|
70
|
1 /* Header for code conversion stuff
|
|
|
2 Copyright (C) 1991, 1995 Free Software Foundation, Inc.
|
|
|
3 Copyright (C) 1995 Sun Microsystems, Inc.
|
|
|
4
|
|
|
5 This file is part of XEmacs.
|
|
|
6
|
|
|
7 XEmacs is free software; you can redistribute it and/or modify it
|
|
|
8 under the terms of the GNU General Public License as published by the
|
|
|
9 Free Software Foundation; either version 2, or (at your option) any
|
|
|
10 later version.
|
|
|
11
|
|
|
12 XEmacs is distributed in the hope that it will be useful, but WITHOUT
|
|
|
13 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
|
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
|
|
15 for more details.
|
|
|
16
|
|
|
17 You should have received a copy of the GNU General Public License
|
|
|
18 along with XEmacs; see the file COPYING. If not, write to
|
|
|
19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
|
20 Boston, MA 02111-1307, USA. */
|
|
|
21
|
|
|
22 /* Synched up with: Mule 2.3. Not in FSF. */
|
|
|
23
|
|
|
24 /* 91.10.09 written by K.Handa <handa@etl.go.jp> */
|
|
|
25 /* Rewritten by Ben Wing <wing@666.com>. */
|
|
|
26
|
|
|
27 #ifndef _XEMACS_MULE_CODING_H_
|
|
|
28 #define _XEMACS_MULE_CODING_H_
|
|
|
29
|
|
|
30 struct decoding_stream;
|
|
|
31 struct encoding_stream;
|
|
|
32
|
|
|
33 /* Coding system types. These go into the TYPE field of a
|
|
|
34 struct Lisp_Coding_System. */
|
|
|
35
|
|
|
36 enum coding_system_type
|
|
|
37 {
|
|
|
38 CODESYS_AUTODETECT, /* Automatic conversion. */
|
|
|
39 CODESYS_SHIFT_JIS, /* Shift-JIS; Hankaku (half-width) KANA
|
|
|
40 is also supported. */
|
|
|
41 CODESYS_ISO2022, /* Any ISO2022-compliant coding system.
|
|
|
42 Includes JIS, EUC, CTEXT */
|
|
|
43 CODESYS_BIG5, /* BIG5 (used for Taiwanese). */
|
|
|
44 CODESYS_CCL, /* Converter written in CCL. */
|
|
|
45 CODESYS_NO_CONVERSION /* "No conversion"; used for binary files.
|
|
|
46 We use quotes because there really
|
|
|
47 is some conversion being applied,
|
|
|
48 but it appears to the user as if
|
|
|
49 the text is read in without conversion. */
|
|
|
50 #ifdef DEBUG_XEMACS
|
|
|
51 ,CODESYS_INTERNAL /* Raw (internally-formatted) data. */
|
|
|
52 #endif
|
|
|
53 };
|
|
|
54
|
|
|
55 enum eol_type
|
|
|
56 {
|
|
|
57 EOL_AUTODETECT,
|
|
|
58 EOL_LF,
|
|
|
59 EOL_CRLF,
|
|
|
60 EOL_CR
|
|
|
61 };
|
|
|
62
|
|
|
63 /* This holds the current state of a running CCL program. */
|
|
|
64 struct ccl_program
|
|
|
65 {
|
|
|
66 Lisp_Object saved_vector;
|
|
|
67 Lisp_Object *prog; /* compiled code */
|
|
|
68 int size; /* size of compiled code */
|
|
|
69 int ic; /* instruction counter */
|
|
|
70 int reg[8]; /* reg[7] is used for `condition' */
|
|
|
71 int end_flag; /* set when processing the last block */
|
|
|
72 int status;
|
|
|
73 };
|
|
|
74
|
|
185
|
75 typedef struct charset_conversion_spec charset_conversion_spec;
|
|
70
|
76 struct charset_conversion_spec
|
|
|
77 {
|
|
|
78 Lisp_Object from_charset;
|
|
|
79 Lisp_Object to_charset;
|
|
|
80 };
|
|
|
81
|
|
185
|
82 typedef struct
|
|
70
|
83 {
|
|
185
|
84 Dynarr_declare (charset_conversion_spec);
|
|
70
|
85 } charset_conversion_spec_dynarr;
|
|
|
86
|
|
|
87 struct Lisp_Coding_System
|
|
|
88 {
|
|
|
89 struct lcrecord_header header;
|
|
|
90
|
|
|
91 /* Name and doc string of this coding system. */
|
|
|
92 Lisp_Object name, doc_string;
|
|
|
93
|
|
|
94 /* This is the major type of the coding system -- one of Big5, ISO2022,
|
|
|
95 Shift-JIS, etc. See the constants above. */
|
|
|
96 enum coding_system_type type;
|
|
|
97
|
|
|
98 /* Mnemonic string displayed in the modeline when this coding
|
|
|
99 system is active for a particular buffer. */
|
|
|
100 Lisp_Object mnemonic;
|
|
|
101
|
|
|
102 Lisp_Object post_read_conversion, pre_write_conversion;
|
|
|
103
|
|
|
104 enum eol_type eol_type;
|
|
|
105
|
|
|
106 /* Subsidiary coding systems that specify a particular type of EOL
|
|
|
107 marking, rather than autodetecting it. These will only be non-nil
|
|
|
108 if (eol_type == EOL_AUTODETECT). */
|
|
|
109 Lisp_Object eol_lf, eol_crlf, eol_cr;
|
|
|
110
|
|
|
111 struct
|
|
187
|
112 {
|
|
|
113 /* What are the charsets to be initially designated to G0, G1,
|
|
|
114 G2, G3? If t, no charset is initially designated. If nil,
|
|
|
115 no charset is initially designated and no charset is allowed
|
|
|
116 to be designated. */
|
|
|
117 Lisp_Object initial_charset[4];
|
|
70
|
118
|
|
187
|
119 /* If true, a designation escape sequence needs to be sent on output
|
|
|
120 for the charset in G[0-3] before that charset is used. */
|
|
|
121 unsigned char force_charset_on_output[4];
|
|
70
|
122
|
|
187
|
123 charset_conversion_spec_dynarr *input_conv;
|
|
|
124 charset_conversion_spec_dynarr *output_conv;
|
|
70
|
125
|
|
187
|
126 unsigned int shoort :1; /* C makes you speak Dutch */
|
|
|
127 unsigned int no_ascii_eol :1;
|
|
|
128 unsigned int no_ascii_cntl :1;
|
|
|
129 unsigned int seven :1;
|
|
|
130 unsigned int lock_shift :1;
|
|
|
131 unsigned int no_iso6429 :1;
|
|
|
132 unsigned int escape_quoted :1;
|
|
|
133 } iso2022;
|
|
70
|
134
|
|
|
135 struct
|
|
187
|
136 {
|
|
|
137 /* For a CCL coding system, these specify the CCL programs used for
|
|
|
138 decoding (input) and encoding (output). */
|
|
|
139 Lisp_Object decode, encode;
|
|
|
140 } ccl;
|
|
70
|
141 };
|
|
|
142
|
|
|
143 DECLARE_LRECORD (coding_system, struct Lisp_Coding_System);
|
|
|
144 #define XCODING_SYSTEM(x) XRECORD (x, coding_system, struct Lisp_Coding_System)
|
|
|
145 #define XSETCODING_SYSTEM(x, p) XSETRECORD (x, p, coding_system)
|
|
|
146 #define CODING_SYSTEMP(x) RECORDP (x, coding_system)
|
|
|
147 #define GC_CODING_SYSTEMP(x) GC_RECORDP (x, coding_system)
|
|
|
148 #define CHECK_CODING_SYSTEM(x) CHECK_RECORD (x, coding_system)
|
|
|
149 #define CONCHECK_CODING_SYSTEM(x) CONCHECK_RECORD (x, coding_system)
|
|
|
150
|
|
|
151 #define CODING_SYSTEM_NAME(codesys) ((codesys)->name)
|
|
|
152 #define CODING_SYSTEM_DOC_STRING(codesys) ((codesys)->doc_string)
|
|
|
153 #define CODING_SYSTEM_TYPE(codesys) ((codesys)->type)
|
|
|
154 #define CODING_SYSTEM_MNEMONIC(codesys) ((codesys)->mnemonic)
|
|
|
155 #define CODING_SYSTEM_POST_READ_CONVERSION(codesys) \
|
|
|
156 ((codesys)->post_read_conversion)
|
|
|
157 #define CODING_SYSTEM_PRE_WRITE_CONVERSION(codesys) \
|
|
|
158 ((codesys)->pre_write_conversion)
|
|
|
159 #define CODING_SYSTEM_EOL_TYPE(codesys) ((codesys)->eol_type)
|
|
|
160 #define CODING_SYSTEM_EOL_LF(codesys) ((codesys)->eol_lf)
|
|
|
161 #define CODING_SYSTEM_EOL_CRLF(codesys) ((codesys)->eol_crlf)
|
|
|
162 #define CODING_SYSTEM_EOL_CR(codesys) ((codesys)->eol_cr)
|
|
|
163 #define CODING_SYSTEM_ISO2022_INITIAL_CHARSET(codesys, g) \
|
|
|
164 ((codesys)->iso2022.initial_charset[g])
|
|
|
165 #define CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT(codesys, g) \
|
|
|
166 ((codesys)->iso2022.force_charset_on_output[g])
|
|
|
167 #define CODING_SYSTEM_ISO2022_SHORT(codesys) ((codesys)->iso2022.shoort)
|
|
|
168 #define CODING_SYSTEM_ISO2022_NO_ASCII_EOL(codesys) \
|
|
|
169 ((codesys)->iso2022.no_ascii_eol)
|
|
|
170 #define CODING_SYSTEM_ISO2022_NO_ASCII_CNTL(codesys) \
|
|
|
171 ((codesys)->iso2022.no_ascii_cntl)
|
|
|
172 #define CODING_SYSTEM_ISO2022_SEVEN(codesys) ((codesys)->iso2022.seven)
|
|
|
173 #define CODING_SYSTEM_ISO2022_LOCK_SHIFT(codesys) \
|
|
|
174 ((codesys)->iso2022.lock_shift)
|
|
|
175 #define CODING_SYSTEM_ISO2022_NO_ISO6429(codesys) \
|
|
|
176 ((codesys)->iso2022.no_iso6429)
|
|
|
177 #define CODING_SYSTEM_ISO2022_ESCAPE_QUOTED(codesys) \
|
|
|
178 ((codesys)->iso2022.escape_quoted)
|
|
|
179 #define CODING_SYSTEM_CCL_DECODE(codesys) ((codesys)->ccl.decode)
|
|
|
180 #define CODING_SYSTEM_CCL_ENCODE(codesys) ((codesys)->ccl.encode)
|
|
|
181
|
|
|
182 #define XCODING_SYSTEM_NAME(codesys) \
|
|
|
183 CODING_SYSTEM_NAME (XCODING_SYSTEM (codesys))
|
|
|
184 #define XCODING_SYSTEM_DOC_STRING(codesys) \
|
|
|
185 CODING_SYSTEM_DOC_STRING (XCODING_SYSTEM (codesys))
|
|
|
186 #define XCODING_SYSTEM_TYPE(codesys) \
|
|
|
187 CODING_SYSTEM_TYPE (XCODING_SYSTEM (codesys))
|
|
|
188 #define XCODING_SYSTEM_MNEMONIC(codesys) \
|
|
|
189 CODING_SYSTEM_MNEMONIC (XCODING_SYSTEM (codesys))
|
|
|
190 #define XCODING_SYSTEM_POST_READ_CONVERSION(codesys) \
|
|
|
191 CODING_SYSTEM_POST_READ_CONVERSION (XCODING_SYSTEM (codesys))
|
|
|
192 #define XCODING_SYSTEM_PRE_WRITE_CONVERSION(codesys) \
|
|
|
193 CODING_SYSTEM_PRE_WRITE_CONVERSION (XCODING_SYSTEM (codesys))
|
|
|
194 #define XCODING_SYSTEM_EOL_TYPE(codesys) \
|
|
|
195 CODING_SYSTEM_EOL_TYPE (XCODING_SYSTEM (codesys))
|
|
|
196 #define XCODING_SYSTEM_EOL_LF(codesys) \
|
|
|
197 CODING_SYSTEM_EOL_LF (XCODING_SYSTEM (codesys))
|
|
|
198 #define XCODING_SYSTEM_EOL_CRLF(codesys) \
|
|
|
199 CODING_SYSTEM_EOL_CRLF (XCODING_SYSTEM (codesys))
|
|
|
200 #define XCODING_SYSTEM_EOL_CR(codesys) \
|
|
|
201 CODING_SYSTEM_EOL_CR (XCODING_SYSTEM (codesys))
|
|
|
202 #define XCODING_SYSTEM_ISO2022_INITIAL_CHARSET(codesys, g) \
|
|
|
203 CODING_SYSTEM_ISO2022_INITIAL_CHARSET (XCODING_SYSTEM (codesys), g)
|
|
|
204 #define XCODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT(codesys, g) \
|
|
|
205 CODING_SYSTEM_ISO2022_FORCE_CHARSET_ON_OUTPUT (XCODING_SYSTEM (codesys), g)
|
|
|
206 #define XCODING_SYSTEM_ISO2022_SHORT(codesys) \
|
|
|
207 CODING_SYSTEM_ISO2022_SHORT (XCODING_SYSTEM (codesys))
|
|
|
208 #define XCODING_SYSTEM_ISO2022_NO_ASCII_EOL(codesys) \
|
|
|
209 CODING_SYSTEM_ISO2022_NO_ASCII_EOL (XCODING_SYSTEM (codesys))
|
|
|
210 #define XCODING_SYSTEM_ISO2022_NO_ASCII_CNTL(codesys) \
|
|
|
211 CODING_SYSTEM_ISO2022_NO_ASCII_CNTL (XCODING_SYSTEM (codesys))
|
|
|
212 #define XCODING_SYSTEM_ISO2022_SEVEN(codesys) \
|
|
|
213 CODING_SYSTEM_ISO2022_SEVEN (XCODING_SYSTEM (codesys))
|
|
|
214 #define XCODING_SYSTEM_ISO2022_LOCK_SHIFT(codesys) \
|
|
|
215 CODING_SYSTEM_ISO2022_LOCK_SHIFT (XCODING_SYSTEM (codesys))
|
|
|
216 #define XCODING_SYSTEM_ISO2022_NO_ISO6429(codesys) \
|
|
|
217 CODING_SYSTEM_ISO2022_NO_ISO6429 (XCODING_SYSTEM (codesys))
|
|
|
218 #define XCODING_SYSTEM_ISO2022_ESCAPE_QUOTED(codesys) \
|
|
|
219 CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (XCODING_SYSTEM (codesys))
|
|
|
220 #define XCODING_SYSTEM_CCL_DECODE(codesys) \
|
|
|
221 CODING_SYSTEM_CCL_DECODE (XCODING_SYSTEM (codesys))
|
|
|
222 #define XCODING_SYSTEM_CCL_ENCODE(codesys) \
|
|
|
223 CODING_SYSTEM_CCL_ENCODE (XCODING_SYSTEM (codesys))
|
|
|
224
|
|
110
|
225 extern Lisp_Object Qbuffer_file_coding_system, Qcoding_system_error;
|
|
70
|
226
|
|
|
227 extern Lisp_Object Vkeyboard_coding_system;
|
|
|
228 extern Lisp_Object Vterminal_coding_system;
|
|
120
|
229 extern Lisp_Object Vcoding_system_for_read;
|
|
|
230 extern Lisp_Object Vcoding_system_for_write;
|
|
70
|
231 extern Lisp_Object Vpathname_coding_system;
|
|
|
232
|
|
|
233 extern Lisp_Object Qescape_quoted;
|
|
|
234
|
|
|
235 /* Flags indicating current state while converting code. */
|
|
|
236
|
|
|
237 /* Used by everyone. */
|
|
|
238
|
|
|
239 #define CODING_STATE_END (1 << 0) /* If set, this is the last chunk of
|
|
|
240 data being processed. When this
|
|
|
241 is finished, output any necessary
|
|
|
242 terminating control characters,
|
|
|
243 escape sequences, etc. */
|
|
|
244 #define CODING_STATE_CR (1 << 1) /* If set, we just saw a CR. */
|
|
|
245
|
|
|
246
|
|
|
247 /* Used by Big 5 on output. */
|
|
|
248
|
|
|
249 #define CODING_STATE_BIG5_1 (1 << 2) /* If set, we just encountered
|
|
|
250 LEADING_BYTE_BIG5_1. */
|
|
|
251 #define CODING_STATE_BIG5_2 (1 << 3) /* If set, we just encountered
|
|
|
252 LEADING_BYTE_BIG5_2. */
|
|
|
253
|
|
|
254
|
|
|
255 /* Used by ISO2022 on input and output. */
|
|
|
256
|
|
|
257 #define CODING_STATE_R2L (1 << 4) /* If set, the current
|
|
|
258 directionality is right-to-left.
|
|
|
259 Otherwise, it's left-to-right. */
|
|
|
260
|
|
|
261
|
|
|
262 /* Used by ISO2022 on input. */
|
|
|
263
|
|
|
264 #define CODING_STATE_ESCAPE (1 << 5) /* If set, we're currently parsing
|
|
|
265 an escape sequence and the upper
|
|
|
266 16 bits should be looked at to
|
|
|
267 indicate what partial escape
|
|
|
268 sequence we've seen so far.
|
|
|
269 Otherwise, we're running
|
|
|
270 through actual text. */
|
|
|
271 #define CODING_STATE_SS2 (1 << 6) /* If set, G2 is invoked into GL, but
|
|
|
272 only for the next character. */
|
|
|
273 #define CODING_STATE_SS3 (1 << 7) /* If set, G3 is invoked into GL,
|
|
|
274 but only for the next character.
|
|
|
275 If both CODING_STATE_SS2 and
|
|
|
276 CODING_STATE_SS3 are set,
|
|
|
277 CODING_STATE_SS2 overrides; but
|
|
|
278 this probably indicates an error
|
|
|
279 in the text encoding. */
|
|
|
280 #define CODING_STATE_COMPOSITE (1 << 8) /* If set, we're currently processing
|
|
|
281 a composite character (i.e. a
|
|
|
282 character constructed by
|
|
|
283 overstriking two or more
|
|
|
284 characters). */
|
|
|
285
|
|
|
286
|
|
|
287 /* CODING_STATE_ISO2022_LOCK is the mask of flags that remain on until
|
|
|
288 explicitly turned off when in the ISO2022 encoder/decoder. Other flags are
|
|
|
289 turned off at the end of processing each character or escape sequence. */
|
|
|
290 # define CODING_STATE_ISO2022_LOCK \
|
|
|
291 (CODING_STATE_END | CODING_STATE_COMPOSITE | CODING_STATE_R2L)
|
|
|
292 #define CODING_STATE_BIG5_LOCK \
|
|
|
293 CODING_STATE_END
|
|
|
294
|
|
|
295 /* Flags indicating what we've seen so far when parsing an
|
|
|
296 ISO2022 escape sequence. */
|
|
|
297 enum iso_esc_flag
|
|
|
298 {
|
|
|
299 /* Partial sequences */
|
|
|
300 ISO_ESC_NOTHING, /* Nothing has been seen. */
|
|
|
301 ISO_ESC, /* We've seen ESC. */
|
|
|
302 ISO_ESC_2_4, /* We've seen ESC $. This indicates
|
|
|
303 that we're designating a multi-byte, rather
|
|
|
304 than a single-byte, character set. */
|
|
|
305 ISO_ESC_2_8, /* We've seen ESC 0x28, i.e. ESC (.
|
|
|
306 This means designate a 94-character
|
|
|
307 character set into G0. */
|
|
|
308 ISO_ESC_2_9, /* We've seen ESC 0x29 -- designate a
|
|
|
309 94-character character set into G1. */
|
|
|
310 ISO_ESC_2_10, /* We've seen ESC 0x2A. */
|
|
|
311 ISO_ESC_2_11, /* We've seen ESC 0x2B. */
|
|
|
312 ISO_ESC_2_12, /* We've seen ESC 0x2C -- designate a
|
|
|
313 96-character character set into G0.
|
|
|
314 (This is not ISO2022-standard.
|
|
|
315 The following 96-character
|
|
|
316 control sequences are standard,
|
|
|
317 though.) */
|
|
|
318 ISO_ESC_2_13, /* We've seen ESC 0x2D -- designate a
|
|
|
319 96-character character set into G1.
|
|
|
320 */
|
|
|
321 ISO_ESC_2_14, /* We've seen ESC 0x2E. */
|
|
|
322 ISO_ESC_2_15, /* We've seen ESC 0x2F. */
|
|
|
323 ISO_ESC_2_4_8, /* We've seen ESC $ 0x28 -- designate
|
|
|
324 a 94^N character set into G0. */
|
|
|
325 ISO_ESC_2_4_9, /* We've seen ESC $ 0x29. */
|
|
|
326 ISO_ESC_2_4_10, /* We've seen ESC $ 0x2A. */
|
|
|
327 ISO_ESC_2_4_11, /* We've seen ESC $ 0x2B. */
|
|
|
328 ISO_ESC_2_4_12, /* We've seen ESC $ 0x2C. */
|
|
|
329 ISO_ESC_2_4_13, /* We've seen ESC $ 0x2D. */
|
|
|
330 ISO_ESC_2_4_14, /* We've seen ESC $ 0x2E. */
|
|
|
331 ISO_ESC_2_4_15, /* We've seen ESC $ 0x2F. */
|
|
|
332 ISO_ESC_5_11, /* We've seen ESC [ or 0x9B. This
|
|
|
333 starts a directionality-control
|
|
|
334 sequence. The next character
|
|
|
335 must be 0, 1, 2, or ]. */
|
|
|
336 ISO_ESC_5_11_0, /* We've seen 0x9B 0. The next
|
|
|
337 character must be ]. */
|
|
|
338 ISO_ESC_5_11_1, /* We've seen 0x9B 1. The next
|
|
|
339 character must be ]. */
|
|
|
340 ISO_ESC_5_11_2, /* We've seen 0x9B 2. The next
|
|
|
341 character must be ]. */
|
|
|
342
|
|
|
343 /* Full sequences. */
|
|
|
344 ISO_ESC_START_COMPOSITE, /* Private usage for START COMPOSING */
|
|
|
345 ISO_ESC_END_COMPOSITE, /* Private usage for END COMPOSING */
|
|
|
346 ISO_ESC_SINGLE_SHIFT, /* We've seen a complete single-shift sequence. */
|
|
|
347 ISO_ESC_LOCKING_SHIFT,/* We've seen a complete locking-shift sequence. */
|
|
|
348 ISO_ESC_DESIGNATE, /* We've seen a complete designation sequence. */
|
|
|
349 ISO_ESC_DIRECTIONALITY,/* We've seen a complete ISO6429 directionality
|
|
|
350 sequence. */
|
|
|
351 ISO_ESC_LITERAL /* We've seen a literal character ala
|
|
|
352 escape-quoting. */
|
|
|
353 };
|
|
|
354
|
|
|
355 /* Macros to define code of control characters for ISO2022's functions. */
|
|
|
356 /* code */ /* function */
|
|
|
357 #define ISO_CODE_LF 0x0A /* line-feed */
|
|
|
358 #define ISO_CODE_CR 0x0D /* carriage-return */
|
|
|
359 #define ISO_CODE_SO 0x0E /* shift-out */
|
|
|
360 #define ISO_CODE_SI 0x0F /* shift-in */
|
|
|
361 #define ISO_CODE_ESC 0x1B /* escape */
|
|
|
362 #define ISO_CODE_DEL 0x7F /* delete */
|
|
|
363 #define ISO_CODE_SS2 0x8E /* single-shift-2 */
|
|
|
364 #define ISO_CODE_SS3 0x8F /* single-shift-3 */
|
|
|
365 #define ISO_CODE_CSI 0x9B /* control-sequence-introduce */
|
|
|
366
|
|
|
367 /* Macros to access an encoding stream or decoding stream */
|
|
|
368
|
|
|
369 #define CODING_STREAM_DECOMPOSE(str, flags, ch) \
|
|
|
370 do { \
|
|
|
371 flags = (str)->flags; \
|
|
|
372 ch = (str)->ch; \
|
|
|
373 } while (0)
|
|
|
374
|
|
|
375 #define CODING_STREAM_COMPOSE(str, flags, ch) \
|
|
|
376 do { \
|
|
|
377 (str)->flags = flags; \
|
|
|
378 (str)->ch = ch; \
|
|
|
379 } while (0)
|
|
|
380
|
|
|
381
|
|
|
382 /* For detecting the encoding of text */
|
|
|
383 enum coding_category_type
|
|
|
384 {
|
|
|
385 CODING_CATEGORY_SHIFT_JIS,
|
|
|
386 CODING_CATEGORY_ISO_7, /* ISO2022 system using only seven-bit bytes,
|
|
|
387 no locking shift */
|
|
|
388 CODING_CATEGORY_ISO_8_DESIGNATE, /* ISO2022 system using eight-bit bytes,
|
|
|
389 no locking shift, no single shift,
|
|
|
390 using designation to switch charsets */
|
|
|
391 CODING_CATEGORY_ISO_8_1, /* ISO2022 system using eight-bit bytes,
|
|
|
392 no locking shift, no designation sequences,
|
|
|
393 one-dimension characters in the upper half. */
|
|
|
394 CODING_CATEGORY_ISO_8_2, /* ISO2022 system using eight-bit bytes,
|
|
|
395 no locking shift, no designation sequences,
|
|
|
396 two-dimension characters in the upper half. */
|
|
|
397 CODING_CATEGORY_ISO_LOCK_SHIFT, /* ISO2022 system using locking shift */
|
|
|
398 CODING_CATEGORY_BIG5,
|
|
|
399 CODING_CATEGORY_NO_CONVERSION
|
|
|
400 };
|
|
|
401
|
|
|
402 #define CODING_CATEGORY_LAST CODING_CATEGORY_NO_CONVERSION
|
|
|
403
|
|
|
404 #define CODING_CATEGORY_SHIFT_JIS_MASK \
|
|
|
405 (1 << CODING_CATEGORY_SHIFT_JIS)
|
|
|
406 #define CODING_CATEGORY_ISO_7_MASK \
|
|
|
407 (1 << CODING_CATEGORY_ISO_7)
|
|
|
408 #define CODING_CATEGORY_ISO_8_DESIGNATE_MASK \
|
|
|
409 (1 << CODING_CATEGORY_ISO_8_DESIGNATE)
|
|
|
410 #define CODING_CATEGORY_ISO_8_1_MASK \
|
|
|
411 (1 << CODING_CATEGORY_ISO_8_1)
|
|
|
412 #define CODING_CATEGORY_ISO_8_2_MASK \
|
|
|
413 (1 << CODING_CATEGORY_ISO_8_2)
|
|
|
414 #define CODING_CATEGORY_ISO_LOCK_SHIFT_MASK \
|
|
|
415 (1 << CODING_CATEGORY_ISO_LOCK_SHIFT)
|
|
|
416 #define CODING_CATEGORY_BIG5_MASK \
|
|
|
417 (1 << CODING_CATEGORY_BIG5)
|
|
|
418 #define CODING_CATEGORY_NO_CONVERSION_MASK \
|
|
|
419 (1 << CODING_CATEGORY_NO_CONVERSION)
|
|
|
420 #define CODING_CATEGORY_NOT_FINISHED_MASK \
|
|
|
421 (1 << 30)
|
|
|
422
|
|
|
423 extern Lisp_Object make_decoding_input_stream (Lstream *stream,
|
|
|
424 Lisp_Object codesys);
|
|
|
425 extern Lisp_Object make_encoding_input_stream (Lstream *stream,
|
|
|
426 Lisp_Object codesys);
|
|
|
427 extern Lisp_Object make_decoding_output_stream (Lstream *stream,
|
|
|
428 Lisp_Object codesys);
|
|
|
429 extern Lisp_Object make_encoding_output_stream (Lstream *stream,
|
|
|
430 Lisp_Object codesys);
|
|
|
431 extern Lisp_Object decoding_stream_coding_system (Lstream *stream);
|
|
|
432 extern Lisp_Object encoding_stream_coding_system (Lstream *stream);
|
|
|
433 extern void set_decoding_stream_coding_system (Lstream *stream,
|
|
|
434 Lisp_Object codesys);
|
|
|
435 extern void set_encoding_stream_coding_system (Lstream *stream,
|
|
|
436 Lisp_Object codesys);
|
|
|
437
|
|
|
438 /* In mule-ccl.c */
|
|
|
439 int ccl_driver (struct ccl_program *ccl, CONST unsigned char *src,
|
|
|
440 unsigned_char_dynarr *dst, int n, int end_flag);
|
|
|
441 void set_ccl_program (struct ccl_program *ccl, Lisp_Object val, int *regs,
|
|
|
442 int numregs, int initial_ic);
|
|
|
443
|
|
|
444 #endif /* _XEMACS_MULE_CODING_H_ */
|
