?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Various cleanups.
[emacs.git] / src / charset.h
blob0d61aa054c0ba923461331f7a701070aaf21811b
1 /* Header for multibyte character handler.
2 Copyright (C) 2001, 2002, 2003, 2004, 2005,
3 2006 Free Software Foundation, Inc.
4 Copyright (C) 1995, 1997, 1998, 2003
5 National Institute of Advanced Industrial Science and Technology (AIST)
6 Registration Number H14PRO021
7
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.
14
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.
19
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., 51 Franklin Street, Fifth Floor,
23 Boston, MA 02110-1301, USA. */
24
25 #ifndef EMACS_CHARSET_H
26 #define EMACS_CHARSET_H
27
28 /* #define BYTE_COMBINING_DEBUG */
29
30 /*** GENERAL NOTE on CHARACTER SET (CHARSET) ***
31
32 A character set ("charset" hereafter) is a meaningful collection
33 (i.e. language, culture, functionality, etc) of characters. Emacs
34 handles multiple charsets at once. Each charset corresponds to one
35 of the ISO charsets. Emacs identifies a charset by a unique
36 identification number, whereas ISO identifies a charset by a triplet
37 of DIMENSION, CHARS and FINAL-CHAR. So, hereafter, just saying
38 "charset" means an identification number (integer value).
39
40 The value range of charsets is 0x00, 0x81..0xFE. There are four
41 kinds of charset depending on DIMENSION (1 or 2) and CHARS (94 or
42 96). For instance, a charset of DIMENSION2_CHARS94 contains 94x94
43 characters.
44
45 Within Emacs Lisp, a charset is treated as a symbol which has a
46 property `charset'. The property value is a vector containing
47 various information about the charset. For readability of C code,
48 we use the following convention for C variable names:
49 charset_symbol: Emacs Lisp symbol of a charset
50 charset_id: Emacs Lisp integer of an identification number of a charset
51 charset: C integer of an identification number of a charset
52
53 Each charset (except for ascii) is assigned a base leading-code
54 (range 0x80..0x9E). In addition, a charset of greater than 0xA0
55 (whose base leading-code is 0x9A..0x9D) is assigned an extended
56 leading-code (range 0xA0..0xFE). In this case, each base
57 leading-code specifies the allowable range of extended leading-code
58 as shown in the table below. A leading-code is used to represent a
59 character in Emacs' buffer and string.
60
61 We call a charset which has extended leading-code a "private
62 charset" because those are mainly for a charset which is not yet
63 registered by ISO. On the contrary, we call a charset which does
64 not have extended leading-code an "official charset".
65
66 ---------------------------------------------------------------------------
67 charset dimension base leading-code extended leading-code
68 ---------------------------------------------------------------------------
69 0x00 official dim1 -- none -- -- none --
70 (ASCII)
71 0x01..0x7F --never used--
72 0x80 official dim1 -- none -- -- none --
73 (eight-bit-graphic)
74 0x81..0x8F official dim1 same as charset -- none --
75 0x90..0x99 official dim2 same as charset -- none --
76 0x9A..0x9D --never used--
77 0x9E official dim1 same as charset -- none --
78 (eight-bit-control)
79 0x9F --never used--
80 0xA0..0xDF private dim1 0x9A same as charset
81 of 1-column width
82 0xE0..0xEF private dim1 0x9B same as charset
83 of 2-column width
84 0xF0..0xF4 private dim2 0x9C same as charset
85 of 1-column width
86 0xF5..0xFE private dim2 0x9D same as charset
87 of 2-column width
88 0xFF --never used--
89 ---------------------------------------------------------------------------
90
91 */
92
93 /* Definition of special leading-codes. */
94 /* Leading-code followed by extended leading-code. */
95 #define LEADING_CODE_PRIVATE_11 0x9A /* for private DIMENSION1 of 1-column */
96 #define LEADING_CODE_PRIVATE_12 0x9B /* for private DIMENSION1 of 2-column */
97 #define LEADING_CODE_PRIVATE_21 0x9C /* for private DIMENSION2 of 1-column */
98 #define LEADING_CODE_PRIVATE_22 0x9D /* for private DIMENSION2 of 2-column */
99
100 #define LEADING_CODE_8_BIT_CONTROL 0x9E /* for `eight-bit-control' */
101
102 /* Extended leading-code. */
103 /* Start of each extended leading-codes. */
104 #define LEADING_CODE_EXT_11 0xA0 /* follows LEADING_CODE_PRIVATE_11 */
105 #define LEADING_CODE_EXT_12 0xE0 /* follows LEADING_CODE_PRIVATE_12 */
106 #define LEADING_CODE_EXT_21 0xF0 /* follows LEADING_CODE_PRIVATE_21 */
107 #define LEADING_CODE_EXT_22 0xF5 /* follows LEADING_CODE_PRIVATE_22 */
108 /* Maximum value of extended leading-codes. */
109 #define LEADING_CODE_EXT_MAX 0xFE
110
111 /* Definition of minimum/maximum charset of each DIMENSION. */
112 #define MIN_CHARSET_OFFICIAL_DIMENSION1 0x80
113 #define MAX_CHARSET_OFFICIAL_DIMENSION1 0x8F
114 #define MIN_CHARSET_OFFICIAL_DIMENSION2 0x90
115 #define MAX_CHARSET_OFFICIAL_DIMENSION2 0x99
116 #define MIN_CHARSET_PRIVATE_DIMENSION1 LEADING_CODE_EXT_11
117 #define MIN_CHARSET_PRIVATE_DIMENSION2 LEADING_CODE_EXT_21
118
119 /* Maximum value of overall charset identification number. */
120 #define MAX_CHARSET 0xFE
121
122 /* Definition of special charsets. */
123 #define CHARSET_ASCII 0 /* 0x00..0x7F */
124 #define CHARSET_8_BIT_CONTROL 0x9E /* 0x80..0x9F */
125 #define CHARSET_8_BIT_GRAPHIC 0x80 /* 0xA0..0xFF */
126
127 extern int charset_latin_iso8859_1; /* ISO8859-1 (Latin-1) */
128 extern int charset_jisx0208_1978; /* JISX0208.1978 (Japanese Kanji old set) */
129 extern int charset_jisx0208; /* JISX0208.1983 (Japanese Kanji) */
130 extern int charset_katakana_jisx0201; /* JISX0201.Kana (Japanese Katakana) */
131 extern int charset_latin_jisx0201; /* JISX0201.Roman (Japanese Roman) */
132 extern int charset_big5_1; /* Big5 Level 1 (Chinese Traditional) */
133 extern int charset_big5_2; /* Big5 Level 2 (Chinese Traditional) */
134 extern int charset_mule_unicode_0100_24ff;
135 extern int charset_mule_unicode_2500_33ff;
136 extern int charset_mule_unicode_e000_ffff;
137
138 /* Check if CH is an ASCII character or a base leading-code.
139 Nowadays, any byte can be the first byte of a character in a
140 multibyte buffer/string. So this macro name is not appropriate. */
141 #define CHAR_HEAD_P(ch) ((unsigned char) (ch) < 0xA0)
142
143 /*** GENERAL NOTE on CHARACTER REPRESENTATION ***
144
145 Firstly, the term "character" or "char" is used for a multilingual
146 character (of course, including ASCII characters), not for a byte in
147 computer memory. We use the term "code" or "byte" for the latter
148 case.
149
150 A character is identified by charset and one or two POSITION-CODEs.
151 POSITION-CODE is the position of the character in the charset. A
152 character of DIMENSION1 charset has one POSITION-CODE: POSITION-CODE-1.
153 A character of DIMENSION2 charset has two POSITION-CODE:
154 POSITION-CODE-1 and POSITION-CODE-2. The code range of
155 POSITION-CODE is 0x20..0x7F.
156
157 Emacs has two kinds of representation of a character: multi-byte
158 form (for buffers and strings) and single-word form (for character
159 objects in Emacs Lisp). The latter is called "character code"
160 hereafter. Both representations encode the information of charset
161 and POSITION-CODE but in a different way (for instance, the MSB of
162 POSITION-CODE is set in multi-byte form).
163
164 For details of the multi-byte form, see the section "2. Emacs
165 internal format handlers" of `coding.c'.
166
167 Emacs uses 19 bits for a character code. The bits are divided into
168 3 fields: FIELD1(5bits):FIELD2(7bits):FIELD3(7bits).
169
170 A character code of DIMENSION1 character uses FIELD2 to hold charset
171 and FIELD3 to hold POSITION-CODE-1. A character code of DIMENSION2
172 character uses FIELD1 to hold charset, FIELD2 and FIELD3 to hold
173 POSITION-CODE-1 and POSITION-CODE-2 respectively.
174
175 More precisely...
176
177 FIELD2 of DIMENSION1 character (except for ascii, eight-bit-control,
178 and eight-bit-graphic) is "charset - 0x70". This is to make all
179 character codes except for ASCII and 8-bit codes greater than 256.
180 So, the range of FIELD2 of DIMENSION1 character is 0, 1, or
181 0x11..0x7F.
182
183 FIELD1 of DIMENSION2 character is "charset - 0x8F" for official
184 charset and "charset - 0xE0" for private charset. So, the range of
185 FIELD1 of DIMENSION2 character is 0x01..0x1E.
186
187 -----------------------------------------------------------------------------
188 charset FIELD1 (5-bit) FIELD2 (7-bit) FIELD3 (7-bit)
189 -----------------------------------------------------------------------------
190 ascii 0 0 0x00..0x7F
191 eight-bit-control 0 1 0x00..0x1F
192 eight-bit-graphic 0 1 0x20..0x7F
193 DIMENSION1 0 charset - 0x70 POSITION-CODE-1
194 DIMENSION2(o) charset - 0x8F POSITION-CODE-1 POSITION-CODE-2
195 DIMENSION2(p) charset - 0xE0 POSITION-CODE-1 POSITION-CODE-2
196 -----------------------------------------------------------------------------
197 "(o)": official, "(p)": private
198 -----------------------------------------------------------------------------
199 */
200
201 /* Masks of each field of character code. */
202 #define CHAR_FIELD1_MASK (0x1F << 14)
203 #define CHAR_FIELD2_MASK (0x7F << 7)
204 #define CHAR_FIELD3_MASK 0x7F
205
206 /* Macros to access each field of character C. */
207 #define CHAR_FIELD1(c) (((c) & CHAR_FIELD1_MASK) >> 14)
208 #define CHAR_FIELD2(c) (((c) & CHAR_FIELD2_MASK) >> 7)
209 #define CHAR_FIELD3(c) ((c) & CHAR_FIELD3_MASK)
210
211 /* Minimum character code of character of each DIMENSION. */
212 #define MIN_CHAR_OFFICIAL_DIMENSION1 \
213 ((0x81 - 0x70) << 7)
214 #define MIN_CHAR_PRIVATE_DIMENSION1 \
215 ((MIN_CHARSET_PRIVATE_DIMENSION1 - 0x70) << 7)
216 #define MIN_CHAR_OFFICIAL_DIMENSION2 \
217 ((MIN_CHARSET_OFFICIAL_DIMENSION2 - 0x8F) << 14)
218 #define MIN_CHAR_PRIVATE_DIMENSION2 \
219 ((MIN_CHARSET_PRIVATE_DIMENSION2 - 0xE0) << 14)
220 /* Maximum character code currently used plus 1. */
221 #define MAX_CHAR (0x1F << 14)
222
223 /* 1 if C is a single byte character, else 0. */
224 #define SINGLE_BYTE_CHAR_P(c) (((unsigned)(c) & 0xFF) == (c))
225
226 /* 1 if BYTE is an ASCII character in itself, in multibyte mode. */
227 #define ASCII_BYTE_P(byte) ((byte) < 0x80)
228
229 /* A char-table containing information on each character set.
230
231 Unlike ordinary char-tables, this doesn't contain any nested tables.
232 Only the top level elements are used. Each element is a vector of
233 the following information:
234 CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION,
235 LEADING-CODE-BASE, LEADING-CODE-EXT,
236 ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE,
237 REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION,
238 PLIST.
239
240 CHARSET-ID (integer) is the identification number of the charset.
241
242 BYTES (integer) is the length of the multi-byte form of a character
243 in the charset: one of 1, 2, 3, and 4.
244
245 DIMENSION (integer) is the number of bytes to represent a character: 1 or 2.
246
247 CHARS (integer) is the number of characters in a dimension: 94 or 96.
248
249 WIDTH (integer) is the number of columns a character in the charset
250 occupies on the screen: one of 0, 1, and 2..
251
252 DIRECTION (integer) is the rendering direction of characters in the
253 charset when rendering. If 0, render from left to right, else
254 render from right to left.
255
256 LEADING-CODE-BASE (integer) is the base leading-code for the
257 charset.
258
259 LEADING-CODE-EXT (integer) is the extended leading-code for the
260 charset. All charsets of less than 0xA0 have the value 0.
261
262 ISO-FINAL-CHAR (character) is the final character of the
263 corresponding ISO 2022 charset. It is -1 for such a character
264 that is used only internally (e.g. `eight-bit-control').
265
266 ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked
267 while encoding to variants of ISO 2022 coding system, one of the
268 following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR). It
269 is -1 for such a character that is used only internally
270 (e.g. `eight-bit-control').
271
272 REVERSE-CHARSET (integer) is the charset which differs only in
273 LEFT-TO-RIGHT value from the charset. If there's no such a
274 charset, the value is -1.
275
276 SHORT-NAME (string) is the short name to refer to the charset.
277
278 LONG-NAME (string) is the long name to refer to the charset.
279
280 DESCRIPTION (string) is the description string of the charset.
281
282 PLIST (property list) may contain any type of information a user
283 wants to put and get by functions `put-charset-property' and
284 `get-charset-property' respectively. */
285 extern Lisp_Object Vcharset_table;
286
287 /* Macros to access various information of CHARSET in Vcharset_table.
288 We provide these macros for efficiency. No range check of CHARSET. */
289
290 /* Return entry of CHARSET (C integer) in Vcharset_table. */
291 #define CHARSET_TABLE_ENTRY(charset) \
292 XCHAR_TABLE (Vcharset_table)->contents[((charset) == CHARSET_ASCII \
293 ? 0 : (charset) + 128)]
294
295 /* Return information INFO-IDX of CHARSET. */
296 #define CHARSET_TABLE_INFO(charset, info_idx) \
297 XVECTOR (CHARSET_TABLE_ENTRY (charset))->contents[info_idx]
298
299 #define CHARSET_ID_IDX (0)
300 #define CHARSET_BYTES_IDX (1)
301 #define CHARSET_DIMENSION_IDX (2)
302 #define CHARSET_CHARS_IDX (3)
303 #define CHARSET_WIDTH_IDX (4)
304 #define CHARSET_DIRECTION_IDX (5)
305 #define CHARSET_LEADING_CODE_BASE_IDX (6)
306 #define CHARSET_LEADING_CODE_EXT_IDX (7)
307 #define CHARSET_ISO_FINAL_CHAR_IDX (8)
308 #define CHARSET_ISO_GRAPHIC_PLANE_IDX (9)
309 #define CHARSET_REVERSE_CHARSET_IDX (10)
310 #define CHARSET_SHORT_NAME_IDX (11)
311 #define CHARSET_LONG_NAME_IDX (12)
312 #define CHARSET_DESCRIPTION_IDX (13)
313 #define CHARSET_PLIST_IDX (14)
314 /* Size of a vector of each entry of Vcharset_table. */
315 #define CHARSET_MAX_IDX (15)
316
317 /* And several more macros to be used frequently. */
318 #define CHARSET_BYTES(charset) \
319 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_BYTES_IDX))
320 #define CHARSET_DIMENSION(charset) \
321 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_DIMENSION_IDX))
322 #define CHARSET_CHARS(charset) \
323 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_CHARS_IDX))
324 #define CHARSET_WIDTH(charset) \
325 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_WIDTH_IDX))
326 #define CHARSET_DIRECTION(charset) \
327 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_DIRECTION_IDX))
328 #define CHARSET_LEADING_CODE_BASE(charset) \
329 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_LEADING_CODE_BASE_IDX))
330 #define CHARSET_LEADING_CODE_EXT(charset) \
331 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_LEADING_CODE_EXT_IDX))
332 #define CHARSET_ISO_FINAL_CHAR(charset) \
333 XINT (CHARSET_TABLE_INFO (charset, CHARSET_ISO_FINAL_CHAR_IDX))
334 #define CHARSET_ISO_GRAPHIC_PLANE(charset) \
335 XINT (CHARSET_TABLE_INFO (charset, CHARSET_ISO_GRAPHIC_PLANE_IDX))
336 #define CHARSET_REVERSE_CHARSET(charset) \
337 XINT (CHARSET_TABLE_INFO (charset, CHARSET_REVERSE_CHARSET_IDX))
338
339 /* Macros to specify direction of a charset. */
340 #define CHARSET_DIRECTION_LEFT_TO_RIGHT 0
341 #define CHARSET_DIRECTION_RIGHT_TO_LEFT 1
342
343 /* A vector of charset symbol indexed by charset-id. This is used
344 only for returning charset symbol from C functions. */
345 extern Lisp_Object Vcharset_symbol_table;
346
347 /* Return symbol of CHARSET. */
348 #define CHARSET_SYMBOL(charset) \
349 XVECTOR (Vcharset_symbol_table)->contents[charset]
350
351 /* 1 if CHARSET is in valid value range, else 0. */
352 #define CHARSET_VALID_P(charset) \
353 ((charset) == 0 \
354 || ((charset) > 0x80 && (charset) <= MAX_CHARSET_OFFICIAL_DIMENSION2) \
355 || ((charset) >= MIN_CHARSET_PRIVATE_DIMENSION1 \
356 && (charset) <= MAX_CHARSET) \
357 || ((charset) == CHARSET_8_BIT_CONTROL) \
358 || ((charset) == CHARSET_8_BIT_GRAPHIC))
359
360 /* 1 if CHARSET is already defined, else 0. */
361 #define CHARSET_DEFINED_P(charset) \
362 (((charset) >= 0) && ((charset) <= MAX_CHARSET) \
363 && !NILP (CHARSET_TABLE_ENTRY (charset)))
364
365 /* Since the information CHARSET-BYTES and CHARSET-WIDTH of
366 Vcharset_table can be retrieved only by the first byte of
367 multi-byte form (an ASCII code or a base leading-code), we provide
368 here tables to be used by macros BYTES_BY_CHAR_HEAD and
369 WIDTH_BY_CHAR_HEAD for faster information retrieval. */
370 extern int bytes_by_char_head[256];
371 extern int width_by_char_head[256];
372
373 #define BYTES_BY_CHAR_HEAD(char_head) \
374 (ASCII_BYTE_P (char_head) ? 1 : bytes_by_char_head[char_head])
375 #define WIDTH_BY_CHAR_HEAD(char_head) \
376 (ASCII_BYTE_P (char_head) ? 1 : width_by_char_head[char_head])
377
378 /* Charset of the character C. */
379 #define CHAR_CHARSET(c) \
380 (SINGLE_BYTE_CHAR_P (c) \
381 ? (ASCII_BYTE_P (c) \
382 ? CHARSET_ASCII \
383 : (c) < 0xA0 ? CHARSET_8_BIT_CONTROL : CHARSET_8_BIT_GRAPHIC) \
384 : ((c) < MIN_CHAR_OFFICIAL_DIMENSION2 \
385 ? CHAR_FIELD2 (c) + 0x70 \
386 : ((c) < MIN_CHAR_PRIVATE_DIMENSION2 \
387 ? CHAR_FIELD1 (c) + 0x8F \
388 : CHAR_FIELD1 (c) + 0xE0)))
389
390 /* Check if two characters C1 and C2 belong to the same charset. */
391 #define SAME_CHARSET_P(c1, c2) \
392 (c1 < MIN_CHAR_OFFICIAL_DIMENSION2 \
393 ? (c1 & CHAR_FIELD2_MASK) == (c2 & CHAR_FIELD2_MASK) \
394 : (c1 & CHAR_FIELD1_MASK) == (c2 & CHAR_FIELD1_MASK))
395
396 /* Return a character of which charset is CHARSET and position-codes
397 are C1 and C2. DIMENSION1 character ignores C2. */
398 #define MAKE_CHAR(charset, c1, c2) \
399 ((charset) == CHARSET_ASCII \
400 ? (c1) & 0x7F \
401 : (((charset) == CHARSET_8_BIT_CONTROL \
402 || (charset) == CHARSET_8_BIT_GRAPHIC) \
403 ? ((c1) & 0x7F) | 0x80 \
404 : ((CHARSET_DEFINED_P (charset) \
405 ? CHARSET_DIMENSION (charset) == 1 \
406 : (charset) < MIN_CHARSET_PRIVATE_DIMENSION2) \
407 ? (((charset) - 0x70) << 7) | ((c1) <= 0 ? 0 : ((c1) & 0x7F)) \
408 : ((((charset) \
409 - ((charset) < MIN_CHARSET_PRIVATE_DIMENSION2 ? 0x8F : 0xE0)) \
410 << 14) \
411 | ((c2) <= 0 ? 0 : ((c2) & 0x7F)) \
412 | ((c1) <= 0 ? 0 : (((c1) & 0x7F) << 7))))))
413
414
415 /* If GENERICP is nonzero, return nonzero iff C is a valid normal or
416 generic character. If GENERICP is zero, return nonzero iff C is a
417 valid normal character. */
418 #define CHAR_VALID_P(c, genericp) \
419 ((c) >= 0 \
420 && (SINGLE_BYTE_CHAR_P (c) || char_valid_p (c, genericp)))
421
422 /* This default value is used when nonascii-translation-table or
423 nonascii-insert-offset fail to convert unibyte character to a valid
424 multibyte character. This makes a Latin-1 character. */
425
426 #define DEFAULT_NONASCII_INSERT_OFFSET 0x800
427
428 /* Parse multibyte string STR of length LENGTH and set BYTES to the
429 byte length of a character at STR. */
430
431 #ifdef BYTE_COMBINING_DEBUG
432
433 #define PARSE_MULTIBYTE_SEQ(str, length, bytes) \
434 do { \
435 int i = 1; \
436 while (i < (length) && ! CHAR_HEAD_P ((str)[i])) i++; \
437 (bytes) = BYTES_BY_CHAR_HEAD ((str)[0]); \
438 if ((bytes) > i) \
439 abort (); \
440 } while (0)
441
442 #else /* not BYTE_COMBINING_DEBUG */
443
444 #define PARSE_MULTIBYTE_SEQ(str, length, bytes) \
445 ((void)(length), (bytes) = BYTES_BY_CHAR_HEAD ((str)[0]))
446
447 #endif /* not BYTE_COMBINING_DEBUG */
448
449 #define VALID_LEADING_CODE_P(code) \
450 (! NILP (CHARSET_TABLE_ENTRY (code)))
451
452 /* Return 1 iff the byte sequence at unibyte string STR (LENGTH bytes)
453 is valid as a multibyte form. If valid, by a side effect, BYTES is
454 set to the byte length of the multibyte form. */
455
456 #define UNIBYTE_STR_AS_MULTIBYTE_P(str, length, bytes) \
457 (((str)[0] < 0x80 || (str)[0] >= 0xA0) \
458 ? ((bytes) = 1) \
459 : (((bytes) = BYTES_BY_CHAR_HEAD ((str)[0])), \
460 ((bytes) <= (length) \
461 && !CHAR_HEAD_P ((str)[1]) \
462 && ((bytes) == 2 \
463 ? (str)[0] != LEADING_CODE_8_BIT_CONTROL \
464 : (!CHAR_HEAD_P ((str)[2]) \
465 && ((bytes) == 3 \
466 ? (((str)[0] != LEADING_CODE_PRIVATE_11 \
467 && (str)[0] != LEADING_CODE_PRIVATE_12) \
468 || VALID_LEADING_CODE_P (str[1])) \
469 : (!CHAR_HEAD_P ((str)[3]) \
470 && VALID_LEADING_CODE_P (str[1]))))))))
471
472
473 /* Return 1 iff the byte sequence at multibyte string STR is valid as
474 a unibyte form. By a side effect, BYTES is set to the byte length
475 of one character at STR. */
476
477 #define MULTIBYTE_STR_AS_UNIBYTE_P(str, bytes) \
478 ((bytes) = BYTES_BY_CHAR_HEAD ((str)[0]), \
479 (str)[0] != LEADING_CODE_8_BIT_CONTROL)
480
481 /* The charset of character C is stored in CHARSET, and the
482 position-codes of C are stored in C1 and C2.
483 We store -1 in C2 if the dimension of the charset is 1. */
484
485 #define SPLIT_CHAR(c, charset, c1, c2) \
486 (SINGLE_BYTE_CHAR_P (c) \
487 ? ((charset \
488 = (ASCII_BYTE_P (c) \
489 ? CHARSET_ASCII \
490 : ((c) < 0xA0 ? CHARSET_8_BIT_CONTROL : CHARSET_8_BIT_GRAPHIC))), \
491 c1 = (c), c2 = -1) \
492 : ((c) & CHAR_FIELD1_MASK \
493 ? (charset = (CHAR_FIELD1 (c) \
494 + ((c) < MIN_CHAR_PRIVATE_DIMENSION2 ? 0x8F : 0xE0)), \
495 c1 = CHAR_FIELD2 (c), \
496 c2 = CHAR_FIELD3 (c)) \
497 : (charset = CHAR_FIELD2 (c) + 0x70, \
498 c1 = CHAR_FIELD3 (c), \
499 c2 = -1)))
500
501 /* Return 1 iff character C has valid printable glyph. */
502 #define CHAR_PRINTABLE_P(c) (ASCII_BYTE_P (c) || char_printable_p (c))
503
504 /* The charset of the character at STR is stored in CHARSET, and the
505 position-codes are stored in C1 and C2.
506 We store -1 in C2 if the character is just 2 bytes. */
507
508 #define SPLIT_STRING(str, len, charset, c1, c2) \
509 ((BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) < 2 \
510 || BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) > len \
511 || split_string (str, len, &charset, &c1, &c2) < 0) \
512 ? c1 = *(str), charset = CHARSET_ASCII \
513 : charset)
514
515 /* Mapping table from ISO2022's charset (specified by DIMENSION,
516 CHARS, and FINAL_CHAR) to Emacs' charset. Should be accessed by
517 macro ISO_CHARSET_TABLE (DIMENSION, CHARS, FINAL_CHAR). */
518 extern int iso_charset_table[2][2][128];
519
520 #define ISO_CHARSET_TABLE(dimension, chars, final_char) \
521 iso_charset_table[XINT (dimension) - 1][XINT (chars) > 94][XINT (final_char)]
522
523 #define BASE_LEADING_CODE_P(c) (BYTES_BY_CHAR_HEAD ((unsigned char) (c)) > 1)
524
525 /* Return how many bytes C will occupy in a multibyte buffer. */
526 #define CHAR_BYTES(c) \
527 (SINGLE_BYTE_CHAR_P (c) \
528 ? ((ASCII_BYTE_P (c) || (c) >= 0xA0) ? 1 : 2) \
529 : char_bytes (c))
530
531 /* The following two macros CHAR_STRING and STRING_CHAR are the main
532 entry points to convert between Emacs's two types of character
533 representations: multi-byte form and single-word form (character
534 code). */
535
536 /* Store multi-byte form of the character C in STR. The caller should
537 allocate at least MAX_MULTIBYTE_LENGTH bytes area at STR in
538 advance. Returns the length of the multi-byte form. If C is an
539 invalid character code, signal an error. */
540
541 #define CHAR_STRING(c, str) \
542 (SINGLE_BYTE_CHAR_P (c) \
543 ? ((ASCII_BYTE_P (c) || c >= 0xA0) \
544 ? (*(str) = (unsigned char)(c), 1) \
545 : (*(str) = LEADING_CODE_8_BIT_CONTROL, *((str)+ 1) = c + 0x20, 2)) \
546 : char_to_string (c, (unsigned char *) str))
547
548 /* Like CHAR_STRING but don't signal an error if C is invalid.
549 Value is -1 in this case. */
550
551 #define CHAR_STRING_NO_SIGNAL(c, str) \
552 (SINGLE_BYTE_CHAR_P (c) \
553 ? ((ASCII_BYTE_P (c) || c >= 0xA0) \
554 ? (*(str) = (unsigned char)(c), 1) \
555 : (*(str) = LEADING_CODE_8_BIT_CONTROL, *((str)+ 1) = c + 0x20, 2)) \
556 : char_to_string_1 (c, (unsigned char *) str))
557
558 /* Return a character code of the character of which multi-byte form
559 is at STR and the length is LEN. If STR doesn't contain valid
560 multi-byte form, only the first byte in STR is returned. */
561
562 #define STRING_CHAR(str, len) \
563 (BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) == 1 \
564 ? (unsigned char) *(str) \
565 : string_to_char (str, len, 0))
566
567 /* This is like STRING_CHAR but the third arg ACTUAL_LEN is set to the
568 length of the multi-byte form. Just to know the length, use
569 MULTIBYTE_FORM_LENGTH. */
570
571 #define STRING_CHAR_AND_LENGTH(str, len, actual_len) \
572 (BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) == 1 \
573 ? ((actual_len) = 1), (unsigned char) *(str) \
574 : string_to_char (str, len, &(actual_len)))
575
576 /* Fetch the "next" character from Lisp string STRING at byte position
577 BYTEIDX, character position CHARIDX. Store it into OUTPUT.
578
579 All the args must be side-effect-free.
580 BYTEIDX and CHARIDX must be lvalues;
581 we increment them past the character fetched. */
582
583 #define FETCH_STRING_CHAR_ADVANCE(OUTPUT, STRING, CHARIDX, BYTEIDX) \
584 if (1) \
585 { \
586 CHARIDX++; \
587 if (STRING_MULTIBYTE (STRING)) \
588 { \
589 const unsigned char *ptr = SDATA (STRING) + BYTEIDX; \
590 int space_left = SBYTES (STRING) - BYTEIDX; \
591 int actual_len; \
592 \
593 OUTPUT = STRING_CHAR_AND_LENGTH (ptr, space_left, actual_len); \
594 BYTEIDX += actual_len; \
595 } \
596 else \
597 OUTPUT = SREF (STRING, BYTEIDX++); \
598 } \
599 else
600
601 /* Like FETCH_STRING_CHAR_ADVANCE but assume STRING is multibyte. */
602
603 #define FETCH_STRING_CHAR_ADVANCE_NO_CHECK(OUTPUT, STRING, CHARIDX, BYTEIDX) \
604 if (1) \
605 { \
606 const unsigned char *fetch_string_char_ptr = SDATA (STRING) + BYTEIDX; \
607 int fetch_string_char_space_left = SBYTES (STRING) - BYTEIDX; \
608 int actual_len; \
609 \
610 OUTPUT \
611 = STRING_CHAR_AND_LENGTH (fetch_string_char_ptr, \
612 fetch_string_char_space_left, actual_len); \
613 \
614 BYTEIDX += actual_len; \
615 CHARIDX++; \
616 } \
617 else
618
619 /* Like FETCH_STRING_CHAR_ADVANCE but fetch character from the current
620 buffer. */
621
622 #define FETCH_CHAR_ADVANCE(OUTPUT, CHARIDX, BYTEIDX) \
623 if (1) \
624 { \
625 CHARIDX++; \
626 if (!NILP (current_buffer->enable_multibyte_characters)) \
627 { \
628 unsigned char *ptr = BYTE_POS_ADDR (BYTEIDX); \
629 int space_left = ((CHARIDX < GPT ? GPT_BYTE : Z_BYTE) - BYTEIDX); \
630 int actual_len; \
631 \
632 OUTPUT= STRING_CHAR_AND_LENGTH (ptr, space_left, actual_len); \
633 BYTEIDX += actual_len; \
634 } \
635 else \
636 { \
637 OUTPUT = *(BYTE_POS_ADDR (BYTEIDX)); \
638 BYTEIDX++; \
639 } \
640 } \
641 else
642
643 /* Return the length of the multi-byte form at string STR of length LEN. */
644
645 #define MULTIBYTE_FORM_LENGTH(str, len) \
646 (BYTES_BY_CHAR_HEAD (*(unsigned char *)(str)) == 1 \
647 ? 1 \
648 : multibyte_form_length (str, len))
649
650 /* If P is before LIMIT, advance P to the next character boundary. It
651 assumes that P is already at a character boundary of the sane
652 mulitbyte form whose end address is LIMIT. */
653
654 #define NEXT_CHAR_BOUNDARY(p, limit) \
655 do { \
656 if ((p) < (limit)) \
657 (p) += BYTES_BY_CHAR_HEAD (*(p)); \
658 } while (0)
659
660
661 /* If P is after LIMIT, advance P to the previous character boundary.
662 It assumes that P is already at a character boundary of the sane
663 mulitbyte form whose beginning address is LIMIT. */
664
665 #define PREV_CHAR_BOUNDARY(p, limit) \
666 do { \
667 if ((p) > (limit)) \
668 { \
669 const unsigned char *p0 = (p); \
670 do { \
671 p0--; \
672 } while (p0 >= limit && ! CHAR_HEAD_P (*p0)); \
673 (p) = (BYTES_BY_CHAR_HEAD (*p0) == (p) - p0) ? p0 : (p) - 1; \
674 } \
675 } while (0)
676
677
678 #ifdef emacs
679
680 /* Increase the buffer byte position POS_BYTE of the current buffer to
681 the next character boundary. This macro relies on the fact that
682 *GPT_ADDR and *Z_ADDR are always accessible and the values are
683 '\0'. No range checking of POS. */
684
685 #ifdef BYTE_COMBINING_DEBUG
686
687 #define INC_POS(pos_byte) \
688 do { \
689 unsigned char *p = BYTE_POS_ADDR (pos_byte); \
690 if (BASE_LEADING_CODE_P (*p)) \
691 { \
692 int len, bytes; \
693 len = Z_BYTE - pos_byte; \
694 PARSE_MULTIBYTE_SEQ (p, len, bytes); \
695 pos_byte += bytes; \
696 } \
697 else \
698 pos_byte++; \
699 } while (0)
700
701 #else /* not BYTE_COMBINING_DEBUG */
702
703 #define INC_POS(pos_byte) \
704 do { \
705 unsigned char *p = BYTE_POS_ADDR (pos_byte); \
706 pos_byte += BYTES_BY_CHAR_HEAD (*p); \
707 } while (0)
708
709 #endif /* not BYTE_COMBINING_DEBUG */
710
711 /* Decrease the buffer byte position POS_BYTE of the current buffer to
712 the previous character boundary. No range checking of POS. */
713 #define DEC_POS(pos_byte) \
714 do { \
715 unsigned char *p, *p_min; \
716 \
717 pos_byte--; \
718 if (pos_byte < GPT_BYTE) \
719 p = BEG_ADDR + pos_byte - BEG_BYTE, p_min = BEG_ADDR; \
720 else \
721 p = BEG_ADDR + GAP_SIZE + pos_byte - BEG_BYTE, p_min = GAP_END_ADDR;\
722 if (p > p_min && !CHAR_HEAD_P (*p)) \
723 { \
724 unsigned char *pend = p--; \
725 int len, bytes; \
726 if (p_min < p - MAX_MULTIBYTE_LENGTH) \
727 p_min = p - MAX_MULTIBYTE_LENGTH; \
728 while (p > p_min && !CHAR_HEAD_P (*p)) p--; \
729 len = pend + 1 - p; \
730 PARSE_MULTIBYTE_SEQ (p, len, bytes); \
731 if (bytes == len) \
732 pos_byte -= len - 1; \
733 } \
734 } while (0)
735
736 /* Increment both CHARPOS and BYTEPOS, each in the appropriate way. */
737
738 #define INC_BOTH(charpos, bytepos) \
739 do \
740 { \
741 (charpos)++; \
742 if (NILP (current_buffer->enable_multibyte_characters)) \
743 (bytepos)++; \
744 else \
745 INC_POS ((bytepos)); \
746 } \
747 while (0)
748
749 /* Decrement both CHARPOS and BYTEPOS, each in the appropriate way. */
750
751 #define DEC_BOTH(charpos, bytepos) \
752 do \
753 { \
754 (charpos)--; \
755 if (NILP (current_buffer->enable_multibyte_characters)) \
756 (bytepos)--; \
757 else \
758 DEC_POS ((bytepos)); \
759 } \
760 while (0)
761
762 /* Increase the buffer byte position POS_BYTE of the current buffer to
763 the next character boundary. This macro relies on the fact that
764 *GPT_ADDR and *Z_ADDR are always accessible and the values are
765 '\0'. No range checking of POS_BYTE. */
766
767 #ifdef BYTE_COMBINING_DEBUG
768
769 #define BUF_INC_POS(buf, pos_byte) \
770 do { \
771 unsigned char *p = BUF_BYTE_ADDRESS (buf, pos_byte); \
772 if (BASE_LEADING_CODE_P (*p)) \
773 { \
774 int len, bytes; \
775 len = BUF_Z_BYTE (buf) - pos_byte; \
776 PARSE_MULTIBYTE_SEQ (p, len, bytes); \
777 pos_byte += bytes; \
778 } \
779 else \
780 pos_byte++; \
781 } while (0)
782
783 #else /* not BYTE_COMBINING_DEBUG */
784
785 #define BUF_INC_POS(buf, pos_byte) \
786 do { \
787 unsigned char *p = BUF_BYTE_ADDRESS (buf, pos_byte); \
788 pos_byte += BYTES_BY_CHAR_HEAD (*p); \
789 } while (0)
790
791 #endif /* not BYTE_COMBINING_DEBUG */
792
793 /* Decrease the buffer byte position POS_BYTE of the current buffer to
794 the previous character boundary. No range checking of POS_BYTE. */
795 #define BUF_DEC_POS(buf, pos_byte) \
796 do { \
797 unsigned char *p, *p_min; \
798 pos_byte--; \
799 if (pos_byte < BUF_GPT_BYTE (buf)) \
800 { \
801 p = BUF_BEG_ADDR (buf) + pos_byte - BEG_BYTE; \
802 p_min = BUF_BEG_ADDR (buf); \
803 } \
804 else \
805 { \
806 p = BUF_BEG_ADDR (buf) + BUF_GAP_SIZE (buf) + pos_byte - BEG_BYTE;\
807 p_min = BUF_GAP_END_ADDR (buf); \
808 } \
809 if (p > p_min && !CHAR_HEAD_P (*p)) \
810 { \
811 unsigned char *pend = p--; \
812 int len, bytes; \
813 if (p_min < p - MAX_MULTIBYTE_LENGTH) \
814 p_min = p - MAX_MULTIBYTE_LENGTH; \
815 while (p > p_min && !CHAR_HEAD_P (*p)) p--; \
816 len = pend + 1 - p; \
817 PARSE_MULTIBYTE_SEQ (p, len, bytes); \
818 if (bytes == len) \
819 pos_byte -= len - 1; \
820 } \
821 } while (0)
822
823 #endif /* emacs */
824
825 /* This is the maximum byte length of multi-byte sequence. */
826 #define MAX_MULTIBYTE_LENGTH 4
827
828 extern void invalid_character P_ ((int));
829
830 extern int translate_char P_ ((Lisp_Object, int, int, int, int));
831 extern int split_string P_ ((const unsigned char *, int, int *,
832 unsigned char *, unsigned char *));
833 extern int char_to_string P_ ((int, unsigned char *));
834 extern int char_to_string_1 P_ ((int, unsigned char *));
835 extern int string_to_char P_ ((const unsigned char *, int, int *));
836 extern int char_printable_p P_ ((int c));
837 extern int multibyte_form_length P_ ((const unsigned char *, int));
838 extern void parse_str_as_multibyte P_ ((const unsigned char *, int, int *,
839 int *));
840 extern int str_as_multibyte P_ ((unsigned char *, int, int, int *));
841 extern int parse_str_to_multibyte P_ ((unsigned char *, int));
842 extern int str_to_multibyte P_ ((unsigned char *, int, int));
843 extern int str_as_unibyte P_ ((unsigned char *, int));
844 extern int get_charset_id P_ ((Lisp_Object));
845 extern int find_charset_in_text P_ ((const unsigned char *, int, int, int *,
846 Lisp_Object));
847 extern int strwidth P_ ((unsigned char *, int));
848 extern int c_string_width P_ ((const unsigned char *, int, int, int *, int *));
849 extern int lisp_string_width P_ ((Lisp_Object, int, int *, int *));
850 extern int char_bytes P_ ((int));
851 extern int char_valid_p P_ ((int, int));
852
853 EXFUN (Funibyte_char_to_multibyte, 1);
854
855 extern Lisp_Object Vtranslation_table_vector;
856
857 /* Return a translation table of id number ID. */
858 #define GET_TRANSLATION_TABLE(id) \
859 (XCDR(XVECTOR(Vtranslation_table_vector)->contents[(id)]))
860
861 /* A char-table for characters which may invoke auto-filling. */
862 extern Lisp_Object Vauto_fill_chars;
863
864 /* Copy LEN bytes from FROM to TO. This macro should be used only
865 when a caller knows that LEN is short and the obvious copy loop is
866 faster than calling bcopy which has some overhead. Copying a
867 multibyte sequence of a multibyte character is the typical case. */
868
869 #define BCOPY_SHORT(from, to, len) \
870 do { \
871 int i = len; \
872 const unsigned char *from_p = from; \
873 unsigned char *to_p = to; \
874 while (i--) *to_p++ = *from_p++; \
875 } while (0)
876
877 #endif /* EMACS_CHARSET_H */
878
879 /* arch-tag: 3b96db55-4961-481d-ac3e-219f46a2b3aa
880 (do not change this comment) */
Emacs repository mirror