Commit | Line | Data |
---|---|---|
4ed46869 KH |
1 | /* Header for coding system handler. |
2 | Ver.1.0 | |
3 | ||
4 | Copyright (C) 1995 Free Software Foundation, Inc. | |
5 | Copyright (C) 1995 Electrotechnical Laboratory, JAPAN. | |
6 | ||
7 | This program is free software; you can redistribute it and/or modify | |
8 | it under the terms of the GNU General Public License as published by | |
9 | the Free Software Foundation; either version 2, or (at your option) | |
10 | any later version. | |
11 | ||
12 | This program is distributed in the hope that it will be useful, | |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | GNU General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU General Public License | |
18 | along with this program; if not, write to the Free Software | |
19 | Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ | |
20 | ||
21 | #ifndef _CODING_H | |
22 | #define _CODING_H | |
23 | ||
24 | #ifndef _CCL_H | |
25 | #include "../src/ccl.h" | |
26 | #endif | |
27 | ||
28 | /*** EMACS' INTERNAL FORMAT section ***/ | |
29 | ||
30 | /* All code (1-byte) of Emacs' internal format is classified into one | |
31 | of the followings. See also `charset.h'. */ | |
32 | enum emacs_code_class_type | |
33 | { | |
34 | EMACS_control_code, /* Control codes in the range | |
35 | 0x00..0x1F and 0x7F except for the | |
36 | following two codes. */ | |
37 | EMACS_linefeed_code, /* 0x0A (linefeed) to denote | |
38 | end-of-line. */ | |
39 | EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used | |
40 | in selective display mode. */ | |
41 | EMACS_ascii_code, /* ASCII characters. */ | |
42 | EMACS_leading_code_composition, /* Leading code of a composite | |
43 | character. */ | |
44 | EMACS_leading_code_2, /* Base leading code of official | |
45 | TYPE9N character. */ | |
46 | EMACS_leading_code_3, /* Base leading code of private TYPE9N | |
47 | or official TYPE9Nx9N character. */ | |
48 | EMACS_leading_code_4, /* Base leading code of private | |
49 | TYPE9Nx9N character. */ | |
50 | EMACS_invalid_code /* Invalid code, i.e. a base leading | |
51 | code not yet assigned to any | |
52 | charset, or a code of the range | |
53 | 0xA0..0xFF. */ | |
54 | }; | |
55 | ||
56 | extern enum emacs_code_class_type emacs_code_class[256]; | |
57 | ||
58 | /*** ISO2022 section ***/ | |
59 | ||
60 | /* Macros to define code of control characters for ISO2022's functions. */ | |
61 | /* code */ /* function */ | |
62 | #define ISO_CODE_LF 0x0A /* line-feed */ | |
63 | #define ISO_CODE_CR 0x0D /* carriage-return */ | |
64 | #define ISO_CODE_SO 0x0E /* shift-out */ | |
65 | #define ISO_CODE_SI 0x0F /* shift-in */ | |
66 | #define ISO_CODE_SS2_7 0x19 /* single-shift-2 for 7-bit code */ | |
67 | #define ISO_CODE_ESC 0x1B /* escape */ | |
68 | #define ISO_CODE_SS2 0x8E /* single-shift-2 */ | |
69 | #define ISO_CODE_SS3 0x8F /* single-shift-3 */ | |
70 | #define ISO_CODE_CSI 0x9B /* control-sequence-introduce */ | |
71 | ||
72 | /* All code (1-byte) of ISO2022 is classified into one of the | |
73 | followings. */ | |
74 | enum iso_code_class_type | |
75 | { | |
76 | ISO_control_code, /* Control codes in the range | |
77 | 0x00..0x1F, 0x7F, and 0x80..0x9F, | |
78 | except for the following seven | |
79 | codes. */ | |
80 | ISO_carriage_return, /* ISO_CODE_CR (0x0D) */ | |
81 | ISO_shift_out, /* ISO_CODE_SO (0x0E) */ | |
82 | ISO_shift_in, /* ISO_CODE_SI (0x0F) */ | |
83 | ISO_single_shift_2_7, /* ISO_CODE_SS2_7 (0x19) */ | |
84 | ISO_escape, /* ISO_CODE_SO (0x1B) */ | |
85 | ISO_single_shift_2, /* ISO_CODE_SS2 (0x8E) */ | |
86 | ISO_single_shift_3, /* ISO_CODE_SS3 (0x8F) */ | |
87 | ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */ | |
88 | ISO_0x20_or_0x7F, /* Codes of the values 0x20 or 0x7F. */ | |
89 | ISO_graphic_plane_0, /* Graphic codes in the range 0x21..0x7E. */ | |
90 | ISO_0xA0_or_0xFF, /* Codes of the values 0xA0 or 0xFF. */ | |
91 | ISO_graphic_plane_1 /* Graphic codes in the range 0xA1..0xFE. */ | |
92 | }; | |
93 | ||
94 | /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags' | |
95 | element in the structure `coding_system'. This information is used | |
96 | while encoding a text to ISO2022. **/ | |
97 | ||
98 | /* If set, produce short-form designation sequence (e.g. ESC $ A) | |
99 | instead of long-form sequence (e.g. ESC $ ( A). */ | |
100 | #define CODING_FLAG_ISO_SHORT_FORM 0x0001 | |
101 | ||
102 | /* If set, reset graphic planes and registers at end-of-line to the | |
103 | initial state. */ | |
104 | #define CODING_FLAG_ISO_RESET_AT_EOL 0x0002 | |
105 | ||
106 | /* If set, reset graphic planes and registers before any control | |
107 | characters to the initial state. */ | |
108 | #define CODING_FLAG_ISO_RESET_AT_CNTL 0x0004 | |
109 | ||
110 | /* If set, encode by 7-bit environment. */ | |
111 | #define CODING_FLAG_ISO_SEVEN_BITS 0x0008 | |
112 | ||
113 | /* If set, use locking-shift function. */ | |
114 | #define CODING_FLAG_ISO_LOCKING_SHIFT 0x0010 | |
115 | ||
116 | /* If set, use single-shift function. Overwrite | |
117 | CODING_FLAG_ISO_LOCKING_SHIFT. */ | |
118 | #define CODING_FLAG_ISO_SINGLE_SHIFT 0x0020 | |
119 | ||
120 | /* If set, designate JISX0201-Roman instead of ASCII. */ | |
121 | #define CODING_FLAG_ISO_USE_ROMAN 0x0040 | |
122 | ||
123 | /* If set, designate JISX0208-1978 instead of JISX0208-1983. */ | |
124 | #define CODING_FLAG_ISO_USE_OLDJIS 0x0080 | |
125 | ||
126 | /* If set, do not produce ISO6429's direction specifying sequence. */ | |
127 | #define CODING_FLAG_ISO_NO_DIRECTION 0x0100 | |
128 | ||
129 | /* Structure of the field `spec.iso2022' in the structure `coding_system'. */ | |
130 | struct iso2022_spec | |
131 | { | |
132 | /* The current graphic register invoked to each graphic plane. */ | |
133 | int current_invocation[2]; | |
134 | ||
135 | /* The current charset designated to each graphic register. */ | |
136 | int current_designation[4]; | |
137 | ||
138 | /* A charset initially designated to each graphic register. */ | |
139 | int initial_designation[4]; | |
140 | ||
141 | /* A graphic register to which each charset should be designated. */ | |
142 | int requested_designation[MAX_CHARSET]; | |
143 | ||
144 | /* Set to 1 temporarily only when graphic register 2 or 3 is invoked | |
145 | by single-shift while encoding. */ | |
146 | int single_shifting; | |
147 | }; | |
148 | ||
149 | /* Macros to access each field in the structure `spec.iso2022'. */ | |
150 | #define CODING_SPEC_ISO_INVOCATION(coding, plane) \ | |
151 | coding->spec.iso2022.current_invocation[plane] | |
152 | #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \ | |
153 | coding->spec.iso2022.current_designation[reg] | |
154 | #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \ | |
155 | coding->spec.iso2022.initial_designation[reg] | |
156 | #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \ | |
157 | coding->spec.iso2022.requested_designation[charset] | |
158 | ||
159 | /* Set to 1 temporarily only when encoding a character with | |
160 | single-shift function. */ | |
161 | #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \ | |
162 | coding->spec.iso2022.single_shifting | |
163 | ||
164 | /* Return a charset which is currently designated to the graphic plane | |
165 | PLANE in the coding-system CODING. */ | |
166 | #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane) \ | |
167 | CODING_SPEC_ISO_DESIGNATION \ | |
168 | (coding, CODING_SPEC_ISO_INVOCATION (coding, plane)) | |
169 | ||
170 | /*** BIG5 section ***/ | |
171 | ||
172 | /* Macros to denote each type of BIG5 coding system. */ | |
173 | #define CODING_FLAG_BIG5_HKU 0x00 /* BIG5-HKU is one of variants of | |
174 | BIG5 developed by Hong Kong | |
175 | University. */ | |
176 | #define CODING_FLAG_BIG5_ETEN 0x01 /* BIG5_ETen is one of variants | |
177 | of BIG5 developed by the | |
178 | company ETen in Taiwan. */ | |
179 | ||
180 | /*** GENERAL section ***/ | |
181 | ||
182 | /* Types of coding system. */ | |
183 | enum coding_type | |
184 | { | |
185 | coding_type_no_conversion, /* A coding system which requires no | |
186 | conversion for reading and writing | |
187 | including end-of-line format. */ | |
188 | coding_type_internal, /* A coding system used in Emacs' | |
189 | buffer and string. Requires no | |
190 | conversion for reading and writing | |
191 | except for end-of-line format. */ | |
192 | coding_type_automatic, /* A coding system which requires | |
193 | automatic detection of a real | |
194 | coding system. */ | |
195 | coding_type_sjis, /* SJIS coding system for Japanese. */ | |
196 | coding_type_iso2022, /* Any coding system of ISO2022 | |
197 | variants. */ | |
198 | coding_type_big5, /* BIG5 coding system for Chinese. */ | |
199 | coding_type_ccl /* The coding system of which decoder | |
200 | and encoder are written in CCL. */ | |
201 | }; | |
202 | ||
203 | /* Formats of end-of-line. */ | |
204 | #define CODING_EOL_LF 0 /* Line-feed only, same as Emacs' | |
205 | internal format. */ | |
206 | #define CODING_EOL_CRLF 1 /* Sequence of carriage-return and | |
207 | line-feed. */ | |
208 | #define CODING_EOL_CR 2 /* Carriage-return only. */ | |
209 | #define CODING_EOL_AUTOMATIC 3 /* This value is used to denote the | |
210 | eol-type is not yet decided. */ | |
211 | ||
212 | /* Character composition status while encoding/decoding. */ | |
213 | #define COMPOSING_NO 0 /* not composing */ | |
214 | #define COMPOSING_WITH_RULE_HEAD 1 /* 1st char of with-rule composing follow */ | |
215 | #define COMPOSING_NO_RULE_HEAD 2 /* 1st char of no-rule composing follow */ | |
216 | #define COMPOSING_WITH_RULE_TAIL 3 /* Nth char of with-rule composing follow */ | |
217 | #define COMPOSING_NO_RULE_TAIL 4 /* Nth char of no-rule composing follow */ | |
218 | #define COMPOSING_WITH_RULE_RULE 5 /* composition rule follow */ | |
219 | ||
220 | /* 1 iff composing. */ | |
221 | #define COMPOSING_P(composing) (composing) | |
222 | /* 1 iff 1st char of composing element follows. */ | |
223 | #define COMPOSING_HEAD_P(composing) \ | |
224 | ((composing) && (composing) <= COMPOSING_NO_RULE_HEAD) | |
225 | /* 1 iff composing with embeded composition rule. */ | |
226 | #define COMPOSING_WITH_RULE_P(composing) ((composing) & 1) | |
227 | ||
228 | struct coding_system | |
229 | { | |
230 | /* Type of the coding system. */ | |
231 | enum coding_type type; | |
232 | ||
233 | /* If the coding system requires specific code to be attached at the | |
234 | tail of converted text, this value should be set to `1'. */ | |
235 | int require_flushing; | |
236 | ||
237 | /* Flag bits of the coding system. The meaning of each bit depends | |
238 | on the type of the coding system. */ | |
239 | unsigned int flags; | |
240 | ||
241 | /* Type of end-of-line format (LF, CRLF, or CR) of the coding system. */ | |
242 | int eol_type; | |
243 | ||
244 | /* Non-zero means that the current source text is the last block of the | |
245 | whole text to be converted. */ | |
246 | int last_block; | |
247 | ||
248 | /* Non-zero means that characters are being composed currently while | |
249 | decoding or encoding. See macros COMPOSING_XXXX above for the | |
250 | meaing of each non-zero value. */ | |
251 | int composing; | |
252 | ||
253 | /* 0 (left-to-right) or 1 (right-to-left): the direction of the text | |
254 | being processed currently. */ | |
255 | int direction; | |
256 | ||
257 | /* Non-zero means that the current source text is in a buffer which | |
258 | enables selective display. */ | |
259 | int selective; | |
260 | ||
261 | /* Detailed information specific to each type of coding system. */ | |
262 | union spec | |
263 | { | |
264 | struct iso2022_spec iso2022; | |
265 | struct ccl_spec ccl; /* Defined in ccl.h. */ | |
266 | } spec; | |
267 | ||
268 | /* Backward pointer to the Lisp symbol of the coding system. */ | |
269 | Lisp_Object symbol; | |
270 | ||
271 | /* Lisp function (symbol) to be called after decoding to do | |
272 | additional conversion. */ | |
273 | Lisp_Object post_read_conversion; | |
274 | ||
275 | /* Lisp function (symbol) to be called before encoding to do | |
276 | additional conversion. */ | |
277 | Lisp_Object pre_write_conversion; | |
278 | ||
279 | /* Carryover yielded by decoding/encoding incomplete source. No | |
280 | coding-system yields more than 7-byte of carryover. This does | |
281 | not include a text which is not processed because of short of | |
282 | output buffer. */ | |
283 | char carryover[8]; | |
284 | ||
285 | /* Actual data length in the above array. */ | |
286 | int carryover_size; | |
287 | }; | |
288 | ||
289 | /* Return 1 if the coding-system CODING requires conversion of | |
290 | representation of a visible character (text). */ | |
291 | #define CODING_REQUIRE_TEXT_CONVERSION(coding) \ | |
292 | ((coding)->type != coding_type_no_conversion \ | |
293 | && (coding)->type != coding_type_internal) | |
294 | ||
295 | /* Return 1 if the coding-system CODING requires conversion of the | |
296 | format of end-of-line. */ | |
297 | #define CODING_REQUIRE_EOL_CONVERSION(coding) \ | |
298 | ((coding)->eol_type != CODING_EOL_AUTOMATIC \ | |
299 | && (coding)->eol_type != CODING_EOL_LF) | |
300 | ||
301 | /* Return 1 if the coding-system CODING requires some conversion. */ | |
302 | #define CODING_REQUIRE_CONVERSION(coding) \ | |
303 | (CODING_REQUIRE_TEXT_CONVERSION (coding) \ | |
304 | || CODING_REQUIRE_EOL_CONVERSION (coding)) | |
305 | ||
306 | /* Index for each coding category in `coding_category_table' */ | |
307 | #define CODING_CATEGORY_IDX_INTERNAL 0 | |
308 | #define CODING_CATEGORY_IDX_SJIS 1 | |
309 | #define CODING_CATEGORY_IDX_ISO_7 2 | |
310 | #define CODING_CATEGORY_IDX_ISO_8_1 3 | |
311 | #define CODING_CATEGORY_IDX_ISO_8_2 4 | |
312 | #define CODING_CATEGORY_IDX_ISO_ELSE 5 | |
313 | #define CODING_CATEGORY_IDX_BIG5 6 | |
314 | #define CODING_CATEGORY_IDX_BINARY 7 | |
315 | #define CODING_CATEGORY_IDX_MAX 8 | |
316 | ||
317 | /* Definitions of flag bits returned by the function | |
318 | detect_coding_mask (). */ | |
319 | #define CODING_CATEGORY_MASK_INTERNAL (1 << CODING_CATEGORY_IDX_INTERNAL) | |
320 | #define CODING_CATEGORY_MASK_SJIS (1 << CODING_CATEGORY_IDX_SJIS) | |
321 | #define CODING_CATEGORY_MASK_ISO_7 (1 << CODING_CATEGORY_IDX_ISO_7) | |
322 | #define CODING_CATEGORY_MASK_ISO_8_1 (1 << CODING_CATEGORY_IDX_ISO_8_1) | |
323 | #define CODING_CATEGORY_MASK_ISO_8_2 (1 << CODING_CATEGORY_IDX_ISO_8_2) | |
324 | #define CODING_CATEGORY_MASK_ISO_ELSE (1 << CODING_CATEGORY_IDX_ISO_ELSE) | |
325 | #define CODING_CATEGORY_MASK_BIG5 (1 << CODING_CATEGORY_IDX_BIG5) | |
326 | ||
327 | /* This value is returned if detect_coding_mask () find nothing other | |
328 | than ASCII characters. */ | |
329 | #define CODING_CATEGORY_MASK_ANY \ | |
330 | ( CODING_CATEGORY_MASK_INTERNAL \ | |
331 | | CODING_CATEGORY_MASK_SJIS \ | |
332 | | CODING_CATEGORY_MASK_ISO_7 \ | |
333 | | CODING_CATEGORY_MASK_ISO_8_1 \ | |
334 | | CODING_CATEGORY_MASK_ISO_8_2 \ | |
335 | | CODING_CATEGORY_MASK_ISO_ELSE \ | |
336 | | CODING_CATEGORY_MASK_BIG5) | |
337 | ||
338 | /* Macros to decode or encode a character of JISX0208 in SJIS. S1 and | |
339 | S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding | |
340 | system. C1 and C2 are the 1st and 2nd position codes of Emacs' | |
341 | internal format. */ | |
342 | ||
343 | #define DECODE_SJIS(s1, s2, c1, c2) \ | |
344 | do { \ | |
345 | if (s2 >= 0x9F) \ | |
346 | c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0), \ | |
347 | c2 = s2 - 0x7E; \ | |
348 | else \ | |
349 | c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1), \ | |
350 | c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F); \ | |
351 | } while (0) | |
352 | ||
353 | #define ENCODE_SJIS(c1, c2, s1, s2) \ | |
354 | do { \ | |
355 | if (c1 & 1) \ | |
356 | s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1), \ | |
357 | s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F); \ | |
358 | else \ | |
359 | s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0), \ | |
360 | s2 = c2 + 0x7E; \ | |
361 | } while (0) | |
362 | ||
363 | /* Extern declarations. */ | |
364 | extern int decode_coding (), encode_coding (); | |
365 | extern int decoding_buffer_size (), encoding_buffer_size (); | |
366 | extern int conversion_buffer_size; | |
367 | extern char *conversion_buffer, *get_conversion_buffer (); | |
368 | extern Lisp_Object Fcheck_coding_system (); | |
369 | extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index; | |
370 | extern Lisp_Object Qbuffer_file_coding_system; | |
371 | extern Lisp_Object Vcoding_category_list; | |
372 | ||
373 | /* Mnemonic character to indicate each type of end-of-line. */ | |
374 | extern int eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac; | |
375 | /* Mnemonic character to indicate type of end-of-line is not yet decided. */ | |
376 | extern int eol_mnemonic_undecided; | |
377 | ||
378 | /* Table of coding-systems currently assigned to each coding-category. */ | |
379 | extern Lisp_Object coding_category_table[CODING_CATEGORY_IDX_MAX]; | |
380 | /* Table of names of symbol for each coding-category. */ | |
381 | extern char *coding_category_name[CODING_CATEGORY_IDX_MAX]; | |
382 | ||
383 | #ifdef emacs | |
384 | extern Lisp_Object Qfile_coding_system; | |
385 | extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument; | |
386 | extern Lisp_Object Qstart_process, Qopen_network_stream; | |
387 | ||
388 | /* Coding-system for reading files and receiving data from process. */ | |
389 | extern Lisp_Object Vcoding_system_for_read; | |
390 | /* Coding-system for writing files and sending data to process. */ | |
391 | extern Lisp_Object Vcoding_system_for_write; | |
392 | /* Coding-system actually used in the latest I/O. */ | |
393 | extern Lisp_Object Vlast_coding_system_used; | |
394 | ||
395 | /* Coding-system to be used for encoding terminal output. This | |
396 | structure contains information of a coding-system specified by the | |
397 | function `set-terminal-coding-system'. */ | |
398 | extern struct coding_system terminal_coding; | |
399 | ||
400 | /* Coding-system of what is sent from terminal keyboard. This | |
401 | structure contains information of a coding-system specified by the | |
402 | function `set-keyboard-coding-system'. */ | |
403 | extern struct coding_system keyboard_coding; | |
404 | ||
405 | extern Lisp_Object Vcoding_system_alist; | |
406 | ||
407 | #endif | |
408 | ||
409 | #endif /* _CODING_H */ |