(CP_DEFAULT): Define.
[bpt/emacs.git] / src / coding.h
CommitLineData
4ed46869 1/* Header for coding system handler.
75c8c592
RS
2 Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
3 Licensed to the Free Software Foundation.
4ed46869 4
369314dc
KH
5This file is part of GNU Emacs.
6
7GNU Emacs is free software; you can redistribute it and/or modify
8it under the terms of the GNU General Public License as published by
9the Free Software Foundation; either version 2, or (at your option)
10any later version.
4ed46869 11
369314dc
KH
12GNU Emacs is distributed in the hope that it will be useful,
13but WITHOUT ANY WARRANTY; without even the implied warranty of
14MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15GNU General Public License for more details.
4ed46869 16
369314dc
KH
17You should have received a copy of the GNU General Public License
18along with GNU Emacs; see the file COPYING. If not, write to
19the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20Boston, MA 02111-1307, USA. */
4ed46869
KH
21
22#ifndef _CODING_H
23#define _CODING_H
24
25#ifndef _CCL_H
2ea6666c 26#include "ccl.h"
4ed46869
KH
27#endif
28
0ef69138 29/*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
4ed46869
KH
30
31/* All code (1-byte) of Emacs' internal format is classified into one
32 of the followings. See also `charset.h'. */
33enum emacs_code_class_type
34 {
35 EMACS_control_code, /* Control codes in the range
36 0x00..0x1F and 0x7F except for the
37 following two codes. */
38 EMACS_linefeed_code, /* 0x0A (linefeed) to denote
39 end-of-line. */
40 EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
41 in selective display mode. */
42 EMACS_ascii_code, /* ASCII characters. */
4ed46869
KH
43 EMACS_leading_code_2, /* Base leading code of official
44 TYPE9N character. */
45 EMACS_leading_code_3, /* Base leading code of private TYPE9N
46 or official TYPE9Nx9N character. */
47 EMACS_leading_code_4, /* Base leading code of private
48 TYPE9Nx9N character. */
49 EMACS_invalid_code /* Invalid code, i.e. a base leading
50 code not yet assigned to any
51 charset, or a code of the range
52 0xA0..0xFF. */
53 };
54
55extern enum emacs_code_class_type emacs_code_class[256];
56
57/*** ISO2022 section ***/
58
59/* Macros to define code of control characters for ISO2022's functions. */
60 /* code */ /* function */
61#define ISO_CODE_LF 0x0A /* line-feed */
62#define ISO_CODE_CR 0x0D /* carriage-return */
63#define ISO_CODE_SO 0x0E /* shift-out */
64#define ISO_CODE_SI 0x0F /* shift-in */
65#define ISO_CODE_SS2_7 0x19 /* single-shift-2 for 7-bit code */
66#define ISO_CODE_ESC 0x1B /* escape */
67#define ISO_CODE_SS2 0x8E /* single-shift-2 */
68#define ISO_CODE_SS3 0x8F /* single-shift-3 */
69#define ISO_CODE_CSI 0x9B /* control-sequence-introduce */
70
71/* All code (1-byte) of ISO2022 is classified into one of the
72 followings. */
73enum iso_code_class_type
74 {
811ea086
KH
75 ISO_control_0, /* Control codes in the range
76 0x00..0x1F and 0x7F, except for the
77 following 5 codes. */
4ed46869
KH
78 ISO_carriage_return, /* ISO_CODE_CR (0x0D) */
79 ISO_shift_out, /* ISO_CODE_SO (0x0E) */
80 ISO_shift_in, /* ISO_CODE_SI (0x0F) */
81 ISO_single_shift_2_7, /* ISO_CODE_SS2_7 (0x19) */
82 ISO_escape, /* ISO_CODE_SO (0x1B) */
811ea086
KH
83 ISO_control_1, /* Control codes in the range
84 0x80..0x9F, except for the
85 following 3 codes. */
4ed46869
KH
86 ISO_single_shift_2, /* ISO_CODE_SS2 (0x8E) */
87 ISO_single_shift_3, /* ISO_CODE_SS3 (0x8F) */
88 ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
89 ISO_0x20_or_0x7F, /* Codes of the values 0x20 or 0x7F. */
90 ISO_graphic_plane_0, /* Graphic codes in the range 0x21..0x7E. */
91 ISO_0xA0_or_0xFF, /* Codes of the values 0xA0 or 0xFF. */
92 ISO_graphic_plane_1 /* Graphic codes in the range 0xA1..0xFE. */
93 };
94
95/** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
96 element in the structure `coding_system'. This information is used
97 while encoding a text to ISO2022. **/
98
99/* If set, produce short-form designation sequence (e.g. ESC $ A)
100 instead of long-form sequence (e.g. ESC $ ( A). */
101#define CODING_FLAG_ISO_SHORT_FORM 0x0001
102
103/* If set, reset graphic planes and registers at end-of-line to the
104 initial state. */
105#define CODING_FLAG_ISO_RESET_AT_EOL 0x0002
106
107/* If set, reset graphic planes and registers before any control
108 characters to the initial state. */
109#define CODING_FLAG_ISO_RESET_AT_CNTL 0x0004
110
111/* If set, encode by 7-bit environment. */
112#define CODING_FLAG_ISO_SEVEN_BITS 0x0008
113
114/* If set, use locking-shift function. */
115#define CODING_FLAG_ISO_LOCKING_SHIFT 0x0010
116
117/* If set, use single-shift function. Overwrite
118 CODING_FLAG_ISO_LOCKING_SHIFT. */
119#define CODING_FLAG_ISO_SINGLE_SHIFT 0x0020
120
121/* If set, designate JISX0201-Roman instead of ASCII. */
122#define CODING_FLAG_ISO_USE_ROMAN 0x0040
123
124/* If set, designate JISX0208-1978 instead of JISX0208-1983. */
125#define CODING_FLAG_ISO_USE_OLDJIS 0x0080
126
127/* If set, do not produce ISO6429's direction specifying sequence. */
128#define CODING_FLAG_ISO_NO_DIRECTION 0x0100
129
8ddb35b2
KH
130/* If set, assume designation states are reset at beginning of line on
131 output. */
132#define CODING_FLAG_ISO_INIT_AT_BOL 0x0200
133
134/* If set, designation sequence should be placed at beginning of line
135 on output. */
136#define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
137
1db9ba06 138/* If set, do not encode unsafe charactes on output. */
fbaa2ed9
KH
139#define CODING_FLAG_ISO_SAFE 0x0800
140
c0c69d45
KH
141/* If set, extra latin codes (128..159) are accepted as a valid code
142 on input. */
143#define CODING_FLAG_ISO_LATIN_EXTRA 0x1000
144
658cc252
KH
145/* If set, use designation escape sequence. */
146#define CODING_FLAG_ISO_DESIGNATION 0x10000
147
fbaa2ed9
KH
148/* A character to be produced on output if encoding of the original
149 character is prohibited by CODING_FLAG_ISO_SAFE. */
c0c69d45 150#define CODING_INHIBIT_CHARACTER_SUBSTITUTION 077 /* 077 == `?' */
fbaa2ed9 151
4ed46869
KH
152/* Structure of the field `spec.iso2022' in the structure `coding_system'. */
153struct iso2022_spec
154{
155 /* The current graphic register invoked to each graphic plane. */
156 int current_invocation[2];
157
158 /* The current charset designated to each graphic register. */
159 int current_designation[4];
160
161 /* A charset initially designated to each graphic register. */
162 int initial_designation[4];
163
658cc252
KH
164 /* If not -1, it is a graphic register specified in an invalid
165 designation sequence. */
166 int last_invalid_designation_register;
167
4ed46869 168 /* A graphic register to which each charset should be designated. */
da30d5b1 169 unsigned char requested_designation[MAX_CHARSET + 1];
4ed46869 170
1db9ba06
KH
171 /* A revision number to be specified for each charset on encoding.
172 The value 255 means no revision number for the corresponding
173 charset. */
174 unsigned char charset_revision_number[MAX_CHARSET + 1];
e6de76f8 175
4ed46869
KH
176 /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
177 by single-shift while encoding. */
178 int single_shifting;
8ddb35b2
KH
179
180 /* Set to 1 temporarily only when processing at beginning of line. */
181 int bol;
4ed46869
KH
182};
183
184/* Macros to access each field in the structure `spec.iso2022'. */
185#define CODING_SPEC_ISO_INVOCATION(coding, plane) \
658cc252 186 (coding)->spec.iso2022.current_invocation[plane]
4ed46869 187#define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
658cc252 188 (coding)->spec.iso2022.current_designation[reg]
4ed46869 189#define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
658cc252 190 (coding)->spec.iso2022.initial_designation[reg]
4ed46869 191#define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
658cc252 192 (coding)->spec.iso2022.requested_designation[charset]
1db9ba06 193#define CODING_SPEC_ISO_REVISION_NUMBER(coding, charset) \
658cc252 194 (coding)->spec.iso2022.charset_revision_number[charset]
4ed46869 195#define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
658cc252 196 (coding)->spec.iso2022.single_shifting
8ddb35b2 197#define CODING_SPEC_ISO_BOL(coding) \
658cc252 198 (coding)->spec.iso2022.bol
4ed46869 199
da30d5b1
KH
200/* A value which may appear in
201 coding->spec.iso2022.requested_designation indicating that the
202 corresponding charset does not request any graphic register to be
203 designated. */
204#define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
205
4ed46869
KH
206/* Return a charset which is currently designated to the graphic plane
207 PLANE in the coding-system CODING. */
ceb58510
KH
208#define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane) \
209 ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0) \
210 ? -1 \
211 : CODING_SPEC_ISO_DESIGNATION (coding, \
212 CODING_SPEC_ISO_INVOCATION (coding, plane)))
4ed46869
KH
213
214/*** BIG5 section ***/
215
216/* Macros to denote each type of BIG5 coding system. */
217#define CODING_FLAG_BIG5_HKU 0x00 /* BIG5-HKU is one of variants of
218 BIG5 developed by Hong Kong
219 University. */
220#define CODING_FLAG_BIG5_ETEN 0x01 /* BIG5_ETen is one of variants
221 of BIG5 developed by the
222 company ETen in Taiwan. */
223
224/*** GENERAL section ***/
225
226/* Types of coding system. */
227enum coding_type
228 {
229 coding_type_no_conversion, /* A coding system which requires no
230 conversion for reading and writing
231 including end-of-line format. */
0ef69138 232 coding_type_emacs_mule, /* A coding system used in Emacs'
4ed46869
KH
233 buffer and string. Requires no
234 conversion for reading and writing
235 except for end-of-line format. */
0ef69138 236 coding_type_undecided, /* A coding system which requires
4ed46869
KH
237 automatic detection of a real
238 coding system. */
239 coding_type_sjis, /* SJIS coding system for Japanese. */
240 coding_type_iso2022, /* Any coding system of ISO2022
241 variants. */
242 coding_type_big5, /* BIG5 coding system for Chinese. */
e80de6b1 243 coding_type_ccl, /* The coding system of which decoder
4ed46869 244 and encoder are written in CCL. */
e80de6b1
KH
245 coding_type_raw_text /* A coding system for a text
246 containing ramdom 8-bit code which
247 does not require code conversion
248 except for end-of-line format. */
4ed46869
KH
249 };
250
251/* Formats of end-of-line. */
252#define CODING_EOL_LF 0 /* Line-feed only, same as Emacs'
253 internal format. */
254#define CODING_EOL_CRLF 1 /* Sequence of carriage-return and
255 line-feed. */
256#define CODING_EOL_CR 2 /* Carriage-return only. */
0ef69138 257#define CODING_EOL_UNDECIDED 3 /* This value is used to denote the
4ed46869 258 eol-type is not yet decided. */
e80de6b1
KH
259#define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
260 eol-type is not consistent
261 through the file. */
4ed46869 262
4ed46869 263/* 1 iff composing. */
279d9f7b
KH
264#define COMPOSING_P(coding) ((int) coding->composing > (int) COMPOSITION_NO)
265
266#define COMPOSITION_DATA_SIZE 4080
267#define COMPOSITION_DATA_MAX_BUNCH_LENGTH (4 + MAX_COMPOSITION_COMPONENTS*2)
268
269/* Data structure to hold information about compositions of text that
270 is being decoded or encode. ISO 2022 base code conversion routines
271 handle special ESC sequences for composition specification. But,
272 they can't get/put such information directly from/to a buffer in
273 the deepest place. So, they store or retrieve the information
274 through this structure.
275
276 The encoder stores the information in this structure when it meets
277 ESC sequences for composition while encoding codes, then, after all
278 text codes are encoded, puts `composition' properties on the text
279 by refering the structure.
280
281 The decoder at first stores the information of a text to be
282 decoded, then, while decoding codes, generates ESC sequences for
283 composition at proper places by refering the structure. */
284
285struct composition_data
286{
287 /* The character position of the first character to be encoded or
288 decoded. START and END (see below) are relative to this
289 position. */
290 int char_offset;
291
292 /* The composition data. These elements are repeated for each
293 composition:
294 LENGTH START END METHOD [ COMPONENT ... ]
295 where,
296 LENGTH is the number of elements for this composition.
297
298 START and END are starting and ending character positions of
299 the composition relative to `char_offset'.
300
301 METHOD is one of `enum cmposing_status' specifying the way of
302 composition.
303
304 COMPONENT is a character or an encoded composition rule. */
305 int data[COMPOSITION_DATA_SIZE];
306
307 /* The number of elements in `data' currently used. */
308 int used;
309
310 /* Pointers to the previous and next structures. When `data' is
311 filled up, another structure is allocated and linked in `next'.
312 The new struture has backward link to this struture in `prev'.
313 The number of chaind structures depends on how many compositions
314 the text being encoded or decoded contains. */
315 struct composition_data *prev, *next;
316};
4ed46869 317
658cc252
KH
318/* Macros used for the member finish_status of the struct
319 coding_system. */
320#define CODING_FINISH_NORMAL 0
321#define CODING_FINISH_INSUFFICIENT_SRC 1
322#define CODING_FINISH_INSUFFICIENT_DST 2
323#define CODING_FINISH_INCONSISTENT_EOL 3
279d9f7b
KH
324#define CODING_FINISH_INSUFFICIENT_CMP 4
325#define CODING_FINISH_INTERRUPT 5
658cc252 326
3b2d77fe 327/* Macros used for the member `mode' of the struct coding_system. */
658cc252
KH
328
329/* If set, recover the original CR or LF of the already decoded text
330 when the decoding routine encounters an inconsistent eol format. */
331#define CODING_MODE_INHIBIT_INCONSISTENT_EOL 0x01
332
333/* If set, the decoding/encoding routines treat the current data as
334 the last block of the whole text to be converted, and do
335 appropriate fisishing job. */
336#define CODING_MODE_LAST_BLOCK 0x02
337
338/* If set, it means that the current source text is in a buffer which
339 enables selective display. */
340#define CODING_MODE_SELECTIVE_DISPLAY 0x04
341
342/* This flag is used by the decoding/encoding routines on the fly. If
343 set, it means that right-to-left text is being processed. */
344#define CODING_MODE_DIRECTION 0x08
345
4ed46869
KH
346struct coding_system
347{
348 /* Type of the coding system. */
349 enum coding_type type;
350
658cc252
KH
351 /* Type of end-of-line format (LF, CRLF, or CR) of the coding system. */
352 int eol_type;
353
a5ee738b 354 /* Flag bits of the coding system. The meaning of each bit is common
658cc252 355 to all types of coding systems. */
a5ee738b 356 unsigned int common_flags;
4ed46869
KH
357
358 /* Flag bits of the coding system. The meaning of each bit depends
359 on the type of the coding system. */
360 unsigned int flags;
361
658cc252
KH
362 /* Mode bits of the coding system. See the comments of the macros
363 CODING_MODE_XXX. */
364 unsigned int mode;
4ed46869 365
1db9ba06 366 /* Table of safe character sets for this coding system. If the Nth
658cc252 367 element is 0, the charset of ID N is not a safe character set.
1db9ba06
KH
368 Such a character set is not encoded when CODING_ISO_FLAG_SAFE is
369 set. */
370 unsigned char safe_charsets[MAX_CHARSET + 1];
371
279d9f7b 372 /* The current status of composition handling. */
4ed46869
KH
373 int composing;
374
279d9f7b
KH
375 /* 1 iff the next character is a composition rule. */
376 int composition_rule_follows;
377
378 /* Information of compositions are stored here on decoding and set
379 in advance on encoding. */
380 struct composition_data *cmp_data;
381
382 /* Index to cmp_data->data for the first element for the current
383 composition. */
384 int cmp_data_start;
385
386 /* Index to cmp_data->data for the current element for the current
387 composition. */
388 int cmp_data_index;
450c60a5 389
4ed46869
KH
390 /* Detailed information specific to each type of coding system. */
391 union spec
392 {
393 struct iso2022_spec iso2022;
394 struct ccl_spec ccl; /* Defined in ccl.h. */
395 } spec;
396
658cc252
KH
397 /* Index number of coding category of the coding system. */
398 int category_idx;
399
d23ee514
KH
400 /* The following two members specify how characters 128..159 are
401 represented in source and destination text respectively. 1 means
402 they are represented by 2-byte sequence, 0 means they are
403 represented by 1-byte as is (see the comment in charset.h). */
811ea086
KH
404 unsigned src_multibyte : 1;
405 unsigned dst_multibyte : 1;
406
a137bb00
KH
407 /* How may heading bytes we can skip for decoding. This is set to
408 -1 in setup_coding_system, and updated by detect_coding. So,
409 when this is equal to the byte length of the text being
410 converted, we can skip the actual conversion process. */
658cc252
KH
411 int heading_ascii;
412
413 /* The following members are set by encoding/decoding routine. */
414 int produced, produced_char, consumed, consumed_char;
415
811ea086
KH
416 /* Number of error source data found in a decoding routine. */
417 int errors;
418
419 /* Finish status of code conversion. It should be one of macros
420 CODING_FINISH_XXXX. */
421 int result;
e6a9a0bc 422
658cc252
KH
423 /* The following members are all Lisp symbols. We don't have to
424 protect them from GC because the current garbage collection
425 doesn't relocate Lisp symbols. But, when it is changed, we must
426 find a way to protect them. */
427
4ed46869
KH
428 /* Backward pointer to the Lisp symbol of the coding system. */
429 Lisp_Object symbol;
430
431 /* Lisp function (symbol) to be called after decoding to do
658cc252 432 additional conversion, or nil. */
4ed46869
KH
433 Lisp_Object post_read_conversion;
434
435 /* Lisp function (symbol) to be called before encoding to do
658cc252 436 additional conversion, or nil. */
4ed46869
KH
437 Lisp_Object pre_write_conversion;
438
ab45712c 439 /* Character translation tables to look up, or nil. */
f967223b
KH
440 Lisp_Object translation_table_for_decode;
441 Lisp_Object translation_table_for_encode;
4ed46869
KH
442};
443
a5ee738b
KH
444#define CODING_REQUIRE_FLUSHING_MASK 1
445#define CODING_REQUIRE_DECODING_MASK 2
446#define CODING_REQUIRE_ENCODING_MASK 4
447#define CODING_REQUIRE_DETECTION_MASK 8
448
449/* Return 1 if the coding system CODING requires specific code to be
450 attached at the tail of converted text. */
451#define CODING_REQUIRE_FLUSHING(coding) \
452 ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
453
454/* Return 1 if the coding system CODING requires code conversion on
455 decoding. */
456#define CODING_REQUIRE_DECODING(coding) \
811ea086
KH
457 ((coding)->dst_multibyte \
458 || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
a5ee738b
KH
459
460/* Return 1 if the coding system CODING requires code conversion on
461 encoding. */
462#define CODING_REQUIRE_ENCODING(coding) \
811ea086
KH
463 ((coding)->src_multibyte \
464 || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
a5ee738b
KH
465
466/* Return 1 if the coding system CODING requires some kind of code
467 detection. */
468#define CODING_REQUIRE_DETECTION(coding) \
469 ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
470
811ea086
KH
471/* Return 1 if the coding system CODING requires code conversion on
472 decoding or some kind of code detection. */
658cc252 473#define CODING_MAY_REQUIRE_DECODING(coding) \
811ea086
KH
474 (CODING_REQUIRE_DECODING (coding) \
475 || CODING_REQUIRE_DETECTION (coding))
4ed46869
KH
476
477/* Index for each coding category in `coding_category_table' */
0ef69138 478#define CODING_CATEGORY_IDX_EMACS_MULE 0
4ed46869
KH
479#define CODING_CATEGORY_IDX_SJIS 1
480#define CODING_CATEGORY_IDX_ISO_7 2
658cc252
KH
481#define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
482#define CODING_CATEGORY_IDX_ISO_8_1 4
483#define CODING_CATEGORY_IDX_ISO_8_2 5
484#define CODING_CATEGORY_IDX_ISO_7_ELSE 6
485#define CODING_CATEGORY_IDX_ISO_8_ELSE 7
8469bb88
KH
486#define CODING_CATEGORY_IDX_CCL 8
487#define CODING_CATEGORY_IDX_BIG5 9
62537270
KH
488#define CODING_CATEGORY_IDX_UTF_8 10
489#define CODING_CATEGORY_IDX_UTF_16_BE 11
490#define CODING_CATEGORY_IDX_UTF_16_LE 12
491#define CODING_CATEGORY_IDX_RAW_TEXT 13
492#define CODING_CATEGORY_IDX_BINARY 14
493#define CODING_CATEGORY_IDX_MAX 15
4ed46869
KH
494
495/* Definitions of flag bits returned by the function
496 detect_coding_mask (). */
0ef69138 497#define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
4ed46869
KH
498#define CODING_CATEGORY_MASK_SJIS (1 << CODING_CATEGORY_IDX_SJIS)
499#define CODING_CATEGORY_MASK_ISO_7 (1 << CODING_CATEGORY_IDX_ISO_7)
658cc252 500#define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
4ed46869
KH
501#define CODING_CATEGORY_MASK_ISO_8_1 (1 << CODING_CATEGORY_IDX_ISO_8_1)
502#define CODING_CATEGORY_MASK_ISO_8_2 (1 << CODING_CATEGORY_IDX_ISO_8_2)
a38ede41
KH
503#define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
504#define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
8469bb88 505#define CODING_CATEGORY_MASK_CCL (1 << CODING_CATEGORY_IDX_CCL)
4ed46869 506#define CODING_CATEGORY_MASK_BIG5 (1 << CODING_CATEGORY_IDX_BIG5)
62537270
KH
507#define CODING_CATEGORY_MASK_UTF_8 (1 << CODING_CATEGORY_IDX_UTF_8)
508#define CODING_CATEGORY_MASK_UTF_16_BE (1 << CODING_CATEGORY_IDX_UTF_16_BE)
509#define CODING_CATEGORY_MASK_UTF_16_LE (1 << CODING_CATEGORY_IDX_UTF_16_LE)
f1651811 510#define CODING_CATEGORY_MASK_RAW_TEXT (1 << CODING_CATEGORY_IDX_RAW_TEXT)
1f312d8a 511#define CODING_CATEGORY_MASK_BINARY (1 << CODING_CATEGORY_IDX_BINARY)
4ed46869
KH
512
513/* This value is returned if detect_coding_mask () find nothing other
514 than ASCII characters. */
515#define CODING_CATEGORY_MASK_ANY \
0ef69138 516 ( CODING_CATEGORY_MASK_EMACS_MULE \
4ed46869
KH
517 | CODING_CATEGORY_MASK_SJIS \
518 | CODING_CATEGORY_MASK_ISO_7 \
658cc252 519 | CODING_CATEGORY_MASK_ISO_7_TIGHT \
4ed46869
KH
520 | CODING_CATEGORY_MASK_ISO_8_1 \
521 | CODING_CATEGORY_MASK_ISO_8_2 \
a38ede41
KH
522 | CODING_CATEGORY_MASK_ISO_7_ELSE \
523 | CODING_CATEGORY_MASK_ISO_8_ELSE \
8469bb88 524 | CODING_CATEGORY_MASK_CCL \
62537270
KH
525 | CODING_CATEGORY_MASK_BIG5 \
526 | CODING_CATEGORY_MASK_UTF_8 \
527 | CODING_CATEGORY_MASK_UTF_16_BE \
528 | CODING_CATEGORY_MASK_UTF_16_LE)
4ed46869 529
658cc252
KH
530#define CODING_CATEGORY_MASK_ISO_7BIT \
531 (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
532
533#define CODING_CATEGORY_MASK_ISO_8BIT \
534 (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
535
536#define CODING_CATEGORY_MASK_ISO_SHIFT \
537 (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
538
539#define CODING_CATEGORY_MASK_ISO \
540 ( CODING_CATEGORY_MASK_ISO_7BIT \
541 | CODING_CATEGORY_MASK_ISO_SHIFT \
542 | CODING_CATEGORY_MASK_ISO_8BIT)
543
62537270
KH
544#define CODING_CATEGORY_MASK_UTF_16_BE_LE \
545 (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
546
4ed46869
KH
547/* Macros to decode or encode a character of JISX0208 in SJIS. S1 and
548 S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
549 system. C1 and C2 are the 1st and 2nd position codes of Emacs'
550 internal format. */
551
552#define DECODE_SJIS(s1, s2, c1, c2) \
553 do { \
554 if (s2 >= 0x9F) \
555 c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0), \
556 c2 = s2 - 0x7E; \
557 else \
558 c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1), \
559 c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F); \
560 } while (0)
561
562#define ENCODE_SJIS(c1, c2, s1, s2) \
563 do { \
564 if (c1 & 1) \
565 s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1), \
566 s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F); \
567 else \
568 s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0), \
569 s2 = c2 + 0x7E; \
570 } while (0)
571
290591c8
KH
572/* Encode the file name NAME using the specified coding system
573 for file names, if any. */
574#define ENCODE_FILE(name) \
575 (! NILP (Vfile_name_coding_system) \
576 && XFASTINT (Vfile_name_coding_system) != 0 \
afee9150 577 ? code_convert_string_norecord (name, Vfile_name_coding_system, 1) \
290591c8 578 : (! NILP (Vdefault_file_name_coding_system) \
afee9150
KH
579 && XFASTINT (Vdefault_file_name_coding_system) != 0 \
580 ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
290591c8
KH
581 : name))
582
583/* Decode the file name NAME using the specified coding system
584 for file names, if any. */
585#define DECODE_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, 0) \
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, 0) \
290591c8
KH
592 : name))
593
4ed46869 594/* Extern declarations. */
c04809fb 595extern int decode_coding P_ ((struct coding_system *, unsigned char *,
658cc252 596 unsigned char *, int, int));
c04809fb 597extern int encode_coding P_ ((struct coding_system *, unsigned char *,
658cc252 598 unsigned char *, int, int));
279d9f7b
KH
599extern void coding_save_composition P_ ((struct coding_system *, int, int,
600 Lisp_Object));
601extern void coding_free_composition_data P_ ((struct coding_system *));
602extern void coding_adjust_composition_offset P_ ((struct coding_system *,
603 int));
55fff3f9
KH
604extern void coding_allocate_composition_data P_ ((struct coding_system *,
605 int));
606extern void coding_restore_composition P_ ((struct coding_system *,
607 Lisp_Object));
ec5d8db7
AS
608extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
609 int, int));
c04809fb
AS
610extern int decoding_buffer_size P_ ((struct coding_system *, int));
611extern int encoding_buffer_size P_ ((struct coding_system *, int));
612extern void detect_coding P_ ((struct coding_system *, unsigned char *, int));
613extern void detect_eol P_ ((struct coding_system *, unsigned char *, int));
4ed46869 614extern int conversion_buffer_size;
c04809fb
AS
615extern char *conversion_buffer;
616extern char *get_conversion_buffer P_ ((int));
617extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
8000e212
KR
618extern Lisp_Object code_convert_string P_ ((Lisp_Object,
619 struct coding_system *, int, int));
e5ff4bc2 620extern void setup_raw_text_coding_system P_ ((struct coding_system *));
4ed46869 621extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
64abe701 622extern Lisp_Object Qraw_text, Qemacs_mule;
4ed46869
KH
623extern Lisp_Object Qbuffer_file_coding_system;
624extern Lisp_Object Vcoding_category_list;
625
f967223b
KH
626extern Lisp_Object Qtranslation_table;
627extern Lisp_Object Qtranslation_table_id;
ab45712c 628
20ee919e
EZ
629/* Mnemonic strings to indicate each type of end-of-line. */
630extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
631/* Mnemonic string to indicate type of end-of-line is not yet decided. */
632extern Lisp_Object eol_mnemonic_undecided;
4ed46869 633
4ed46869
KH
634#ifdef emacs
635extern Lisp_Object Qfile_coding_system;
636extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
637extern Lisp_Object Qstart_process, Qopen_network_stream;
638
68c45bf0
PE
639extern char *emacs_strerror P_ ((int));
640
4ed46869
KH
641/* Coding-system for reading files and receiving data from process. */
642extern Lisp_Object Vcoding_system_for_read;
643/* Coding-system for writing files and sending data to process. */
644extern Lisp_Object Vcoding_system_for_write;
645/* Coding-system actually used in the latest I/O. */
646extern Lisp_Object Vlast_coding_system_used;
68c45bf0
PE
647/* Coding-system to use with system messages (e.g. strerror). */
648extern Lisp_Object Vlocale_coding_system;
4ed46869 649
77a9bc9a
EZ
650/* If non-zero, process buffer inherits the coding system used to decode
651 the subprocess output. */
652extern int inherit_process_coding_system;
653
4ed46869
KH
654/* Coding-system to be used for encoding terminal output. This
655 structure contains information of a coding-system specified by the
656 function `set-terminal-coding-system'. */
657extern struct coding_system terminal_coding;
658
fbaa2ed9
KH
659/* Coding system to be used to encode text for terminal display when
660 terminal coding system is nil. */
661extern struct coding_system safe_terminal_coding;
662
4ed46869
KH
663/* Coding-system of what is sent from terminal keyboard. This
664 structure contains information of a coding-system specified by the
665 function `set-keyboard-coding-system'. */
666extern struct coding_system keyboard_coding;
667
0827f88d
KH
668/* Default coding system to be used to write a file. */
669extern struct coding_system default_buffer_file_coding;
670
a5825a24
KH
671/* Default coding systems used for process I/O. */
672extern Lisp_Object Vdefault_process_coding_system;
4ed46869 673
658cc252
KH
674/* Function to call to force a user to force select a propert coding
675 system. */
676extern Lisp_Object Vselect_safe_coding_system_function;
677
31406df1
RS
678/* Coding system for file names, or nil if none. */
679extern Lisp_Object Vfile_name_coding_system;
680
681/* Coding system for file names used only when
682 Vfile_name_coding_system is nil. */
683extern Lisp_Object Vdefault_file_name_coding_system;
4ed46869
KH
684#endif
685
686#endif /* _CODING_H */