 | Lisp_Object, Lisp_Object, |
643 | int, int)); | |
644 | extern Lisp_Object code_convert_string P_ ((Lisp_Object, Lisp_Object, | |
645 | Lisp_Object, int, int, int)); | |
eb545596 DL |
646 | extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object, |
647 | int)); | |
df7492f9 KH |
648 | extern Lisp_Object raw_text_coding_system P_ ((Lisp_Object)); |
649 | extern Lisp_Object coding_inherit_eol_type P_ ((Lisp_Object, Lisp_Object)); | |
650 | ||
651 | extern int decode_coding_gap P_ ((struct coding_system *, | |
652 | EMACS_INT, EMACS_INT)); | |
653 | extern int encode_coding_gap P_ ((struct coding_system *, | |
654 | EMACS_INT, EMACS_INT)); | |
655 | extern void decode_coding_object P_ ((struct coding_system *, | |
656 | Lisp_Object, EMACS_INT, EMACS_INT, | |
657 | EMACS_INT, EMACS_INT, Lisp_Object)); | |
658 | extern void encode_coding_object P_ ((struct coding_system *, | |
659 | Lisp_Object, EMACS_INT, EMACS_INT, | |
660 | EMACS_INT, EMACS_INT, Lisp_Object)); | |
661 | ||
933373ed KH |
662 | /* Macros for backward compatibility. */ |
663 | ||
df7492f9 KH |
664 | #define decode_coding_region(coding, from, to) \ |
665 | decode_coding_object (coding, Fcurrent_buffer (), \ | |
666 | from, CHAR_TO_BYTE (from), \ | |
667 | to, CHAR_TO_BYTE (to), Fcurrent_buffer ()) | |
668 | ||
669 | ||
670 | #define encode_coding_region(coding, from, to) \ | |
671 | encode_coding_object (coding, Fcurrent_buffer (), \ | |
672 | from, CHAR_TO_BYTE (from), \ | |
673 | to, CHAR_TO_BYTE (to), Fcurrent_buffer ()) | |
674 | ||
675 | ||
676 | #define decode_coding_string(coding, string, nocopy) \ | |
677 | decode_coding_object (coding, string, 0, 0, XSTRING (string)->size, \ | |
678 | STRING_BYTES (XSTRING (string)), Qt) | |
679 | ||
680 | #define encode_coding_string(coding, string, nocopy) \ | |
681 | (encode_coding_object (coding, string, 0, 0, XSTRING (string)->size, \ | |
682 | STRING_BYTES (XSTRING (string)), Qt), \ | |
683 | (coding)->dst_object) | |
684 | ||
685 | ||
686 | #define decode_coding_c_string(coding, src, bytes, dst_object) \ | |
687 | do { \ | |
688 | (coding)->source = (src); \ | |
689 | (coding)->src_chars = (coding)->src_bytes = (bytes); \ | |
690 | decode_coding_object ((coding), Qnil, 0, 0, (bytes), (bytes), \ | |
691 | (dst_object)); \ | |
692 | } while (0) | |
693 | ||
694 | ||
695 | extern Lisp_Object preferred_coding_system P_ (()); | |
696 | ||
697 | ||
8f924df7 KH |
698 | extern Lisp_Object Qutf_8, Qutf_8_emacs; |
699 | ||
4ed46869 | 700 | extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index; |
df7492f9 KH |
701 | extern Lisp_Object Qcoding_system_p; |
702 | extern Lisp_Object Qraw_text, Qemacs_mule, Qno_conversion, Qundecided; | |
703 | extern Lisp_Object Qiso_2022; | |
4ed46869 | 704 | extern Lisp_Object Qbuffer_file_coding_system; |
df7492f9 KH |
705 | |
706 | extern Lisp_Object Qunix, Qdos, Qmac; | |
4ed46869 | 707 | |
f967223b KH |
708 | extern Lisp_Object Qtranslation_table; |
709 | extern Lisp_Object Qtranslation_table_id; | |
ab45712c | 710 | |
20ee919e EZ |
711 | /* Mnemonic strings to indicate each type of end-of-line. */ |
712 | extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac; | |
713 | /* Mnemonic string to indicate type of end-of-line is not yet decided. */ | |
714 | extern Lisp_Object eol_mnemonic_undecided; | |
4ed46869 | 715 | |
4ed46869 KH |
716 | #ifdef emacs |
717 | extern Lisp_Object Qfile_coding_system; | |
718 | extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument; | |
719 | extern Lisp_Object Qstart_process, Qopen_network_stream; | |
d008a7cc | 720 | extern Lisp_Object Qwrite_region; |
4ed46869 | 721 | |
68c45bf0 PE |
722 | extern char *emacs_strerror P_ ((int)); |
723 | ||
4ed46869 KH |
724 | /* Coding-system for reading files and receiving data from process. */ |
725 | extern Lisp_Object Vcoding_system_for_read; | |
726 | /* Coding-system for writing files and sending data to process. */ | |
727 | extern Lisp_Object Vcoding_system_for_write; | |
728 | /* Coding-system actually used in the latest I/O. */ | |
729 | extern Lisp_Object Vlast_coding_system_used; | |
68c45bf0 PE |
730 | /* Coding-system to use with system messages (e.g. strerror). */ |
731 | extern Lisp_Object Vlocale_coding_system; | |
4ed46869 | 732 | |
77a9bc9a EZ |
733 | /* If non-zero, process buffer inherits the coding system used to decode |
734 | the subprocess output. */ | |
735 | extern int inherit_process_coding_system; | |
736 | ||
4ed46869 KH |
737 | /* Coding-system to be used for encoding terminal output. This |
738 | structure contains information of a coding-system specified by the | |
739 | function `set-terminal-coding-system'. */ | |
740 | extern struct coding_system terminal_coding; | |
741 | ||
fbaa2ed9 KH |
742 | /* Coding system to be used to encode text for terminal display when |
743 | terminal coding system is nil. */ | |
744 | extern struct coding_system safe_terminal_coding; | |
745 | ||
4ed46869 KH |
746 | /* Coding-system of what is sent from terminal keyboard. This |
747 | structure contains information of a coding-system specified by the | |
748 | function `set-keyboard-coding-system'. */ | |
749 | extern struct coding_system keyboard_coding; | |
750 | ||
a5825a24 KH |
751 | /* Default coding systems used for process I/O. */ |
752 | extern Lisp_Object Vdefault_process_coding_system; | |
4ed46869 | 753 | |
df7492f9 | 754 | /* Function to call to force a user to force select a propert coding |
658cc252 KH |
755 | system. */ |
756 | extern Lisp_Object Vselect_safe_coding_system_function; | |
757 | ||
6926d591 KH |
758 | /* If nonzero, on writing a file, Vselect_safe_coding_system_function |
759 | is called even if Vcoding_system_for_write is non-nil. */ | |
760 | extern int coding_system_require_warning; | |
761 | ||
31406df1 RS |
762 | /* Coding system for file names, or nil if none. */ |
763 | extern Lisp_Object Vfile_name_coding_system; | |
764 | ||
765 | /* Coding system for file names used only when | |
766 | Vfile_name_coding_system is nil. */ | |
767 | extern Lisp_Object Vdefault_file_name_coding_system; | |
2dfda962 | 768 | |
4ed46869 KH |
769 | #endif |
770 | ||
d008a7cc GM |
771 | /* Error signaled when there's a problem with detecting coding system */ |
772 | extern Lisp_Object Qcoding_system_error; | |
773 | ||
df7492f9 KH |
774 | extern char emacs_mule_bytes[256]; |
775 | extern int emacs_mule_string_char P_ ((unsigned char *)); | |
776 | ||
6f776e81 | 777 | #endif /* EMACS_CODING_H */ |