Mercurial > emacs
annotate src/character.h @ 95656:32785fc68f99
(Ffontset_info): Fix typo.
author | Andreas Schwab <schwab@suse.de> |
---|---|
date | Sun, 08 Jun 2008 09:01:28 +0000 |
parents | abcb40dd43a2 |
children | f13a77e0e34f |
rev | line source |
---|---|
88363 | 1 /* Header for multibyte character handler. |
2 Copyright (C) 1995, 1997, 1998 Electrotechnical Laboratory, JAPAN. | |
89483 | 3 Licensed to the Free Software Foundation. |
91444
1d067b286f05
Update copyright years and GPL version.
Glenn Morris <rgm@gnu.org>
parents:
90798
diff
changeset
|
4 Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008 |
88363 | 5 National Institute of Advanced Industrial Science and Technology (AIST) |
6 Registration Number H13PRO009 | |
7 | |
8 This file is part of GNU Emacs. | |
9 | |
94994
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
91805
diff
changeset
|
10 GNU Emacs is free software: you can redistribute it and/or modify |
88363 | 11 it under the terms of the GNU General Public License as published by |
94994
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
91805
diff
changeset
|
12 the Free Software Foundation, either version 3 of the License, or |
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
91805
diff
changeset
|
13 (at your option) any later version. |
88363 | 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 | |
94994
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
91805
diff
changeset
|
21 along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */ |
88363 | 22 |
23 #ifndef EMACS_CHARACTER_H | |
24 #define EMACS_CHARACTER_H | |
25 | |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
26 /* character code 1st byte byte sequence |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
27 -------------- -------- ------------- |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
28 0-7F 00..7F 0xxxxxxx |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
29 80-7FF C2..DF 110xxxxx 10xxxxxx |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
30 800-FFFF E0..EF 1110xxxx 10xxxxxx 10xxxxxx |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
31 10000-1FFFFF F0..F7 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
32 200000-3FFF7F F8 11111000 1000xxxx 10xxxxxx 10xxxxxx 10xxxxxx |
89692 | 33 3FFF80-3FFFFF C0..C1 1100000x 10xxxxxx (for eight-bit-char) |
34 400000-... invalid | |
88363 | 35 |
89692 | 36 invalid 1st byte 80..BF 10xxxxxx |
37 F9..FF 11111xxx (xxx != 000) | |
88363 | 38 */ |
39 | |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
40 /* Maximum character code ((1 << CHARACTERBITS) - 1). */ |
88363 | 41 #define MAX_CHAR 0x3FFFFF |
42 | |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
43 /* Maximum Unicode character code. */ |
88363 | 44 #define MAX_UNICODE_CHAR 0x10FFFF |
45 | |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
46 /* Maximum N-byte character codes. */ |
88363 | 47 #define MAX_1_BYTE_CHAR 0x7F |
48 #define MAX_2_BYTE_CHAR 0x7FF | |
49 #define MAX_3_BYTE_CHAR 0xFFFF | |
50 #define MAX_4_BYTE_CHAR 0x1FFFFF | |
51 #define MAX_5_BYTE_CHAR 0x3FFF7F | |
52 | |
90791
0603ad3252c9
(MIN_MULTIBYTE_LEADING_CODE)
Kenichi Handa <handa@m17n.org>
parents:
90533
diff
changeset
|
53 /* Minimum leading code of multibyte characters. */ |
0603ad3252c9
(MIN_MULTIBYTE_LEADING_CODE)
Kenichi Handa <handa@m17n.org>
parents:
90533
diff
changeset
|
54 #define MIN_MULTIBYTE_LEADING_CODE 0xC0 |
0603ad3252c9
(MIN_MULTIBYTE_LEADING_CODE)
Kenichi Handa <handa@m17n.org>
parents:
90533
diff
changeset
|
55 /* Maximum leading code of multibyte characters. */ |
0603ad3252c9
(MIN_MULTIBYTE_LEADING_CODE)
Kenichi Handa <handa@m17n.org>
parents:
90533
diff
changeset
|
56 #define MAX_MULTIBYTE_LEADING_CODE 0xF8 |
0603ad3252c9
(MIN_MULTIBYTE_LEADING_CODE)
Kenichi Handa <handa@m17n.org>
parents:
90533
diff
changeset
|
57 |
88946
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
58 /* Nonzero iff C is a character that corresponds to a raw 8-bit |
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
59 byte. */ |
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
60 #define CHAR_BYTE8_P(c) ((c) > MAX_5_BYTE_CHAR) |
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
61 |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
62 /* Return the character code for raw 8-bit byte BYTE. */ |
88363 | 63 #define BYTE8_TO_CHAR(byte) ((byte) + 0x3FFF00) |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
64 |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
65 /* Return the raw 8-bit byte for character C. */ |
88946
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
66 #define CHAR_TO_BYTE8(c) \ |
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
67 (CHAR_BYTE8_P (c) \ |
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
68 ? (c) - 0x3FFF00 \ |
233c080b5756
(CHAR_TO_BYTE8): If C is not eight-bit char, call
Kenichi Handa <handa@m17n.org>
parents:
88915
diff
changeset
|
69 : multibyte_char_to_unibyte (c, Qnil)) |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
70 |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
71 /* Nonzero iff BYTE is the 1st byte of a multibyte form of a character |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
72 that corresponds to a raw 8-bit byte. */ |
88363 | 73 #define CHAR_BYTE8_HEAD_P(byte) ((byte) == 0xC0 || (byte) == 0xC1) |
74 | |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
75 /* Mapping table from unibyte chars to multibyte chars. */ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
76 extern int unibyte_to_multibyte_table[256]; |
89018
a9f683a73092
(MAKE_CHAR_UNIBYTE, MAKE_CHAR_MULTIBYTE): New macros.
Kenichi Handa <handa@m17n.org>
parents:
88946
diff
changeset
|
77 |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
78 /* Convert the unibyte character C to the corresponding multibyte |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
79 character. If C can't be converted, return C. */ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
80 #define unibyte_char_to_multibyte(c) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
81 ((c) < 256 ? unibyte_to_multibyte_table[(c)] : (c)) |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
82 |
90019
1987dfad4543
(unibyte_has_multibyte_table): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
89911
diff
changeset
|
83 /* Nth element is 1 iff unibyte char N can be mapped to a multibyte |
1987dfad4543
(unibyte_has_multibyte_table): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
89911
diff
changeset
|
84 char. */ |
1987dfad4543
(unibyte_has_multibyte_table): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
89911
diff
changeset
|
85 extern char unibyte_has_multibyte_table[256]; |
1987dfad4543
(unibyte_has_multibyte_table): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
89911
diff
changeset
|
86 |
1987dfad4543
(unibyte_has_multibyte_table): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
89911
diff
changeset
|
87 #define UNIBYTE_CHAR_HAS_MULTIBYTE_P(c) (unibyte_has_multibyte_table[(c)]) |
1987dfad4543
(unibyte_has_multibyte_table): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
89911
diff
changeset
|
88 |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
89 /* If C is not ASCII, make it unibyte. */ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
90 #define MAKE_CHAR_UNIBYTE(c) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
91 do { \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
92 if (! ASCII_CHAR_P (c)) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
93 c = CHAR_TO_BYTE8 (c); \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
94 } while (0) |
89018
a9f683a73092
(MAKE_CHAR_UNIBYTE, MAKE_CHAR_MULTIBYTE): New macros.
Kenichi Handa <handa@m17n.org>
parents:
88946
diff
changeset
|
95 |
a9f683a73092
(MAKE_CHAR_UNIBYTE, MAKE_CHAR_MULTIBYTE): New macros.
Kenichi Handa <handa@m17n.org>
parents:
88946
diff
changeset
|
96 |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
97 /* If C is not ASCII, make it multibyte. It assumes C < 256. */ |
95576
abcb40dd43a2
(MAKE_CHAR_MULTIBYTE): Check the arg is a (uni)byte.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
94994
diff
changeset
|
98 #define MAKE_CHAR_MULTIBYTE(c) \ |
abcb40dd43a2
(MAKE_CHAR_MULTIBYTE): Check the arg is a (uni)byte.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
94994
diff
changeset
|
99 (eassert ((c) >= 0 && (c) < 256), (c) = unibyte_to_multibyte_table[(c)]) |
89018
a9f683a73092
(MAKE_CHAR_UNIBYTE, MAKE_CHAR_MULTIBYTE): New macros.
Kenichi Handa <handa@m17n.org>
parents:
88946
diff
changeset
|
100 |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
101 /* This is the maximum byte length of multibyte form. */ |
88363 | 102 #define MAX_MULTIBYTE_LENGTH 5 |
103 | |
89887
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
104 /* Return a Lisp character whose character code is C. It assumes C is |
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
105 a valid character code. */ |
88363 | 106 #define make_char(c) make_number (c) |
107 | |
108 /* Nonzero iff C is an ASCII byte. */ | |
109 #define ASCII_BYTE_P(c) ((unsigned) (c) < 0x80) | |
110 | |
111 /* Nonzero iff X is a character. */ | |
112 #define CHARACTERP(x) (NATNUMP (x) && XFASTINT (x) <= MAX_CHAR) | |
113 | |
88832 | 114 /* Nonzero iff C is valid as a character code. GENERICP is not used |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
115 now. */ |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
116 #define CHAR_VALID_P(c, genericp) ((unsigned) (c) <= MAX_CHAR) |
88363 | 117 |
118 /* Check if Lisp object X is a character or not. */ | |
90533 | 119 #define CHECK_CHARACTER(x) \ |
120 CHECK_TYPE (CHARACTERP (x), Qcharacterp, x) | |
88363 | 121 |
89483 | 122 #define CHECK_CHARACTER_CAR(x) \ |
123 do { \ | |
124 Lisp_Object tmp = XCAR (x); \ | |
125 CHECK_CHARACTER (tmp); \ | |
126 XSETCAR ((x), tmp); \ | |
127 } while (0) | |
128 | |
129 #define CHECK_CHARACTER_CDR(x) \ | |
130 do { \ | |
131 Lisp_Object tmp = XCDR (x); \ | |
132 CHECK_CHARACTER (tmp); \ | |
133 XSETCDR ((x), tmp); \ | |
134 } while (0) | |
135 | |
88363 | 136 /* Nonzero iff C is an ASCII character. */ |
137 #define ASCII_CHAR_P(c) ((unsigned) (c) < 0x80) | |
138 | |
139 /* Nonzero iff C is a character of code less than 0x100. */ | |
140 #define SINGLE_BYTE_CHAR_P(c) ((unsigned) (c) < 0x100) | |
141 | |
142 /* Nonzero if character C has a printable glyph. */ | |
143 #define CHAR_PRINTABLE_P(c) \ | |
144 (((c) >= 32 && ((c) < 127) \ | |
145 || ! NILP (CHAR_TABLE_REF (Vprintable_chars, (c))))) | |
146 | |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
147 /* Return byte length of multibyte form for character C. */ |
88363 | 148 #define CHAR_BYTES(c) \ |
149 ( (c) <= MAX_1_BYTE_CHAR ? 1 \ | |
150 : (c) <= MAX_2_BYTE_CHAR ? 2 \ | |
151 : (c) <= MAX_3_BYTE_CHAR ? 3 \ | |
152 : (c) <= MAX_4_BYTE_CHAR ? 4 \ | |
153 : (c) <= MAX_5_BYTE_CHAR ? 5 \ | |
154 : 2) | |
155 | |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
156 |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
157 /* Return the leading code of multibyte form of C. */ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
158 #define CHAR_LEADING_CODE(c) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
159 ((c) <= MAX_1_BYTE_CHAR ? c \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
160 : (c) <= MAX_2_BYTE_CHAR ? (0xC0 | ((c) >> 6)) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
161 : (c) <= MAX_3_BYTE_CHAR ? (0xE0 | ((c) >> 12)) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
162 : (c) <= MAX_4_BYTE_CHAR ? (0xF0 | ((c) >> 18)) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
163 : (c) <= MAX_5_BYTE_CHAR ? 0xF8 \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
164 : (0xC0 | (((c) >> 6) & 0x01))) |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
165 |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
166 |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
167 /* Store multibyte form of the character C in P. The caller should |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
168 allocate at least MAX_MULTIBYTE_LENGTH bytes area at P in advance. |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
169 Returns the length of the multibyte form. */ |
88363 | 170 |
171 #define CHAR_STRING(c, p) \ | |
172 ((unsigned) (c) <= MAX_1_BYTE_CHAR \ | |
173 ? ((p)[0] = (c), \ | |
174 1) \ | |
175 : (unsigned) (c) <= MAX_2_BYTE_CHAR \ | |
176 ? ((p)[0] = (0xC0 | ((c) >> 6)), \ | |
177 (p)[1] = (0x80 | ((c) & 0x3F)), \ | |
178 2) \ | |
179 : (unsigned) (c) <= MAX_3_BYTE_CHAR \ | |
180 ? ((p)[0] = (0xE0 | ((c) >> 12)), \ | |
181 (p)[1] = (0x80 | (((c) >> 6) & 0x3F)), \ | |
182 (p)[2] = (0x80 | ((c) & 0x3F)), \ | |
183 3) \ | |
90798
ad70c7654800
(CHAR_STRING): Cast C to unsigned on calling
Kenichi Handa <handa@m17n.org>
parents:
90791
diff
changeset
|
184 : char_string ((unsigned) c, p)) |
88363 | 185 |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
186 /* Store multibyte form of byte B in P. The caller should allocate at |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
187 least MAX_MULTIBYTE_LENGTH bytes area at P in advance. Returns the |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
188 length of the multibyte form. */ |
89038
f60ed671d6e4
(BYTE8_STRING): New macro.
Kenichi Handa <handa@m17n.org>
parents:
89018
diff
changeset
|
189 |
f60ed671d6e4
(BYTE8_STRING): New macro.
Kenichi Handa <handa@m17n.org>
parents:
89018
diff
changeset
|
190 #define BYTE8_STRING(b, p) \ |
f60ed671d6e4
(BYTE8_STRING): New macro.
Kenichi Handa <handa@m17n.org>
parents:
89018
diff
changeset
|
191 ((p)[0] = (0xC0 | (((b) >> 6) & 0x01)), \ |
90051 | 192 (p)[1] = (0x80 | ((b) & 0x3F)), \ |
89038
f60ed671d6e4
(BYTE8_STRING): New macro.
Kenichi Handa <handa@m17n.org>
parents:
89018
diff
changeset
|
193 2) |
f60ed671d6e4
(BYTE8_STRING): New macro.
Kenichi Handa <handa@m17n.org>
parents:
89018
diff
changeset
|
194 |
88363 | 195 |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
196 /* Store multibyte form of the character C in P. The caller should |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
197 allocate at least MAX_MULTIBYTE_LENGTH bytes area at P in advance. |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
198 And, advance P to the end of the multibyte form. */ |
88363 | 199 |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
200 #define CHAR_STRING_ADVANCE(c, p) \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
201 do { \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
202 if ((c) <= MAX_1_BYTE_CHAR) \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
203 *(p)++ = (c); \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
204 else if ((c) <= MAX_2_BYTE_CHAR) \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
205 *(p)++ = (0xC0 | ((c) >> 6)), \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
206 *(p)++ = (0x80 | ((c) & 0x3F)); \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
207 else if ((c) <= MAX_3_BYTE_CHAR) \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
208 *(p)++ = (0xE0 | ((c) >> 12)), \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
209 *(p)++ = (0x80 | (((c) >> 6) & 0x3F)), \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
210 *(p)++ = (0x80 | ((c) & 0x3F)); \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
211 else \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
212 (p) += char_string ((c), (p)); \ |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
213 } while (0) |
88363 | 214 |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
215 |
88363 | 216 /* Nonzero iff BYTE starts a non-ASCII character in a multibyte |
217 form. */ | |
218 #define LEADING_CODE_P(byte) (((byte) & 0xC0) == 0xC0) | |
219 | |
88873
7d441bc35e9b
(TRAILING_CODE_P): New macro.
Kenichi Handa <handa@m17n.org>
parents:
88832
diff
changeset
|
220 /* Nonzero iff BYTE is a trailing code of a non-ASCII character in a |
7d441bc35e9b
(TRAILING_CODE_P): New macro.
Kenichi Handa <handa@m17n.org>
parents:
88832
diff
changeset
|
221 multibyte form. */ |
7d441bc35e9b
(TRAILING_CODE_P): New macro.
Kenichi Handa <handa@m17n.org>
parents:
88832
diff
changeset
|
222 #define TRAILING_CODE_P(byte) (((byte) & 0xC0) == 0x80) |
7d441bc35e9b
(TRAILING_CODE_P): New macro.
Kenichi Handa <handa@m17n.org>
parents:
88832
diff
changeset
|
223 |
88428
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
224 /* Nonzero iff BYTE starts a character in a multibyte form. |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
225 This is equivalent to: |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
226 (ASCII_BYTE_P (byte) || LEADING_CODE_P (byte)) */ |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
227 #define CHAR_HEAD_P(byte) (((byte) & 0xC0) != 0x80) |
5eaa8c11ab45
(CHAR_VALID_P): Don't call CHARACTERP.
Kenichi Handa <handa@m17n.org>
parents:
88363
diff
changeset
|
228 |
88363 | 229 /* Just kept for backward compatibility. This macro will be removed |
230 in the future. */ | |
231 #define BASE_LEADING_CODE_P LEADING_CODE_P | |
232 | |
233 /* How many bytes a character that starts with BYTE occupies in a | |
234 multibyte form. */ | |
235 #define BYTES_BY_CHAR_HEAD(byte) \ | |
236 (!((byte) & 0x80) ? 1 \ | |
237 : !((byte) & 0x20) ? 2 \ | |
238 : !((byte) & 0x10) ? 3 \ | |
239 : !((byte) & 0x08) ? 4 \ | |
240 : 5) | |
241 | |
242 | |
243 /* Return the length of the multi-byte form at string STR of length | |
244 LEN while assuming that STR points a valid multi-byte form. As | |
245 this macro isn't necessary anymore, all callers will be changed to | |
246 use BYTES_BY_CHAR_HEAD directly in the future. */ | |
247 | |
248 #define MULTIBYTE_FORM_LENGTH(str, len) \ | |
249 BYTES_BY_CHAR_HEAD (*(str)) | |
250 | |
251 /* Parse multibyte string STR of length LENGTH and set BYTES to the | |
252 byte length of a character at STR while assuming that STR points a | |
253 valid multibyte form. As this macro isn't necessary anymore, all | |
254 callers will be changed to use BYTES_BY_CHAR_HEAD directly in the | |
255 future. */ | |
256 | |
257 #define PARSE_MULTIBYTE_SEQ(str, length, bytes) \ | |
258 (bytes) = BYTES_BY_CHAR_HEAD (*(str)) | |
259 | |
260 /* The byte length of multibyte form at unibyte string P ending at | |
261 PEND. If STR doesn't point a valid multibyte form, return 0. */ | |
262 | |
263 #define MULTIBYTE_LENGTH(p, pend) \ | |
264 (p >= pend ? 0 \ | |
265 : !((p)[0] & 0x80) ? 1 \ | |
266 : ((p + 1 >= pend) || (((p)[1] & 0xC0) != 0x80)) ? 0 \ | |
267 : ((p)[0] & 0xE0) == 0xC0 ? 2 \ | |
268 : ((p + 2 >= pend) || (((p)[2] & 0xC0) != 0x80)) ? 0 \ | |
269 : ((p)[0] & 0xF0) == 0xE0 ? 3 \ | |
270 : ((p + 3 >= pend) || (((p)[3] & 0xC0) != 0x80)) ? 0 \ | |
271 : ((p)[0] & 0xF8) == 0xF0 ? 4 \ | |
272 : ((p + 4 >= pend) || (((p)[4] & 0xC0) != 0x80)) ? 0 \ | |
273 : (p)[0] == 0xF8 && ((p)[1] & 0xF0) == 0x80 ? 5 \ | |
274 : 0) | |
275 | |
276 | |
277 /* Like MULTIBYTE_LENGTH but don't check the ending address. */ | |
278 | |
279 #define MULTIBYTE_LENGTH_NO_CHECK(p) \ | |
280 (!((p)[0] & 0x80) ? 1 \ | |
281 : ((p)[1] & 0xC0) != 0x80 ? 0 \ | |
282 : ((p)[0] & 0xE0) == 0xC0 ? 2 \ | |
283 : ((p)[2] & 0xC0) != 0x80 ? 0 \ | |
284 : ((p)[0] & 0xF0) == 0xE0 ? 3 \ | |
285 : ((p)[3] & 0xC0) != 0x80 ? 0 \ | |
286 : ((p)[0] & 0xF8) == 0xF0 ? 4 \ | |
287 : ((p)[4] & 0xC0) != 0x80 ? 0 \ | |
288 : (p)[0] == 0xF8 && ((p)[1] & 0xF0) == 0x80 ? 5 \ | |
289 : 0) | |
290 | |
89483 | 291 /* If P is before LIMIT, advance P to the next character boundary. It |
292 assumes that P is already at a character boundary of the sane | |
293 mulitbyte form whose end address is LIMIT. */ | |
294 | |
295 #define NEXT_CHAR_BOUNDARY(p, limit) \ | |
296 do { \ | |
297 if ((p) < (limit)) \ | |
298 (p) += BYTES_BY_CHAR_HEAD (*(p)); \ | |
299 } while (0) | |
300 | |
301 | |
302 /* If P is after LIMIT, advance P to the previous character boundary. | |
303 It assumes that P is already at a character boundary of the sane | |
304 mulitbyte form whose beginning address is LIMIT. */ | |
305 | |
306 #define PREV_CHAR_BOUNDARY(p, limit) \ | |
307 do { \ | |
308 if ((p) > (limit)) \ | |
309 { \ | |
310 const unsigned char *p0 = (p); \ | |
311 do { \ | |
312 p0--; \ | |
313 } while (p0 >= limit && ! CHAR_HEAD_P (*p0)); \ | |
314 (p) = (BYTES_BY_CHAR_HEAD (*p0) == (p) - p0) ? p0 : (p) - 1; \ | |
315 } \ | |
316 } while (0) | |
88363 | 317 |
318 /* Return the character code of character whose multibyte form is at | |
319 P. The argument LEN is ignored. It will be removed in the | |
320 future. */ | |
321 | |
322 #define STRING_CHAR(p, len) \ | |
323 (!((p)[0] & 0x80) \ | |
324 ? (p)[0] \ | |
325 : ! ((p)[0] & 0x20) \ | |
326 ? (((((p)[0] & 0x1F) << 6) \ | |
327 | ((p)[1] & 0x3F)) \ | |
328 + (((unsigned char) (p)[0]) < 0xC2 ? 0x3FFF80 : 0)) \ | |
329 : ! ((p)[0] & 0x10) \ | |
330 ? ((((p)[0] & 0x0F) << 12) \ | |
331 | (((p)[1] & 0x3F) << 6) \ | |
332 | ((p)[2] & 0x3F)) \ | |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
333 : string_char ((p), NULL, NULL)) |
88363 | 334 |
335 | |
336 /* Like STRING_CHAR but set ACTUAL_LEN to the length of multibyte | |
337 form. The argument LEN is ignored. It will be removed in the | |
338 future. */ | |
339 | |
340 #define STRING_CHAR_AND_LENGTH(p, len, actual_len) \ | |
341 (!((p)[0] & 0x80) \ | |
342 ? ((actual_len) = 1, (p)[0]) \ | |
343 : ! ((p)[0] & 0x20) \ | |
344 ? ((actual_len) = 2, \ | |
345 (((((p)[0] & 0x1F) << 6) \ | |
346 | ((p)[1] & 0x3F)) \ | |
347 + (((unsigned char) (p)[0]) < 0xC2 ? 0x3FFF80 : 0))) \ | |
348 : ! ((p)[0] & 0x10) \ | |
349 ? ((actual_len) = 3, \ | |
350 ((((p)[0] & 0x0F) << 12) \ | |
351 | (((p)[1] & 0x3F) << 6) \ | |
352 | ((p)[2] & 0x3F))) \ | |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
353 : string_char ((p), NULL, &actual_len)) |
88363 | 354 |
355 | |
89887
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
356 /* Like STRING_CHAR but advance P to the end of multibyte form. */ |
88363 | 357 |
358 #define STRING_CHAR_ADVANCE(p) \ | |
359 (!((p)[0] & 0x80) \ | |
360 ? *(p)++ \ | |
361 : ! ((p)[0] & 0x20) \ | |
362 ? ((p) += 2, \ | |
363 ((((p)[-2] & 0x1F) << 6) \ | |
364 | ((p)[-1] & 0x3F) \ | |
89483 | 365 | ((unsigned char) ((p)[-2]) < 0xC2 ? 0x3FFF80 : 0))) \ |
88363 | 366 : ! ((p)[0] & 0x10) \ |
367 ? ((p) += 3, \ | |
368 ((((p)[-3] & 0x0F) << 12) \ | |
369 | (((p)[-2] & 0x3F) << 6) \ | |
370 | ((p)[-1] & 0x3F))) \ | |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
371 : string_char ((p), &(p), NULL)) |
88363 | 372 |
373 | |
374 /* Fetch the "next" character from Lisp string STRING at byte position | |
375 BYTEIDX, character position CHARIDX. Store it into OUTPUT. | |
376 | |
377 All the args must be side-effect-free. | |
378 BYTEIDX and CHARIDX must be lvalues; | |
379 we increment them past the character fetched. */ | |
380 | |
381 #define FETCH_STRING_CHAR_ADVANCE(OUTPUT, STRING, CHARIDX, BYTEIDX) \ | |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
382 do \ |
88363 | 383 { \ |
384 CHARIDX++; \ | |
385 if (STRING_MULTIBYTE (STRING)) \ | |
386 { \ | |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
387 unsigned char *ptr = &SDATA (STRING)[BYTEIDX]; \ |
88363 | 388 int len; \ |
389 \ | |
390 OUTPUT = STRING_CHAR_AND_LENGTH (ptr, 0, len); \ | |
391 BYTEIDX += len; \ | |
392 } \ | |
393 else \ | |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
394 { \ |
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
395 OUTPUT = SREF (STRING, BYTEIDX); \ |
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
396 BYTEIDX++; \ |
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
397 } \ |
88363 | 398 } \ |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
399 while (0) |
88363 | 400 |
89887
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
401 /* Like FETCH_STRING_CHAR_ADVANCE but return a multibyte character eve |
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
402 if STRING is unibyte. */ |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
403 |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
404 #define FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE(OUTPUT, STRING, CHARIDX, BYTEIDX) \ |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
405 do \ |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
406 { \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
407 CHARIDX++; \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
408 if (STRING_MULTIBYTE (STRING)) \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
409 { \ |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
410 unsigned char *ptr = &SDATA (STRING)[BYTEIDX]; \ |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
411 int len; \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
412 \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
413 OUTPUT = STRING_CHAR_AND_LENGTH (ptr, 0, len); \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
414 BYTEIDX += len; \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
415 } \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
416 else \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
417 { \ |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
418 OUTPUT = SREF (STRING, BYTEIDX); \ |
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
419 BYTEIDX++; \ |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
420 MAKE_CHAR_MULTIBYTE (OUTPUT); \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
421 } \ |
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
422 } \ |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
423 while (0) |
89053
06a2cbbeaee9
(LEADING_CODE_LATIN_1_MIN)
Kenichi Handa <handa@m17n.org>
parents:
89038
diff
changeset
|
424 |
88363 | 425 |
426 /* Like FETCH_STRING_CHAR_ADVANCE but assumes STRING is multibyte. */ | |
427 | |
428 #define FETCH_STRING_CHAR_ADVANCE_NO_CHECK(OUTPUT, STRING, CHARIDX, BYTEIDX) \ | |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
429 do \ |
88363 | 430 { \ |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
431 unsigned char *ptr = &SDATA (STRING)[BYTEIDX]; \ |
88363 | 432 int len; \ |
433 \ | |
434 OUTPUT = STRING_CHAR_AND_LENGTH (ptr, 0, len); \ | |
435 BYTEIDX += len; \ | |
436 CHARIDX++; \ | |
437 } \ | |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
438 while (0) |
88363 | 439 |
440 | |
441 /* Like FETCH_STRING_CHAR_ADVANCE but fetch character from the current | |
442 buffer. */ | |
443 | |
444 #define FETCH_CHAR_ADVANCE(OUTPUT, CHARIDX, BYTEIDX) \ | |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
445 do \ |
88363 | 446 { \ |
447 CHARIDX++; \ | |
448 if (!NILP (current_buffer->enable_multibyte_characters)) \ | |
449 { \ | |
450 unsigned char *ptr = BYTE_POS_ADDR (BYTEIDX); \ | |
451 int len; \ | |
452 \ | |
453 OUTPUT= STRING_CHAR_AND_LENGTH (ptr, 0, len); \ | |
454 BYTEIDX += len; \ | |
455 } \ | |
456 else \ | |
457 { \ | |
458 OUTPUT = *(BYTE_POS_ADDR (BYTEIDX)); \ | |
459 BYTEIDX++; \ | |
460 } \ | |
461 } \ | |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
462 while (0) |
88363 | 463 |
464 | |
89887
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
465 /* Like FETCH_CHAR_ADVANCE but assumes the current buffer is multibyte. */ |
88363 | 466 |
467 #define FETCH_CHAR_ADVANCE_NO_CHECK(OUTPUT, CHARIDX, BYTEIDX) \ | |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
468 do \ |
88363 | 469 { \ |
470 unsigned char *ptr = BYTE_POS_ADDR (BYTEIDX); \ | |
471 int len; \ | |
472 \ | |
473 OUTPUT= STRING_CHAR_AND_LENGTH (ptr, 0, len); \ | |
474 BYTEIDX += len; \ | |
475 CHARIDX++; \ | |
476 } \ | |
91620
acb32817f0e8
Use "do...while (0)", not "if (1)..else" in macro definitions.
Jan Djärv <jan.h.d@swipnet.se>
parents:
91444
diff
changeset
|
477 while (0) |
88363 | 478 |
479 | |
480 /* Increase the buffer byte position POS_BYTE of the current buffer to | |
481 the next character boundary. No range checking of POS. */ | |
482 | |
483 #define INC_POS(pos_byte) \ | |
484 do { \ | |
485 unsigned char *p = BYTE_POS_ADDR (pos_byte); \ | |
486 pos_byte += BYTES_BY_CHAR_HEAD (*p); \ | |
487 } while (0) | |
488 | |
489 | |
490 /* Decrease the buffer byte position POS_BYTE of the current buffer to | |
491 the previous character boundary. No range checking of POS. */ | |
492 | |
493 #define DEC_POS(pos_byte) \ | |
494 do { \ | |
495 unsigned char *p; \ | |
496 \ | |
497 pos_byte--; \ | |
498 if (pos_byte < GPT_BYTE) \ | |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
499 p = BEG_ADDR + pos_byte - BEG_BYTE; \ |
88363 | 500 else \ |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
501 p = BEG_ADDR + GAP_SIZE + pos_byte - BEG_BYTE;\ |
88363 | 502 while (!CHAR_HEAD_P (*p)) \ |
503 { \ | |
504 p--; \ | |
505 pos_byte--; \ | |
506 } \ | |
507 } while (0) | |
508 | |
509 /* Increment both CHARPOS and BYTEPOS, each in the appropriate way. */ | |
510 | |
511 #define INC_BOTH(charpos, bytepos) \ | |
512 do \ | |
513 { \ | |
514 (charpos)++; \ | |
515 if (NILP (current_buffer->enable_multibyte_characters)) \ | |
516 (bytepos)++; \ | |
517 else \ | |
518 INC_POS ((bytepos)); \ | |
519 } \ | |
520 while (0) | |
521 | |
522 | |
523 /* Decrement both CHARPOS and BYTEPOS, each in the appropriate way. */ | |
524 | |
525 #define DEC_BOTH(charpos, bytepos) \ | |
526 do \ | |
527 { \ | |
528 (charpos)--; \ | |
529 if (NILP (current_buffer->enable_multibyte_characters)) \ | |
530 (bytepos)--; \ | |
531 else \ | |
532 DEC_POS ((bytepos)); \ | |
533 } \ | |
534 while (0) | |
535 | |
536 | |
537 /* Increase the buffer byte position POS_BYTE of the current buffer to | |
538 the next character boundary. This macro relies on the fact that | |
539 *GPT_ADDR and *Z_ADDR are always accessible and the values are | |
540 '\0'. No range checking of POS_BYTE. */ | |
541 | |
542 #define BUF_INC_POS(buf, pos_byte) \ | |
543 do { \ | |
544 unsigned char *p = BUF_BYTE_ADDRESS (buf, pos_byte); \ | |
545 pos_byte += BYTES_BY_CHAR_HEAD (*p); \ | |
546 } while (0) | |
547 | |
548 | |
549 /* Decrease the buffer byte position POS_BYTE of the current buffer to | |
550 the previous character boundary. No range checking of POS_BYTE. */ | |
551 | |
552 #define BUF_DEC_POS(buf, pos_byte) \ | |
553 do { \ | |
554 unsigned char *p; \ | |
555 pos_byte--; \ | |
556 if (pos_byte < BUF_GPT_BYTE (buf)) \ | |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
557 p = BUF_BEG_ADDR (buf) + pos_byte - BEG_BYTE; \ |
88363 | 558 else \ |
91805
c330bf7419fd
(FETCH_STRING_CHAR_ADVANCE, FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
91620
diff
changeset
|
559 p = BUF_BEG_ADDR (buf) + BUF_GAP_SIZE (buf) + pos_byte - BEG_BYTE;\ |
88363 | 560 while (!CHAR_HEAD_P (*p)) \ |
561 { \ | |
562 p--; \ | |
563 pos_byte--; \ | |
564 } \ | |
565 } while (0) | |
566 | |
567 | |
89887
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
568 /* If C is a character to be unified with a Unicode character, return |
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
569 the unified Unicode character. */ |
028a1f06f612
(LEADING_CODE_LATIN_1_MIN) (LEADING_CODE_LATIN_1_MAX): Delete these
Kenichi Handa <handa@m17n.org>
parents:
89692
diff
changeset
|
570 |
88742
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
571 #define MAYBE_UNIFY_CHAR(c) \ |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
572 if (c > MAX_UNICODE_CHAR \ |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
573 && CHAR_TABLE_P (Vchar_unify_table)) \ |
88742
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
574 { \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
575 Lisp_Object val; \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
576 int unified; \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
577 \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
578 val = CHAR_TABLE_REF (Vchar_unify_table, c); \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
579 if (! NILP (val)) \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
580 { \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
581 if (SYMBOLP (val)) \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
582 { \ |
88873
7d441bc35e9b
(TRAILING_CODE_P): New macro.
Kenichi Handa <handa@m17n.org>
parents:
88832
diff
changeset
|
583 Funify_charset (val, Qnil, Qnil); \ |
88742
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
584 val = CHAR_TABLE_REF (Vchar_unify_table, c); \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
585 } \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
586 if ((unified = XINT (val)) >= 0) \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
587 c = unified; \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
588 } \ |
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
589 } \ |
88363 | 590 else |
591 | |
88742
55e36a0cf0ee
(MAYBE_UNIFY_CHAR): Adjusted for the change of
Kenichi Handa <handa@m17n.org>
parents:
88545
diff
changeset
|
592 |
88363 | 593 /* Return the width of ASCII character C. The width is measured by |
594 how many columns occupied on the screen when displayed in the | |
595 current buffer. */ | |
596 | |
597 #define ASCII_CHAR_WIDTH(c) \ | |
598 (c < 0x20 \ | |
599 ? (c == '\t' \ | |
600 ? XFASTINT (current_buffer->tab_width) \ | |
601 : (c == '\n' ? 0 : (NILP (current_buffer->ctl_arrow) ? 4 : 2))) \ | |
602 : (c < 0x7f \ | |
603 ? 1 \ | |
604 : ((NILP (current_buffer->ctl_arrow) ? 4 : 2)))) | |
605 | |
606 /* Return the width of character C. The width is measured by how many | |
607 columns occupied on the screen when displayed in the current | |
608 buffer. */ | |
609 | |
610 #define CHAR_WIDTH(c) \ | |
611 (ASCII_CHAR_P (c) \ | |
612 ? ASCII_CHAR_WIDTH (c) \ | |
613 : XINT (CHAR_TABLE_REF (Vchar_width_table, c))) | |
614 | |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
615 extern int char_resolve_modifier_mask P_ ((int)); |
90798
ad70c7654800
(CHAR_STRING): Cast C to unsigned on calling
Kenichi Handa <handa@m17n.org>
parents:
90791
diff
changeset
|
616 extern int char_string P_ ((unsigned, unsigned char *)); |
89180
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
617 extern int string_char P_ ((const unsigned char *, |
1d29c2b108e6
(CHAR_STRING): Call char_string if C is greater than
Kenichi Handa <handa@m17n.org>
parents:
89053
diff
changeset
|
618 const unsigned char **, int *)); |
88363 | 619 |
620 extern int translate_char P_ ((Lisp_Object, int c)); | |
621 extern int char_printable_p P_ ((int c)); | |
89483 | 622 extern void parse_str_as_multibyte P_ ((const unsigned char *, int, int *, |
623 int *)); | |
88363 | 624 extern int parse_str_to_multibyte P_ ((unsigned char *, int)); |
625 extern int str_as_multibyte P_ ((unsigned char *, int, int, int *)); | |
626 extern int str_to_multibyte P_ ((unsigned char *, int, int)); | |
627 extern int str_as_unibyte P_ ((unsigned char *, int)); | |
628 extern int strwidth P_ ((unsigned char *, int)); | |
89483 | 629 extern int c_string_width P_ ((const unsigned char *, int, int, int *, int *)); |
88363 | 630 extern int lisp_string_width P_ ((Lisp_Object, int, int *, int *)); |
631 | |
632 extern Lisp_Object Vprintable_chars; | |
633 | |
634 extern Lisp_Object Qcharacterp, Qauto_fill_chars; | |
635 extern Lisp_Object Vtranslation_table_vector; | |
636 extern Lisp_Object Vchar_width_table; | |
637 extern Lisp_Object Vchar_direction_table; | |
638 extern Lisp_Object Vchar_unify_table; | |
639 | |
88545 | 640 extern Lisp_Object string_escape_byte8 P_ ((Lisp_Object)); |
641 | |
88363 | 642 /* Return a translation table of id number ID. */ |
643 #define GET_TRANSLATION_TABLE(id) \ | |
644 (XCDR(XVECTOR(Vtranslation_table_vector)->contents[(id)])) | |
645 | |
646 /* A char-table for characters which may invoke auto-filling. */ | |
647 extern Lisp_Object Vauto_fill_chars; | |
648 | |
88915
94184802d0cc
(Vchar_script_table): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
88873
diff
changeset
|
649 extern Lisp_Object Vchar_script_table; |
90402
69ac9cbd4be5
(Vscript_representative_chars): Extern it.
Kenichi Handa <handa@m17n.org>
parents:
90051
diff
changeset
|
650 extern Lisp_Object Vscript_representative_chars; |
88873
7d441bc35e9b
(TRAILING_CODE_P): New macro.
Kenichi Handa <handa@m17n.org>
parents:
88832
diff
changeset
|
651 |
88363 | 652 /* Copy LEN bytes from FROM to TO. This macro should be used only |
653 when a caller knows that LEN is short and the obvious copy loop is | |
654 faster than calling bcopy which has some overhead. Copying a | |
655 multibyte sequence of a character is the typical case. */ | |
656 | |
657 #define BCOPY_SHORT(from, to, len) \ | |
658 do { \ | |
659 int i = len; \ | |
660 unsigned char *from_p = from, *to_p = to; \ | |
661 while (i--) *to_p++ = *from_p++; \ | |
662 } while (0) | |
663 | |
664 #define DEFSYM(sym, name) \ | |
665 do { (sym) = intern ((name)); staticpro (&(sym)); } while (0) | |
666 | |
667 #endif /* EMACS_CHARACTER_H */ | |
89911 | 668 |
669 /* arch-tag: 4ef86004-2eff-4073-8cea-cfcbcf7188ac | |
670 (do not change this comment) */ |