u 0.5.0
Sign up to get free protection for your applications and to get access to all the features.
- data/README +38 -0
- data/Rakefile +64 -0
- data/ext/encoding/character/utf-8/break.c +25 -0
- data/ext/encoding/character/utf-8/data/break.h +22931 -0
- data/ext/encoding/character/utf-8/data/character-tables.h +14358 -0
- data/ext/encoding/character/utf-8/data/compose.h +1607 -0
- data/ext/encoding/character/utf-8/data/decompose.h +10926 -0
- data/ext/encoding/character/utf-8/data/generate-unicode-data.rb +1070 -0
- data/ext/encoding/character/utf-8/decompose.c +444 -0
- data/ext/encoding/character/utf-8/depend +65 -0
- data/ext/encoding/character/utf-8/extconf.rb +67 -0
- data/ext/encoding/character/utf-8/private.c +62 -0
- data/ext/encoding/character/utf-8/private.h +51 -0
- data/ext/encoding/character/utf-8/properties.c +1056 -0
- data/ext/encoding/character/utf-8/rb_includes.h +19 -0
- data/ext/encoding/character/utf-8/rb_methods.h +49 -0
- data/ext/encoding/character/utf-8/rb_private.h +52 -0
- data/ext/encoding/character/utf-8/rb_utf_aref.c +111 -0
- data/ext/encoding/character/utf-8/rb_utf_aset.c +105 -0
- data/ext/encoding/character/utf-8/rb_utf_casecmp.c +24 -0
- data/ext/encoding/character/utf-8/rb_utf_chomp.c +114 -0
- data/ext/encoding/character/utf-8/rb_utf_chop.c +44 -0
- data/ext/encoding/character/utf-8/rb_utf_collate.c +13 -0
- data/ext/encoding/character/utf-8/rb_utf_count.c +30 -0
- data/ext/encoding/character/utf-8/rb_utf_delete.c +60 -0
- data/ext/encoding/character/utf-8/rb_utf_downcase.c +13 -0
- data/ext/encoding/character/utf-8/rb_utf_each_char.c +27 -0
- data/ext/encoding/character/utf-8/rb_utf_foldcase.c +13 -0
- data/ext/encoding/character/utf-8/rb_utf_hex.c +14 -0
- data/ext/encoding/character/utf-8/rb_utf_index.c +50 -0
- data/ext/encoding/character/utf-8/rb_utf_insert.c +48 -0
- data/ext/encoding/character/utf-8/rb_utf_internal_bignum.c +332 -0
- data/ext/encoding/character/utf-8/rb_utf_internal_bignum.h +12 -0
- data/ext/encoding/character/utf-8/rb_utf_internal_tr.c +142 -0
- data/ext/encoding/character/utf-8/rb_utf_internal_tr.h +41 -0
- data/ext/encoding/character/utf-8/rb_utf_justify.c +96 -0
- data/ext/encoding/character/utf-8/rb_utf_length.c +14 -0
- data/ext/encoding/character/utf-8/rb_utf_lstrip.c +41 -0
- data/ext/encoding/character/utf-8/rb_utf_normalize.c +51 -0
- data/ext/encoding/character/utf-8/rb_utf_oct.c +14 -0
- data/ext/encoding/character/utf-8/rb_utf_reverse.c +13 -0
- data/ext/encoding/character/utf-8/rb_utf_rindex.c +88 -0
- data/ext/encoding/character/utf-8/rb_utf_rstrip.c +51 -0
- data/ext/encoding/character/utf-8/rb_utf_squeeze.c +70 -0
- data/ext/encoding/character/utf-8/rb_utf_strip.c +27 -0
- data/ext/encoding/character/utf-8/rb_utf_to_i.c +25 -0
- data/ext/encoding/character/utf-8/rb_utf_tr.c +250 -0
- data/ext/encoding/character/utf-8/rb_utf_upcase.c +13 -0
- data/ext/encoding/character/utf-8/tables.h +38 -0
- data/ext/encoding/character/utf-8/unicode.c +319 -0
- data/ext/encoding/character/utf-8/unicode.h +216 -0
- data/ext/encoding/character/utf-8/utf.c +1334 -0
- data/lib/encoding/character/utf-8.rb +201 -0
- data/lib/u.rb +16 -0
- data/lib/u/string.rb +185 -0
- data/lib/u/version.rb +5 -0
- data/test/unit/u.rb +5 -0
- data/test/unit/u/string.rb +91 -0
- metadata +174 -0
@@ -0,0 +1,67 @@
|
|
1
|
+
require 'mkmf'
|
2
|
+
|
3
|
+
def try_compiler_option(opt, &block)
|
4
|
+
checking_for "#{opt} option to compiler" do
|
5
|
+
$CFLAGS += " #{opt}" if try_compile '', opt, &block
|
6
|
+
end
|
7
|
+
end
|
8
|
+
|
9
|
+
try_compiler_option '-std=c99'
|
10
|
+
try_compiler_option '-finline-functions'
|
11
|
+
try_compiler_option '-fno-common'
|
12
|
+
try_compiler_option '-Wall'
|
13
|
+
try_compiler_option '-Waggregate-return'
|
14
|
+
try_compiler_option '-Wcast-align'
|
15
|
+
try_compiler_option '-Wextra'
|
16
|
+
try_compiler_option '-Wformat=2'
|
17
|
+
try_compiler_option '-Winit-self'
|
18
|
+
try_compiler_option '-Winline'
|
19
|
+
try_compiler_option '-Wmissing-declarations'
|
20
|
+
try_compiler_option '-Wmissing-format-attribute'
|
21
|
+
try_compiler_option '-Wmissing-include-dirs'
|
22
|
+
try_compiler_option '-Wmissing-noreturn'
|
23
|
+
try_compiler_option '-Wmissing-prototypes'
|
24
|
+
try_compiler_option '-Wnested-externs'
|
25
|
+
try_compiler_option '-Wold-style-definition'
|
26
|
+
try_compiler_option '-Wpacked'
|
27
|
+
try_compiler_option '-Wp,-D_FORTIFY_SOURCE=2'
|
28
|
+
try_compiler_option '-Wpointer-arith'
|
29
|
+
try_compiler_option '-Wsign-compare'
|
30
|
+
try_compiler_option '-Wstrict-aliasing=2'
|
31
|
+
try_compiler_option '-Wswitch-default'
|
32
|
+
try_compiler_option '-Wswitch-enum'
|
33
|
+
try_compiler_option '-Wundef'
|
34
|
+
try_compiler_option '-Wunsafe-loop-optimizations'
|
35
|
+
try_compiler_option '-Wwrite-strings'
|
36
|
+
|
37
|
+
checking_for 'GNUC visibility attribute' do
|
38
|
+
$defs.push('-DHAVE_GNUC_VISIBILITY') if try_compile <<EOC, '-Werror'
|
39
|
+
void f_hidden(void);
|
40
|
+
void __attribute__((visibility("hidden")))
|
41
|
+
f_hidden(void)
|
42
|
+
{
|
43
|
+
}
|
44
|
+
int main(void)
|
45
|
+
{
|
46
|
+
f_hidden();
|
47
|
+
return 0;
|
48
|
+
}
|
49
|
+
EOC
|
50
|
+
end
|
51
|
+
|
52
|
+
have_header 'assert.h'
|
53
|
+
have_header 'limits.h'
|
54
|
+
have_header 'locale.h'
|
55
|
+
have_header 'stdbool.h'
|
56
|
+
have_header 'stddef.h'
|
57
|
+
have_header 'stdint.h'
|
58
|
+
have_header 'stdio.h'
|
59
|
+
have_header 'stdlib.h'
|
60
|
+
have_header 'string.h'
|
61
|
+
have_header 'sys/types.h'
|
62
|
+
have_header 'wchar.h'
|
63
|
+
|
64
|
+
$INSTALLFILES ||= []
|
65
|
+
$INSTALLFILES << ['unicode.h', '$(RUBYARCHDIR)', 'lib']
|
66
|
+
|
67
|
+
create_makefile 'encoding/character/utf-8/utf8'
|
@@ -0,0 +1,62 @@
|
|
1
|
+
/*
|
2
|
+
* contents: Private functions used by the UTF-8 character-encoding library.
|
3
|
+
*
|
4
|
+
* Copyright © 2007 Nikolai Weibull <now@bitwi.se>
|
5
|
+
*/
|
6
|
+
|
7
|
+
#include <ruby.h>
|
8
|
+
#include <stdbool.h>
|
9
|
+
#include <stddef.h>
|
10
|
+
#include <stdint.h>
|
11
|
+
#include <stdlib.h>
|
12
|
+
|
13
|
+
#include "unicode.h"
|
14
|
+
|
15
|
+
#include "private.h"
|
16
|
+
|
17
|
+
/* Lookup C in the sorted TABLE using binary search. TABLE consists of N
|
18
|
+
* entries, where each entry is SIZEOF_ENTRY bytes in size and the first
|
19
|
+
* component is a unichar of size SIZEOF_CHAR. If C is found in TABLE, its
|
20
|
+
* index is stored in INDEX and true is returned. Otherwise, false is returned
|
21
|
+
* and INDEX is left untouched. */
|
22
|
+
bool
|
23
|
+
binary_search_unicode_table(const void *table, size_t n, size_t sizeof_entry, size_t sizeof_char, unichar c, int *index)
|
24
|
+
{
|
25
|
+
#define ENTRY(index) ((unichar)(*(unichar *)((const char *)table + ((index) * sizeof_entry))) & char_mask)
|
26
|
+
|
27
|
+
int begin = 0;
|
28
|
+
int end = n - 1;
|
29
|
+
int middle;
|
30
|
+
|
31
|
+
/* This is ugly, but not all tables use unichars as their lookup
|
32
|
+
* character. The casefold table, for example, uses uint16_t-sized
|
33
|
+
* characters. To only get the interesting part of our table entry
|
34
|
+
* we’ll have to mask the retrieved value. */
|
35
|
+
int char_mask = (1 << (8 * sizeof_char)) - 1;
|
36
|
+
|
37
|
+
/* Drop out early if we know for certain that C can’t be in the
|
38
|
+
* decomposition table. */
|
39
|
+
if (c < ENTRY(0) || c > ENTRY(end))
|
40
|
+
return false;
|
41
|
+
|
42
|
+
while (begin <= end) {
|
43
|
+
middle = binary_search_middle_of(begin, end);
|
44
|
+
|
45
|
+
unichar probe = ENTRY(middle);
|
46
|
+
if (c < probe)
|
47
|
+
end = middle - 1;
|
48
|
+
else if (c > probe)
|
49
|
+
begin = middle + 1;
|
50
|
+
else
|
51
|
+
break;
|
52
|
+
}
|
53
|
+
|
54
|
+
if (begin > end)
|
55
|
+
return false;
|
56
|
+
|
57
|
+
*index = middle;
|
58
|
+
|
59
|
+
return true;
|
60
|
+
|
61
|
+
#undef ENTRY
|
62
|
+
}
|
@@ -0,0 +1,51 @@
|
|
1
|
+
/*
|
2
|
+
* contents: Private Unicode related information.
|
3
|
+
*
|
4
|
+
* Copyright (C) 2004 Nikolai Weibull <source@pcppopper.org>
|
5
|
+
*/
|
6
|
+
|
7
|
+
#ifndef PRIVATE_H
|
8
|
+
#define PRIVATE_H
|
9
|
+
|
10
|
+
#define NUL '\0'
|
11
|
+
#define lengthof(ary) (sizeof(ary) / sizeof((ary)[0]))
|
12
|
+
|
13
|
+
#if defined(HAVE_GNUC_VISIBILITY)
|
14
|
+
# define HIDDEN \
|
15
|
+
__attribute__((visibility("hidden")))
|
16
|
+
#else
|
17
|
+
# define HIDDEN
|
18
|
+
#endif
|
19
|
+
|
20
|
+
#if defined(__GNUC__)
|
21
|
+
# define UNUSED(u) \
|
22
|
+
u __attribute__((__unused__))
|
23
|
+
#else
|
24
|
+
# define UNUSED(u) \
|
25
|
+
u
|
26
|
+
#endif
|
27
|
+
|
28
|
+
#define binary_search_middle_of(begin, end) \
|
29
|
+
(((unsigned)((begin) + (end))) >> 1)
|
30
|
+
|
31
|
+
#define unicode_table_lookup(table, c, index) \
|
32
|
+
binary_search_unicode_table(table, lengthof(table), sizeof((table)[0]), sizeof((table)[0].ch), c, index)
|
33
|
+
|
34
|
+
bool binary_search_unicode_table(const void *table, size_t n, size_t sizeof_entry, size_t sizeof_char, unichar c, int *index) HIDDEN;
|
35
|
+
|
36
|
+
#define SPLIT_UNICODE_TABLE_LOOKUP_PAGE(data, part, page, c) \
|
37
|
+
((part[page] >= UNICODE_MAX_TABLE_INDEX) \
|
38
|
+
? (part[page] - UNICODE_MAX_TABLE_INDEX) \
|
39
|
+
: (data[part[page]][(c) & 0xff]))
|
40
|
+
|
41
|
+
#define SPLIT_UNICODE_TABLE_LOOKUP(data, part1, part2, c, fallback) \
|
42
|
+
(((c) <= UNICODE_LAST_CHAR_PART1) \
|
43
|
+
? SPLIT_UNICODE_TABLE_LOOKUP_PAGE(data, part1, (c) >> 8, c) \
|
44
|
+
: (((c) >= UNICODE_FIRST_CHAR_PART2 && (c) <= UNICODE_LAST_CHAR) \
|
45
|
+
? SPLIT_UNICODE_TABLE_LOOKUP_PAGE(data, part2, ((c) - UNICODE_FIRST_CHAR_PART2) >> 8, c) \
|
46
|
+
: (fallback)))
|
47
|
+
|
48
|
+
unichar *_utf_normalize_wc(const char *str, size_t max_len, bool use_len,
|
49
|
+
NormalizeMode mode) HIDDEN;
|
50
|
+
|
51
|
+
#endif /* PRIVATE_H */
|
@@ -0,0 +1,1056 @@
|
|
1
|
+
/*
|
2
|
+
* contents: Unicode character properties.
|
3
|
+
*
|
4
|
+
* Copyright (C) 2004 Nikolai Weibull <source@pcppopper.org>
|
5
|
+
*/
|
6
|
+
|
7
|
+
#include <ruby.h>
|
8
|
+
#include <assert.h>
|
9
|
+
#include <locale.h>
|
10
|
+
#include <stdbool.h>
|
11
|
+
#include <stddef.h>
|
12
|
+
#include <stdint.h>
|
13
|
+
#include <string.h>
|
14
|
+
#include "unicode.h"
|
15
|
+
#include "private.h"
|
16
|
+
#include "data/character-tables.h"
|
17
|
+
|
18
|
+
|
19
|
+
#define COMBINING_DOT_ABOVE ((unichar)0x0307)
|
20
|
+
#define LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE ((unichar)0x0130)
|
21
|
+
#define COMBINING_GREEK_YPOGEGRAMMENI ((unichar)0x0345)
|
22
|
+
#define GREEK_CAPITAL_LETTER_IOTA ((unichar)0x0399)
|
23
|
+
#define LATIN_SMALL_LETTER_I ((unichar)0x0069)
|
24
|
+
#define LATIN_SMALL_LETTER_DOTLESS_I ((unichar)0x0131)
|
25
|
+
#define LATIN_CAPITAL_LETTER_I_WITH_GRAVE ((unichar)0x00cc)
|
26
|
+
#define LATIN_CAPITAL_LETTER_I_WITH_ACUTE ((unichar)0x00cd)
|
27
|
+
#define LATIN_CAPITAL_LETTER_I_WITH_TILDE ((unichar)0x0128)
|
28
|
+
#define LATIN_CAPITAL_LETTER_I_WITH_OGONEK ((unichar)0x012e)
|
29
|
+
#define COMBINING_GRAVE_ACCENT ((unichar)0x0300)
|
30
|
+
#define COMBINING_ACUTE_ACCENT ((unichar)0x0301)
|
31
|
+
#define COMBINING_TILDE ((unichar)0x0303)
|
32
|
+
#define GREEK_CAPITAL_LETTER_SIGMA ((unichar)0x03a3)
|
33
|
+
#define GREEK_SMALL_LETTER_SIGMA ((unichar)0x03c3)
|
34
|
+
#define GREEK_SMALL_LETTER_FINAL_SIGMA ((unichar)0x03c2)
|
35
|
+
|
36
|
+
#define OFFSET_IF(buf, len) (((buf) != NULL) ? (buf) + (len) : NULL)
|
37
|
+
|
38
|
+
/* {{{1
|
39
|
+
* Macros for accessing the Unicode character attribute table.
|
40
|
+
*
|
41
|
+
* TODO: Turn these macros into full-fledged functions, as this is rather silly
|
42
|
+
* when we have ‹inline› in C99.
|
43
|
+
*/
|
44
|
+
#define ATTR_TABLE(page) \
|
45
|
+
(((page) <= UNICODE_LAST_PAGE_PART1) \
|
46
|
+
? attr_table_part1[page] \
|
47
|
+
: attr_table_part2[(page) - 0xe00])
|
48
|
+
|
49
|
+
#define ATTTABLE(page, char) \
|
50
|
+
((ATTR_TABLE(page) == UNICODE_MAX_TABLE_INDEX) \
|
51
|
+
? 0 : (attr_data[ATTR_TABLE(page)][char]))
|
52
|
+
|
53
|
+
|
54
|
+
/* {{{1
|
55
|
+
* Internal function used for figuring out the type of a given character.
|
56
|
+
*/
|
57
|
+
static inline int
|
58
|
+
s_type(unichar c)
|
59
|
+
{
|
60
|
+
const int16_t *table;
|
61
|
+
unsigned int page;
|
62
|
+
|
63
|
+
if (c <= UNICODE_LAST_CHAR_PART1) {
|
64
|
+
page = c >> 8;
|
65
|
+
table = type_table_part1;
|
66
|
+
} else if (c >= UNICODE_FIRST_CHAR_PART2 && c <= UNICODE_LAST_CHAR) {
|
67
|
+
page = (c - UNICODE_FIRST_CHAR_PART2) >> 8;
|
68
|
+
table = type_table_part2;
|
69
|
+
} else {
|
70
|
+
return UNICODE_UNASSIGNED;
|
71
|
+
}
|
72
|
+
|
73
|
+
if (table[page] >= UNICODE_MAX_TABLE_INDEX)
|
74
|
+
return table[page] - UNICODE_MAX_TABLE_INDEX;
|
75
|
+
else
|
76
|
+
return type_data[table[page]][c & 0xff];
|
77
|
+
}
|
78
|
+
|
79
|
+
|
80
|
+
/* {{{1
|
81
|
+
* Bit-fiddling macros for testing the class of a type.
|
82
|
+
*/
|
83
|
+
#define IS(type, class) (((unsigned int)1 << (type)) & (class))
|
84
|
+
#define OR(type, rest) (((unsigned int)1 << (type)) | (rest))
|
85
|
+
|
86
|
+
|
87
|
+
/* {{{1
|
88
|
+
* Internal function used to check if the given type represents a digit type.
|
89
|
+
*/
|
90
|
+
static inline bool
|
91
|
+
s_isdigit(int type)
|
92
|
+
{
|
93
|
+
return IS(type,
|
94
|
+
OR(UNICODE_DECIMAL_NUMBER,
|
95
|
+
OR(UNICODE_LETTER_NUMBER,
|
96
|
+
OR(UNICODE_OTHER_NUMBER, 0))));
|
97
|
+
}
|
98
|
+
|
99
|
+
|
100
|
+
/* {{{1
|
101
|
+
* Internal function used to check if the given type represents an alphabetic
|
102
|
+
* type.
|
103
|
+
*/
|
104
|
+
static inline bool
|
105
|
+
s_isalpha(int type)
|
106
|
+
{
|
107
|
+
return IS(type,
|
108
|
+
OR(UNICODE_LOWERCASE_LETTER,
|
109
|
+
OR(UNICODE_UPPERCASE_LETTER,
|
110
|
+
OR(UNICODE_TITLECASE_LETTER,
|
111
|
+
OR(UNICODE_MODIFIER_LETTER,
|
112
|
+
OR(UNICODE_OTHER_LETTER, 0))))));
|
113
|
+
}
|
114
|
+
|
115
|
+
|
116
|
+
/* {{{1
|
117
|
+
* Internal function used to check if the given type represents a mark type.
|
118
|
+
*/
|
119
|
+
static inline bool
|
120
|
+
s_ismark(int type)
|
121
|
+
{
|
122
|
+
return IS(type,
|
123
|
+
OR(UNICODE_NON_SPACING_MARK,
|
124
|
+
OR(UNICODE_COMBINING_MARK,
|
125
|
+
OR(UNICODE_ENCLOSING_MARK, 0))));
|
126
|
+
}
|
127
|
+
|
128
|
+
|
129
|
+
/* {{{1
|
130
|
+
* Determine whether ‘c’ is an alphanumeric, such as A, B, C, 0, 1, or 2.
|
131
|
+
*/
|
132
|
+
bool
|
133
|
+
unichar_isalnum(unichar c)
|
134
|
+
{
|
135
|
+
int type = s_type(c);
|
136
|
+
|
137
|
+
return s_isdigit(type) || s_isalpha(type);
|
138
|
+
}
|
139
|
+
|
140
|
+
|
141
|
+
/* {{{1
|
142
|
+
* Determine whether ‘c’ is an alphabetic (i.e. a letter), such as A, B, or C.
|
143
|
+
*/
|
144
|
+
bool
|
145
|
+
unichar_isalpha(unichar c)
|
146
|
+
{
|
147
|
+
return s_isalpha(s_type(c));
|
148
|
+
}
|
149
|
+
|
150
|
+
|
151
|
+
/* {{{1
|
152
|
+
* Determine whether ‘c’ is a control character, such as ‹NUL›.
|
153
|
+
*/
|
154
|
+
bool
|
155
|
+
unichar_iscntrl(unichar c)
|
156
|
+
{
|
157
|
+
return s_type(c) == UNICODE_CONTROL;
|
158
|
+
}
|
159
|
+
|
160
|
+
|
161
|
+
/* {{{1
|
162
|
+
* Determine whether ‘c’ is a digit, such as 0, 1, or 2.
|
163
|
+
*/
|
164
|
+
bool
|
165
|
+
unichar_isdigit(unichar c)
|
166
|
+
{
|
167
|
+
return s_type(c) == UNICODE_DECIMAL_NUMBER;
|
168
|
+
}
|
169
|
+
|
170
|
+
|
171
|
+
/* {{{1
|
172
|
+
* Determine whether ‘c’ is printable and not a space or control character such
|
173
|
+
* as tab or <NUL›, such as A, B, or C.
|
174
|
+
*/
|
175
|
+
bool
|
176
|
+
unichar_isgraph(unichar c)
|
177
|
+
{
|
178
|
+
return !IS(s_type(c),
|
179
|
+
OR(UNICODE_CONTROL,
|
180
|
+
OR(UNICODE_FORMAT,
|
181
|
+
OR(UNICODE_UNASSIGNED,
|
182
|
+
OR(UNICODE_PRIVATE_USE,
|
183
|
+
OR(UNICODE_SURROGATE,
|
184
|
+
OR(UNICODE_SPACE_SEPARATOR, 0)))))));
|
185
|
+
}
|
186
|
+
|
187
|
+
|
188
|
+
/* {{{1
|
189
|
+
* Determine whether ‘c’ is a lowercase letter, such as a, b, or c.
|
190
|
+
*/
|
191
|
+
bool
|
192
|
+
unichar_islower(unichar c)
|
193
|
+
{
|
194
|
+
return s_type(c) == UNICODE_LOWERCASE_LETTER;
|
195
|
+
}
|
196
|
+
|
197
|
+
|
198
|
+
/* {{{1
|
199
|
+
* Determine whether ‘c’ is printable, which works the same as
|
200
|
+
* unichar_isgraph(), except that space characters are also printable.
|
201
|
+
*/
|
202
|
+
bool
|
203
|
+
unichar_isprint(unichar c)
|
204
|
+
{
|
205
|
+
return !IS(s_type(c),
|
206
|
+
OR(UNICODE_CONTROL,
|
207
|
+
OR(UNICODE_FORMAT,
|
208
|
+
OR(UNICODE_UNASSIGNED,
|
209
|
+
OR(UNICODE_PRIVATE_USE,
|
210
|
+
OR(UNICODE_SURROGATE, 0))))));
|
211
|
+
}
|
212
|
+
|
213
|
+
|
214
|
+
/* {{{1
|
215
|
+
* Determine whether ‘c’ is some form of punctuation or other symbol.
|
216
|
+
*/
|
217
|
+
bool
|
218
|
+
unichar_ispunct(unichar c)
|
219
|
+
{
|
220
|
+
return IS(s_type(c),
|
221
|
+
OR(UNICODE_CONNECT_PUNCTUATION,
|
222
|
+
OR(UNICODE_DASH_PUNCTUATION,
|
223
|
+
OR(UNICODE_OPEN_PUNCTUATION,
|
224
|
+
OR(UNICODE_CLOSE_PUNCTUATION,
|
225
|
+
OR(UNICODE_INITIAL_PUNCTUATION,
|
226
|
+
OR(UNICODE_FINAL_PUNCTUATION,
|
227
|
+
OR(UNICODE_OTHER_PUNCTUATION,
|
228
|
+
OR(UNICODE_MODIFIER_SYMBOL,
|
229
|
+
OR(UNICODE_MATH_SYMBOL,
|
230
|
+
OR(UNICODE_CURRENCY_SYMBOL,
|
231
|
+
OR(UNICODE_OTHER_SYMBOL, 0)))))))))))) ? true : false;
|
232
|
+
}
|
233
|
+
|
234
|
+
|
235
|
+
/* {{{1
|
236
|
+
* Determine whether ‘c’ is some form of whitespace, such as space, tab or a
|
237
|
+
* line separator (newline, carriage return, etc.).
|
238
|
+
*/
|
239
|
+
bool
|
240
|
+
unichar_isspace(unichar c)
|
241
|
+
{
|
242
|
+
switch (c) {
|
243
|
+
case '\t':
|
244
|
+
case '\n':
|
245
|
+
case '\r':
|
246
|
+
case '\f':
|
247
|
+
return true;
|
248
|
+
default:
|
249
|
+
return IS(s_type(c),
|
250
|
+
OR(UNICODE_SPACE_SEPARATOR,
|
251
|
+
OR(UNICODE_LINE_SEPARATOR,
|
252
|
+
OR(UNICODE_PARAGRAPH_SEPARATOR, 0)))) ? true : false;
|
253
|
+
}
|
254
|
+
}
|
255
|
+
|
256
|
+
|
257
|
+
/* {{{1
|
258
|
+
* Determine whether ‘c’ is an uppeercase letter, such as A, B, or C
|
259
|
+
*/
|
260
|
+
bool
|
261
|
+
unichar_isupper(unichar c)
|
262
|
+
{
|
263
|
+
return s_type(c) == UNICODE_UPPERCASE_LETTER;
|
264
|
+
}
|
265
|
+
|
266
|
+
|
267
|
+
/* {{{1
|
268
|
+
* Determine whether ‘c’ is a titlecase letter, such as the slavic digraph DZ,
|
269
|
+
* which at the beginning of a word is written as Dz, where only the initial D
|
270
|
+
* is capitalized. (Complicated huh?)
|
271
|
+
*/
|
272
|
+
bool
|
273
|
+
unichar_istitle(unichar c)
|
274
|
+
{
|
275
|
+
/* TODO: binary search helpful? */
|
276
|
+
for (size_t i = 0; i < lengthof(title_table); i++)
|
277
|
+
if (title_table[i][0] == c)
|
278
|
+
return true;
|
279
|
+
|
280
|
+
return false;
|
281
|
+
}
|
282
|
+
|
283
|
+
|
284
|
+
/* {{{1
|
285
|
+
* Determine whether ‘c’ is a new-line.
|
286
|
+
*/
|
287
|
+
#define UNICHAR_NEXT_LINE ((unichar)0x0085)
|
288
|
+
#define UNICHAR_LINE_SEPARATOR ((unichar)0x2028)
|
289
|
+
#define UNICHAR_PARAGRAPH_SEPARATOR ((unichar)0x2029)
|
290
|
+
|
291
|
+
bool
|
292
|
+
unichar_isnewline(unichar c)
|
293
|
+
{
|
294
|
+
switch (c) {
|
295
|
+
case '\n': case '\f': case '\r': case UNICHAR_NEXT_LINE:
|
296
|
+
case UNICHAR_LINE_SEPARATOR: case UNICHAR_PARAGRAPH_SEPARATOR:
|
297
|
+
return true;
|
298
|
+
default:
|
299
|
+
return false;
|
300
|
+
}
|
301
|
+
}
|
302
|
+
|
303
|
+
/* {{{1
|
304
|
+
* Determine whether ‘c’ is a hexadecimal digit, such as 0, 1, ..., 9, a, b,
|
305
|
+
* ..., f, or A, B, ..., F.
|
306
|
+
*/
|
307
|
+
#define UNICHAR_FULLWIDTH_A 0xff21
|
308
|
+
#define UNICHAR_FULLWIDTH_F 0xff26
|
309
|
+
#define UNICHAR_FULLWIDTH_a 0xff41
|
310
|
+
#define UNICHAR_FULLWIDTH_f 0xff46
|
311
|
+
bool
|
312
|
+
unichar_isxdigit(unichar c)
|
313
|
+
{
|
314
|
+
return ((c >= 'a' && c <= 'f') ||
|
315
|
+
(c >= 'A' && c <= 'F') ||
|
316
|
+
(c >= UNICHAR_FULLWIDTH_a && c <= UNICHAR_FULLWIDTH_f) ||
|
317
|
+
(c >= UNICHAR_FULLWIDTH_A && c <= UNICHAR_FULLWIDTH_F) ||
|
318
|
+
(s_type(c) == UNICODE_DECIMAL_NUMBER));
|
319
|
+
// s_isdigit(s_type(c)));
|
320
|
+
}
|
321
|
+
|
322
|
+
|
323
|
+
/* {{{1
|
324
|
+
* Determine whether code point ‘c’ has been assigned a code value.
|
325
|
+
*/
|
326
|
+
bool
|
327
|
+
unichar_isassigned(unichar c)
|
328
|
+
{
|
329
|
+
return s_type(c) != UNICODE_UNASSIGNED;
|
330
|
+
}
|
331
|
+
|
332
|
+
|
333
|
+
/* {{{1
|
334
|
+
* Determine whether ‘c’ is a wide character, thus is typically rendered in a
|
335
|
+
* double-width cell on a terminal.
|
336
|
+
*/
|
337
|
+
bool
|
338
|
+
unichar_iswide(unichar c)
|
339
|
+
{
|
340
|
+
if (c < 0x1100)
|
341
|
+
return false;
|
342
|
+
|
343
|
+
return (c <= 0x115f || /* Hangul Jamo init. consonants */
|
344
|
+
c == 0x2329 || c == 0x232a || /* angle brackets */
|
345
|
+
(c >= 0x2e80 && c <= 0xa4cf && /* CJK ... Yi */
|
346
|
+
(c < 0x302a || c > 0x302f) &&
|
347
|
+
c != 0x303f && c != 0x3099 && c != 0x309a) ||
|
348
|
+
(c >= 0xac00 && c <= 0xd7a3) || /* Hangul syllables */
|
349
|
+
(c >= 0xf900 && c <= 0xfaff) || /* CJK comp. graphs */
|
350
|
+
(c >= 0xfe30 && c <= 0xfe6f) || /* CJK comp. forms */
|
351
|
+
(c >= 0xff00 && c <= 0xff60) || /* fullwidth forms */
|
352
|
+
(c >= 0xffe0 && c <= 0xffe6) || /* -"- */
|
353
|
+
(c >= 0x20000 && c <= 0x2fffd) || /* CJK extra stuff */
|
354
|
+
(c >= 0x30000 && c <= 0x3fffd)); /* -"- */
|
355
|
+
}
|
356
|
+
|
357
|
+
|
358
|
+
/* {{{1
|
359
|
+
* Convert ‘c’ to its uppercase representation (if any).
|
360
|
+
*/
|
361
|
+
static unichar
|
362
|
+
special_case_table_lookup(unichar c)
|
363
|
+
{
|
364
|
+
unichar tv = ATTTABLE(c >> 8, c & 0xff);
|
365
|
+
|
366
|
+
if (tv >= UNICODE_SPECIAL_CASE_TABLE_START)
|
367
|
+
tv = utf_char(special_case_table +
|
368
|
+
tv - UNICODE_SPECIAL_CASE_TABLE_START);
|
369
|
+
|
370
|
+
if (tv == '\0')
|
371
|
+
return c;
|
372
|
+
|
373
|
+
return tv;
|
374
|
+
}
|
375
|
+
|
376
|
+
static unichar
|
377
|
+
titlecase_table_lookup(unichar c, bool want_upper)
|
378
|
+
{
|
379
|
+
for (size_t i = 0; i < lengthof(title_table); i++)
|
380
|
+
if (title_table[i][0] == c)
|
381
|
+
return title_table[i][want_upper ? 1 : 2];
|
382
|
+
|
383
|
+
return c;
|
384
|
+
}
|
385
|
+
|
386
|
+
unichar
|
387
|
+
unichar_toupper(unichar c)
|
388
|
+
{
|
389
|
+
int type = s_type(c);
|
390
|
+
|
391
|
+
if (type == UNICODE_LOWERCASE_LETTER)
|
392
|
+
return special_case_table_lookup(c);
|
393
|
+
|
394
|
+
if (type == UNICODE_TITLECASE_LETTER)
|
395
|
+
return titlecase_table_lookup(c, true);
|
396
|
+
|
397
|
+
return c;
|
398
|
+
}
|
399
|
+
|
400
|
+
|
401
|
+
/* {{{1
|
402
|
+
* Convert ‘c’ to its lowercase representation (if any).
|
403
|
+
*/
|
404
|
+
unichar
|
405
|
+
unichar_tolower(unichar c)
|
406
|
+
{
|
407
|
+
int type = s_type(c);
|
408
|
+
|
409
|
+
if (type == UNICODE_UPPERCASE_LETTER)
|
410
|
+
return special_case_table_lookup(c);
|
411
|
+
|
412
|
+
if (type == UNICODE_TITLECASE_LETTER)
|
413
|
+
return titlecase_table_lookup(c, false);
|
414
|
+
|
415
|
+
return c;
|
416
|
+
}
|
417
|
+
|
418
|
+
|
419
|
+
/* {{{1
|
420
|
+
* Convert ‘c’ to its titlecase representation (if any).
|
421
|
+
*/
|
422
|
+
unichar
|
423
|
+
unichar_totitle(unichar c)
|
424
|
+
{
|
425
|
+
for (size_t i = 0; i < lengthof(title_table); i++)
|
426
|
+
if (title_table[i][0] == c ||
|
427
|
+
title_table[i][1] == c ||
|
428
|
+
title_table[i][2] == c)
|
429
|
+
return title_table[i][0];
|
430
|
+
|
431
|
+
if (s_type(c) == UNICODE_LOWERCASE_LETTER)
|
432
|
+
return unichar_toupper(c);
|
433
|
+
|
434
|
+
return c;
|
435
|
+
}
|
436
|
+
|
437
|
+
|
438
|
+
/* {{{1
|
439
|
+
* Return the numeric value of ‘c’ if it's a decimal digit, or -1 if not.
|
440
|
+
*/
|
441
|
+
int
|
442
|
+
unichar_digit_value(unichar c)
|
443
|
+
{
|
444
|
+
if (s_type(c) == UNICODE_DECIMAL_NUMBER)
|
445
|
+
return ATTTABLE(c >> 8, c & 0xff);
|
446
|
+
|
447
|
+
return -1;
|
448
|
+
}
|
449
|
+
|
450
|
+
|
451
|
+
/* {{{1
|
452
|
+
* Return the numeric value of ‘c’ if it's a hexadecimal digit, or -1 if not.
|
453
|
+
*/
|
454
|
+
int
|
455
|
+
unichar_xdigit_value(unichar c)
|
456
|
+
{
|
457
|
+
if (c >= 'a' && c <= 'f')
|
458
|
+
return c - 'a' + 10;
|
459
|
+
else if (c >= 'A' && c <= 'F')
|
460
|
+
return c - 'A' + 10;
|
461
|
+
else if (c >= UNICHAR_FULLWIDTH_a && c <= UNICHAR_FULLWIDTH_f)
|
462
|
+
return c - UNICHAR_FULLWIDTH_a + 10;
|
463
|
+
else if (c >= UNICHAR_FULLWIDTH_A && c <= UNICHAR_FULLWIDTH_F)
|
464
|
+
return c - UNICHAR_FULLWIDTH_A + 10;
|
465
|
+
else
|
466
|
+
return unichar_digit_value(c);
|
467
|
+
}
|
468
|
+
|
469
|
+
|
470
|
+
/* {{{1
|
471
|
+
* Determine the Unicode character type of ‘c’.
|
472
|
+
*/
|
473
|
+
UnicodeType
|
474
|
+
unichar_type(unichar c)
|
475
|
+
{
|
476
|
+
return s_type(c);
|
477
|
+
}
|
478
|
+
|
479
|
+
|
480
|
+
/* {{{1
|
481
|
+
* LocaleType: This ‹enum› is used for dealing with different locales for
|
482
|
+
* turning strings into uppercase or lowercase.
|
483
|
+
*/
|
484
|
+
typedef enum {
|
485
|
+
LOCALE_NORMAL,
|
486
|
+
LOCALE_TURKIC,
|
487
|
+
LOCALE_LITHUANIAN
|
488
|
+
} LocaleType;
|
489
|
+
|
490
|
+
|
491
|
+
/* {{{1
|
492
|
+
* Retrieve the locale type from the environment (LC_CTYPE).
|
493
|
+
*/
|
494
|
+
static LocaleType
|
495
|
+
get_locale_type(void)
|
496
|
+
{
|
497
|
+
const char *locale = setlocale(LC_CTYPE, NULL);
|
498
|
+
|
499
|
+
if ((locale[0] == 'a' && locale[1] == 'z') ||
|
500
|
+
(locale[0] == 't' && locale[1] == 'r'))
|
501
|
+
return LOCALE_TURKIC;
|
502
|
+
|
503
|
+
if (locale[0] == 'l' && locale[1] == 't')
|
504
|
+
return LOCALE_LITHUANIAN;
|
505
|
+
|
506
|
+
return LOCALE_NORMAL;
|
507
|
+
}
|
508
|
+
|
509
|
+
|
510
|
+
/* {{{1
|
511
|
+
* Put character marks found in ‘p_inout’ into itself. If ‘remove_dot’ is
|
512
|
+
* true, remove the dot over an uppercase I for a turkish locale.
|
513
|
+
*/
|
514
|
+
static size_t
|
515
|
+
output_marks(const char **p_inout, char *buf, bool remove_dot)
|
516
|
+
{
|
517
|
+
size_t len = 0;
|
518
|
+
const char *p = *p_inout;
|
519
|
+
|
520
|
+
for ( ; *p != '\0'; p = utf_next(p)) {
|
521
|
+
unichar c = utf_char(p);
|
522
|
+
|
523
|
+
if (!s_ismark(s_type(c)))
|
524
|
+
break;
|
525
|
+
|
526
|
+
if (!remove_dot || c != COMBINING_DOT_ABOVE)
|
527
|
+
len += unichar_to_utf(c, (buf != NULL) ? buf + len : NULL);
|
528
|
+
}
|
529
|
+
|
530
|
+
*p_inout = p;
|
531
|
+
|
532
|
+
return len;
|
533
|
+
}
|
534
|
+
|
535
|
+
/* {{{1
|
536
|
+
* Output titlecases where appropriate.
|
537
|
+
*/
|
538
|
+
static size_t
|
539
|
+
output_special_case(char *buf, int offset, int type, bool upper)
|
540
|
+
{
|
541
|
+
const char *p = special_case_table + offset;
|
542
|
+
|
543
|
+
if (type != UNICODE_TITLECASE_LETTER)
|
544
|
+
p = utf_next(p);
|
545
|
+
|
546
|
+
if (upper)
|
547
|
+
p += utf_byte_length(p) + 1;
|
548
|
+
|
549
|
+
size_t len = utf_byte_length(p);
|
550
|
+
|
551
|
+
if (buf != NULL)
|
552
|
+
memcpy(buf, p, len);
|
553
|
+
|
554
|
+
return len;
|
555
|
+
}
|
556
|
+
|
557
|
+
/* {{{1
|
558
|
+
* Do uppercasing of ‘p’ for Lithuanian locales.
|
559
|
+
*/
|
560
|
+
static size_t
|
561
|
+
remove_all_combining_dot_above(unichar c, char *buf)
|
562
|
+
{
|
563
|
+
size_t decomp_len;
|
564
|
+
unichar *decomp = unicode_canonical_decomposition(c, &decomp_len);
|
565
|
+
|
566
|
+
size_t len = 0;
|
567
|
+
for (size_t i = 0; i < decomp_len; i++)
|
568
|
+
if (decomp[i] != COMBINING_DOT_ABOVE)
|
569
|
+
len += unichar_to_utf(unichar_toupper(decomp[i]),
|
570
|
+
OFFSET_IF(buf, len));
|
571
|
+
|
572
|
+
free(decomp);
|
573
|
+
|
574
|
+
return len;
|
575
|
+
}
|
576
|
+
|
577
|
+
static size_t
|
578
|
+
real_toupper_lithuanian(const char **p, unichar c, int type, char *buf,
|
579
|
+
bool *was_i)
|
580
|
+
{
|
581
|
+
if (c == 'i') {
|
582
|
+
*was_i = true;
|
583
|
+
return 0;
|
584
|
+
}
|
585
|
+
|
586
|
+
if (*was_i) {
|
587
|
+
size_t len = remove_all_combining_dot_above(c, buf);
|
588
|
+
return len + output_marks(p, OFFSET_IF(buf, len), true);
|
589
|
+
}
|
590
|
+
|
591
|
+
if (!s_ismark(type))
|
592
|
+
*was_i = false;
|
593
|
+
|
594
|
+
return 0;
|
595
|
+
}
|
596
|
+
|
597
|
+
/* {{{1
|
598
|
+
* Do real upcasing. */
|
599
|
+
static inline size_t
|
600
|
+
real_do_toupper(unichar c, int type, char *buf)
|
601
|
+
{
|
602
|
+
bool upper = (type != UNICODE_LOWERCASE_LETTER);
|
603
|
+
unichar tv = ATTTABLE(c >> 8, c & 0xff);
|
604
|
+
|
605
|
+
if (tv >= UNICODE_SPECIAL_CASE_TABLE_START)
|
606
|
+
return output_special_case(buf,
|
607
|
+
tv - UNICODE_SPECIAL_CASE_TABLE_START,
|
608
|
+
type, upper);
|
609
|
+
|
610
|
+
/* TODO: this should really use titlecase_table_lookup somehow. */
|
611
|
+
if (type == UNICODE_TITLECASE_LETTER)
|
612
|
+
for (size_t i = 0; i < lengthof(title_table); i++)
|
613
|
+
if (title_table[i][0] == c)
|
614
|
+
return unichar_to_utf(title_table[i][1], buf);
|
615
|
+
|
616
|
+
return unichar_to_utf(tv != '\0' ? tv : c, buf);
|
617
|
+
}
|
618
|
+
|
619
|
+
/* {{{1
|
620
|
+
* Do real uppercasing of ‘str’.
|
621
|
+
*/
|
622
|
+
static size_t
|
623
|
+
real_toupper_one(const char **p, const char *prev, char *buf,
|
624
|
+
LocaleType locale_type, bool *was_i)
|
625
|
+
{
|
626
|
+
unichar c = utf_char(prev);
|
627
|
+
int type = s_type(c);
|
628
|
+
|
629
|
+
if (locale_type == LOCALE_LITHUANIAN) {
|
630
|
+
size_t len = real_toupper_lithuanian(p, c, type, buf, was_i);
|
631
|
+
if (len > 0)
|
632
|
+
return len;
|
633
|
+
}
|
634
|
+
|
635
|
+
if (locale_type == LOCALE_TURKIC && c == 'i')
|
636
|
+
return unichar_to_utf(LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE,
|
637
|
+
buf);
|
638
|
+
|
639
|
+
if (c == COMBINING_GREEK_YPOGEGRAMMENI) {
|
640
|
+
/* Nasty, need to move it after other combining marks...this
|
641
|
+
* would go away if we normalized first. */
|
642
|
+
/* TODO: don’t we need to make sure we don’t go beyond the end
|
643
|
+
* of ‘p’? */
|
644
|
+
size_t len = output_marks(p, buf, false);
|
645
|
+
return len + unichar_to_utf(GREEK_CAPITAL_LETTER_IOTA,
|
646
|
+
OFFSET_IF(buf, len));
|
647
|
+
}
|
648
|
+
|
649
|
+
if (IS(type, OR(UNICODE_LOWERCASE_LETTER,
|
650
|
+
OR(UNICODE_TITLECASE_LETTER, 0))))
|
651
|
+
return real_do_toupper(c, type, buf);
|
652
|
+
|
653
|
+
size_t len = s_utf_skip_lengths[*(const unsigned char *)prev];
|
654
|
+
|
655
|
+
if (buf != NULL)
|
656
|
+
memcpy(buf, prev, len);
|
657
|
+
|
658
|
+
return len;
|
659
|
+
}
|
660
|
+
|
661
|
+
static size_t
|
662
|
+
real_toupper(const char *str, size_t max, bool use_max, char *buf,
|
663
|
+
LocaleType locale_type)
|
664
|
+
{
|
665
|
+
const char *p = str;
|
666
|
+
size_t len = 0;
|
667
|
+
bool p_was_i = false;
|
668
|
+
|
669
|
+
while ((!use_max || p < str + max) && *p != '\0') {
|
670
|
+
const char *prev = p;
|
671
|
+
p = utf_next(p);
|
672
|
+
|
673
|
+
len += real_toupper_one(&p, prev, OFFSET_IF(buf, len),
|
674
|
+
locale_type, &p_was_i);
|
675
|
+
}
|
676
|
+
|
677
|
+
return len;
|
678
|
+
}
|
679
|
+
|
680
|
+
/* {{{1
|
681
|
+
* Wrapper around real_toupper() for dealing with memory allocation and such.
|
682
|
+
*/
|
683
|
+
static char *
|
684
|
+
utf_upcase_impl(const char *str, size_t max, bool use_max)
|
685
|
+
{
|
686
|
+
assert(str != NULL);
|
687
|
+
|
688
|
+
LocaleType locale_type = get_locale_type();
|
689
|
+
|
690
|
+
size_t len = real_toupper(str, max, use_max, NULL, locale_type);
|
691
|
+
char *result = ALLOC_N(char, len + 1);
|
692
|
+
real_toupper(str, max, use_max, result, locale_type);
|
693
|
+
result[len] = '\0';
|
694
|
+
|
695
|
+
return result;
|
696
|
+
}
|
697
|
+
|
698
|
+
|
699
|
+
/* {{{1
|
700
|
+
* Convert all characters in ‘str’ to their uppercase representation if
|
701
|
+
* applicable. Returns the freshly allocated representation.
|
702
|
+
*/
|
703
|
+
char *
|
704
|
+
utf_upcase(const char *str)
|
705
|
+
{
|
706
|
+
return utf_upcase_impl(str, 0, false);
|
707
|
+
}
|
708
|
+
|
709
|
+
|
710
|
+
/* {{{1
|
711
|
+
* Convert all characters in ‘str’ to their uppercase representation if
|
712
|
+
* applicable. Returns the freshly allocated representation. Do this for at
|
713
|
+
* most ‘len˚ bytes from ‘str’.
|
714
|
+
*/
|
715
|
+
char *
|
716
|
+
utf_upcase_n(const char *str, size_t len)
|
717
|
+
{
|
718
|
+
return utf_upcase_impl(str, len, true);
|
719
|
+
}
|
720
|
+
|
721
|
+
|
722
|
+
/* {{{1
|
723
|
+
* Traverse the string checking for characters with combining class == 230
|
724
|
+
* until a base character is found.
|
725
|
+
*/
|
726
|
+
static bool
|
727
|
+
has_more_above(const char *str)
|
728
|
+
{
|
729
|
+
for (const char *p = str; *p != '\0'; p = utf_next(p)) {
|
730
|
+
int c_class = unichar_combining_class(utf_char(p));
|
731
|
+
|
732
|
+
if (c_class == 230)
|
733
|
+
return true;
|
734
|
+
|
735
|
+
if (c_class == 0)
|
736
|
+
return false;
|
737
|
+
}
|
738
|
+
|
739
|
+
return false;
|
740
|
+
}
|
741
|
+
|
742
|
+
static inline size_t
|
743
|
+
real_do_tolower(unichar c, int type, char *buf)
|
744
|
+
{
|
745
|
+
unichar tv = ATTTABLE(c >> 8, c & 0xff);
|
746
|
+
|
747
|
+
if (tv >= UNICODE_SPECIAL_CASE_TABLE_START)
|
748
|
+
return output_special_case(buf,
|
749
|
+
tv - UNICODE_SPECIAL_CASE_TABLE_START,
|
750
|
+
type, false);
|
751
|
+
|
752
|
+
/* TODO: this should really use titlecase_table_lookup somehow. */
|
753
|
+
if (type == UNICODE_TITLECASE_LETTER)
|
754
|
+
for (size_t i = 0; i < lengthof(title_table); i++)
|
755
|
+
if (title_table[i][0] == c)
|
756
|
+
return unichar_to_utf(title_table[i][2], buf);
|
757
|
+
|
758
|
+
return unichar_to_utf(tv != '\0' ? tv : c, buf);
|
759
|
+
}
|
760
|
+
|
761
|
+
/* {{{1
|
762
|
+
* The real implementation of downcase.
|
763
|
+
*/
|
764
|
+
static size_t
|
765
|
+
tolower_turkic_i(const char **p, char *buf)
|
766
|
+
{
|
767
|
+
unichar i = LATIN_SMALL_LETTER_DOTLESS_I;
|
768
|
+
|
769
|
+
if (utf_char(*p) == COMBINING_DOT_ABOVE) {
|
770
|
+
/* TODO: don’t we need to make sure we don’t go beyond the end
|
771
|
+
* of ‘p’? */
|
772
|
+
*p = utf_next(*p);
|
773
|
+
i = LATIN_SMALL_LETTER_I;
|
774
|
+
}
|
775
|
+
|
776
|
+
return unichar_to_utf(i, buf);
|
777
|
+
}
|
778
|
+
|
779
|
+
static size_t
|
780
|
+
tolower_lithuianian_i(char *buf, unichar base, unichar combiner)
|
781
|
+
{
|
782
|
+
size_t len = unichar_to_utf(base, buf);
|
783
|
+
len += unichar_to_utf(COMBINING_DOT_ABOVE, OFFSET_IF(buf, len));
|
784
|
+
if (combiner != '\0')
|
785
|
+
len += unichar_to_utf(combiner, OFFSET_IF(buf, len));
|
786
|
+
|
787
|
+
return len;
|
788
|
+
}
|
789
|
+
|
790
|
+
static size_t
|
791
|
+
tolower_sigma(const char **p, char *buf, const char *end, bool use_end)
|
792
|
+
{
|
793
|
+
unichar sigma = GREEK_SMALL_LETTER_FINAL_SIGMA;
|
794
|
+
|
795
|
+
/* SIGMA maps differently depending on whether it is final or not. The
|
796
|
+
* following simplified test would fail in the case of combining marks
|
797
|
+
* following the sigma, but I don't think that occurs in real text.
|
798
|
+
* The test here matches that in ICU. */
|
799
|
+
if ((!use_end || *p < end) && **p != '\0' && s_isalpha(s_type(utf_char(*p))))
|
800
|
+
sigma = GREEK_SMALL_LETTER_SIGMA;
|
801
|
+
|
802
|
+
return unichar_to_utf(sigma, buf);
|
803
|
+
}
|
804
|
+
|
805
|
+
static size_t
|
806
|
+
real_tolower_one(const char **p, const char *prev, char *buf,
|
807
|
+
LocaleType locale_type, const char *end, bool use_end)
|
808
|
+
{
|
809
|
+
unichar c = utf_char(prev);
|
810
|
+
int type = s_type(c);
|
811
|
+
|
812
|
+
if (locale_type == LOCALE_TURKIC && c == 'I')
|
813
|
+
return tolower_turkic_i(p, buf);
|
814
|
+
|
815
|
+
/* Introduce an explicit dot above the lowercasing capital I’s
|
816
|
+
* and J’s whenever there are more accents above.
|
817
|
+
* [SpecialCasing.txt] */
|
818
|
+
if (locale_type == LOCALE_LITHUANIAN) {
|
819
|
+
unichar base = LATIN_SMALL_LETTER_I;
|
820
|
+
unichar combiner = '\0';
|
821
|
+
|
822
|
+
switch (c) {
|
823
|
+
case LATIN_CAPITAL_LETTER_I_WITH_GRAVE:
|
824
|
+
combiner = COMBINING_GRAVE_ACCENT;
|
825
|
+
break;
|
826
|
+
case LATIN_CAPITAL_LETTER_I_WITH_ACUTE:
|
827
|
+
combiner = COMBINING_ACUTE_ACCENT;
|
828
|
+
break;
|
829
|
+
case LATIN_CAPITAL_LETTER_I_WITH_TILDE:
|
830
|
+
combiner = COMBINING_TILDE;
|
831
|
+
break;
|
832
|
+
case 'I':
|
833
|
+
case 'J':
|
834
|
+
case LATIN_CAPITAL_LETTER_I_WITH_OGONEK:
|
835
|
+
if (!has_more_above(*p))
|
836
|
+
goto no_lithuanian_i_casing;
|
837
|
+
|
838
|
+
base = unichar_tolower(c);
|
839
|
+
break;
|
840
|
+
default:
|
841
|
+
goto no_lithuanian_i_casing;
|
842
|
+
}
|
843
|
+
|
844
|
+
return tolower_lithuianian_i(buf, base, combiner);
|
845
|
+
}
|
846
|
+
|
847
|
+
no_lithuanian_i_casing:
|
848
|
+
|
849
|
+
if (c == GREEK_CAPITAL_LETTER_SIGMA)
|
850
|
+
return tolower_sigma(p, buf, end, use_end);
|
851
|
+
|
852
|
+
if (IS(type, OR(UNICODE_UPPERCASE_LETTER,
|
853
|
+
OR(UNICODE_TITLECASE_LETTER, 0))))
|
854
|
+
return real_do_tolower(c, type, buf);
|
855
|
+
|
856
|
+
size_t len = s_utf_skip_lengths[*(const unsigned char *)prev];
|
857
|
+
|
858
|
+
if (buf != NULL)
|
859
|
+
memcpy(buf, prev, len);
|
860
|
+
|
861
|
+
return len;
|
862
|
+
}
|
863
|
+
|
864
|
+
static size_t
|
865
|
+
real_tolower(const char *str, size_t max, bool use_max, char *buf,
|
866
|
+
LocaleType locale_type)
|
867
|
+
{
|
868
|
+
const char *p = str;
|
869
|
+
const char *end = str + max;
|
870
|
+
size_t len = 0;
|
871
|
+
|
872
|
+
while ((!use_max || p < end) && *p != '\0') {
|
873
|
+
const char *prev = p;
|
874
|
+
p = utf_next(p);
|
875
|
+
|
876
|
+
len += real_tolower_one(&p, prev, OFFSET_IF(buf, len),
|
877
|
+
locale_type, end, use_max);
|
878
|
+
}
|
879
|
+
|
880
|
+
return len;
|
881
|
+
}
|
882
|
+
|
883
|
+
|
884
|
+
/* {{{1 */
|
885
|
+
static char *
|
886
|
+
utf_downcase_impl(const char *str, size_t max, bool use_max)
|
887
|
+
{
|
888
|
+
assert(str != NULL);
|
889
|
+
|
890
|
+
LocaleType locale_type = get_locale_type();
|
891
|
+
|
892
|
+
size_t len = real_tolower(str, max, use_max, NULL, locale_type);
|
893
|
+
char *result = ALLOC_N(char, len + 1);
|
894
|
+
real_tolower(str, max, use_max, result, locale_type);
|
895
|
+
result[len] = '\0';
|
896
|
+
|
897
|
+
return result;
|
898
|
+
}
|
899
|
+
|
900
|
+
|
901
|
+
/* {{{1
|
902
|
+
* Convert all characters in ‘str’ to their lowercase representation if
|
903
|
+
* applicable. Returns the freshly allocated representation.
|
904
|
+
*/
|
905
|
+
char *
|
906
|
+
utf_downcase(const char *str)
|
907
|
+
{
|
908
|
+
return utf_downcase_impl(str, 0, false);
|
909
|
+
}
|
910
|
+
|
911
|
+
|
912
|
+
/* {{{1
|
913
|
+
* Convert all characters in ‘str’ to their lowercase representation if
|
914
|
+
* applicable. Returns the freshly allocated representation. Do this for at
|
915
|
+
* most ‘len˚ bytes from ‘str’.
|
916
|
+
*/
|
917
|
+
char *
|
918
|
+
utf_downcase_n(const char *str, size_t len)
|
919
|
+
{
|
920
|
+
return utf_downcase_impl(str, len, true);
|
921
|
+
}
|
922
|
+
|
923
|
+
|
924
|
+
/* {{{1
|
925
|
+
* The real implementation of case folding below.
|
926
|
+
*/
|
927
|
+
|
928
|
+
static bool
|
929
|
+
casefold_table_lookup(unichar c, char *folded, size_t *len)
|
930
|
+
{
|
931
|
+
int index;
|
932
|
+
|
933
|
+
if (!unicode_table_lookup(casefold_table, c, &index))
|
934
|
+
return false;
|
935
|
+
|
936
|
+
char const *folded_c = casefold_table[index].data;
|
937
|
+
|
938
|
+
if (folded != NULL)
|
939
|
+
strcpy(folded, folded_c);
|
940
|
+
|
941
|
+
*len += utf_byte_length(folded_c);
|
942
|
+
|
943
|
+
return true;
|
944
|
+
}
|
945
|
+
|
946
|
+
static char *
|
947
|
+
utf_foldcase_impl(const char *str, size_t max, bool use_max)
|
948
|
+
{
|
949
|
+
assert(str != NULL);
|
950
|
+
|
951
|
+
char *folded = NULL;
|
952
|
+
size_t len = 0;
|
953
|
+
|
954
|
+
again:
|
955
|
+
for (const char *p = str; (!use_max || p < str + max) && *p != '\0'; p = utf_next(p)) {
|
956
|
+
unichar c = utf_char(p);
|
957
|
+
|
958
|
+
if (casefold_table_lookup(c, OFFSET_IF(folded, len), &len))
|
959
|
+
continue;
|
960
|
+
|
961
|
+
len += unichar_to_utf(unichar_tolower(c), OFFSET_IF(folded, len));
|
962
|
+
}
|
963
|
+
|
964
|
+
if (folded == NULL) {
|
965
|
+
folded = ALLOC_N(char, len + 1);
|
966
|
+
folded[0] = NUL;
|
967
|
+
len = 0;
|
968
|
+
goto again;
|
969
|
+
}
|
970
|
+
|
971
|
+
folded[len] = '\0';
|
972
|
+
|
973
|
+
return folded;
|
974
|
+
}
|
975
|
+
|
976
|
+
|
977
|
+
/* {{{1
|
978
|
+
* Convert a string into a form that is independent of case. Return the
|
979
|
+
* freshly allocated representation.
|
980
|
+
*/
|
981
|
+
char *
|
982
|
+
utf_foldcase(const char *str)
|
983
|
+
{
|
984
|
+
return utf_foldcase_impl(str, 0, false);
|
985
|
+
}
|
986
|
+
|
987
|
+
|
988
|
+
/* {{{1
|
989
|
+
* Convert a string into a form that is independent of case. Return the
|
990
|
+
* freshly allocated representation. Do this for at most ‘len’ bytes from the
|
991
|
+
* string.
|
992
|
+
*/
|
993
|
+
char *
|
994
|
+
utf_foldcase_n(const char *str, size_t len)
|
995
|
+
{
|
996
|
+
return utf_foldcase_impl(str, len, true);
|
997
|
+
}
|
998
|
+
|
999
|
+
|
1000
|
+
/* {{{1
|
1001
|
+
* The real implementation of utf_width() and utf_width_n() below.
|
1002
|
+
*/
|
1003
|
+
static size_t
|
1004
|
+
utf_width_impl(const char *str, size_t len, bool use_len)
|
1005
|
+
{
|
1006
|
+
assert(str != NULL);
|
1007
|
+
|
1008
|
+
size_t width = 0;
|
1009
|
+
|
1010
|
+
for (const char *p = str; (!use_len || p < str + len) && *p != NUL; p = utf_next(p))
|
1011
|
+
width += unichar_iswide(utf_char(p)) ? 2 : 1;
|
1012
|
+
|
1013
|
+
return width;
|
1014
|
+
}
|
1015
|
+
|
1016
|
+
|
1017
|
+
/* {{{1
|
1018
|
+
* Calculate the width in cells of ‘str’.
|
1019
|
+
*/
|
1020
|
+
size_t
|
1021
|
+
utf_width(const char *str)
|
1022
|
+
{
|
1023
|
+
return utf_width_impl(str, 0, false);
|
1024
|
+
}
|
1025
|
+
|
1026
|
+
|
1027
|
+
/* {{{1
|
1028
|
+
* Calculate the width in cells of ‘str’, which is of length ‘len’.
|
1029
|
+
*/
|
1030
|
+
size_t
|
1031
|
+
utf_width_n(const char *str, size_t len)
|
1032
|
+
{
|
1033
|
+
return utf_width_impl(str, len, true);
|
1034
|
+
}
|
1035
|
+
|
1036
|
+
|
1037
|
+
/* {{{1
|
1038
|
+
* Retrieve the mirrored representation of ‘c’ (if any) and store it in
|
1039
|
+
* ‘mirrored’.
|
1040
|
+
*/
|
1041
|
+
bool
|
1042
|
+
unichar_mirror(unichar c, unichar *mirrored)
|
1043
|
+
{
|
1044
|
+
int index;
|
1045
|
+
|
1046
|
+
if (!unicode_table_lookup(bidi_mirroring_table, c, &index))
|
1047
|
+
return false;
|
1048
|
+
|
1049
|
+
if (mirrored != NULL)
|
1050
|
+
*mirrored = bidi_mirroring_table[index].mirrored_ch;
|
1051
|
+
|
1052
|
+
return true;
|
1053
|
+
}
|
1054
|
+
|
1055
|
+
|
1056
|
+
/* }}}1 */
|