bare-buffer 2.5.9 → 2.5.11
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CMakeLists.txt +4 -11
- package/package.json +3 -8
- package/prebuilds/android-arm/bare-buffer.bare +0 -0
- package/prebuilds/android-arm64/bare-buffer.bare +0 -0
- package/prebuilds/android-ia32/bare-buffer.bare +0 -0
- package/prebuilds/android-x64/bare-buffer.bare +0 -0
- package/prebuilds/darwin-arm64/bare-buffer.bare +0 -0
- package/prebuilds/darwin-x64/bare-buffer.bare +0 -0
- package/prebuilds/ios-arm64/bare-buffer.bare +0 -0
- package/prebuilds/ios-arm64-simulator/bare-buffer.bare +0 -0
- package/prebuilds/ios-x64-simulator/bare-buffer.bare +0 -0
- package/prebuilds/linux-arm64/bare-buffer.bare +0 -0
- package/prebuilds/linux-x64/bare-buffer.bare +0 -0
- package/prebuilds/win32-arm64/bare-buffer.bare +0 -0
- package/prebuilds/win32-x64/bare-buffer.bare +0 -0
- package/vendor/libbase64/CMakeLists.txt +0 -70
- package/vendor/libbase64/include/base64.h +0 -22
- package/vendor/libbase64/src/base64.c +0 -107
- package/vendor/libhex/CMakeLists.txt +0 -71
- package/vendor/libhex/include/hex.h +0 -22
- package/vendor/libhex/src/hex.c +0 -67
- package/vendor/libutf/CMakeLists.txt +0 -93
- package/vendor/libutf/include/utf/string.h +0 -786
- package/vendor/libutf/include/utf.h +0 -132
- package/vendor/libutf/src/ascii/validate.c +0 -47
- package/vendor/libutf/src/endianness.c +0 -19
- package/vendor/libutf/src/endianness.h +0 -54
- package/vendor/libutf/src/latin1/convert-to-utf16.c +0 -37
- package/vendor/libutf/src/latin1/convert-to-utf32.c +0 -34
- package/vendor/libutf/src/latin1/convert-to-utf8.c +0 -58
- package/vendor/libutf/src/latin1/length-from-utf16.c +0 -26
- package/vendor/libutf/src/latin1/length-from-utf32.c +0 -26
- package/vendor/libutf/src/latin1/length-from-utf8.c +0 -34
- package/vendor/libutf/src/utf16/convert-to-latin1.c +0 -41
- package/vendor/libutf/src/utf16/convert-to-utf32.c +0 -56
- package/vendor/libutf/src/utf16/convert-to-utf8.c +0 -75
- package/vendor/libutf/src/utf16/length-from-latin1.c +0 -26
- package/vendor/libutf/src/utf16/length-from-utf32.c +0 -33
- package/vendor/libutf/src/utf16/length-from-utf8.c +0 -38
- package/vendor/libutf/src/utf16/validate.c +0 -53
- package/vendor/libutf/src/utf32/convert-to-latin1.c +0 -40
- package/vendor/libutf/src/utf32/convert-to-utf16.c +0 -56
- package/vendor/libutf/src/utf32/convert-to-utf8.c +0 -71
- package/vendor/libutf/src/utf32/length-from-latin1.c +0 -26
- package/vendor/libutf/src/utf32/length-from-utf16.c +0 -35
- package/vendor/libutf/src/utf32/length-from-utf8.c +0 -35
- package/vendor/libutf/src/utf32/validate.c +0 -39
- package/vendor/libutf/src/utf8/convert-to-latin1.c +0 -70
- package/vendor/libutf/src/utf8/convert-to-utf16.c +0 -95
- package/vendor/libutf/src/utf8/convert-to-utf32.c +0 -110
- package/vendor/libutf/src/utf8/length-from-latin1.c +0 -32
- package/vendor/libutf/src/utf8/length-from-utf16.c +0 -44
- package/vendor/libutf/src/utf8/length-from-utf32.c +0 -35
- package/vendor/libutf/src/utf8/string.c +0 -149
- package/vendor/libutf/src/utf8/validate.c +0 -107
|
@@ -1,132 +0,0 @@
|
|
|
1
|
-
#ifndef UTF_H
|
|
2
|
-
#define UTF_H
|
|
3
|
-
|
|
4
|
-
#ifdef __cplusplus
|
|
5
|
-
extern "C" {
|
|
6
|
-
#endif
|
|
7
|
-
|
|
8
|
-
#include <stdbool.h>
|
|
9
|
-
#include <stddef.h>
|
|
10
|
-
#include <stdint.h>
|
|
11
|
-
|
|
12
|
-
typedef uint_least8_t utf8_t;
|
|
13
|
-
typedef uint_least16_t utf16_t;
|
|
14
|
-
typedef uint_least32_t utf32_t;
|
|
15
|
-
typedef uint_least8_t latin1_t;
|
|
16
|
-
typedef uint_least8_t ascii_t;
|
|
17
|
-
|
|
18
|
-
/**
|
|
19
|
-
* UTF-8
|
|
20
|
-
*/
|
|
21
|
-
|
|
22
|
-
bool
|
|
23
|
-
utf8_validate (const utf8_t *data, size_t len);
|
|
24
|
-
|
|
25
|
-
size_t
|
|
26
|
-
utf8_length_from_utf16le (const utf16_t *data, size_t len);
|
|
27
|
-
|
|
28
|
-
size_t
|
|
29
|
-
utf8_length_from_utf32 (const utf32_t *data, size_t len);
|
|
30
|
-
|
|
31
|
-
size_t
|
|
32
|
-
utf8_length_from_latin1 (const latin1_t *data, size_t len);
|
|
33
|
-
|
|
34
|
-
size_t
|
|
35
|
-
utf8_convert_to_utf16le (const utf8_t *data, size_t len, utf16_t *result);
|
|
36
|
-
|
|
37
|
-
size_t
|
|
38
|
-
utf8_convert_to_utf32 (const utf8_t *data, size_t len, utf32_t *result);
|
|
39
|
-
|
|
40
|
-
size_t
|
|
41
|
-
utf8_convert_to_latin1 (const utf8_t *data, size_t len, latin1_t *result);
|
|
42
|
-
|
|
43
|
-
/**
|
|
44
|
-
* UTF-16
|
|
45
|
-
*/
|
|
46
|
-
|
|
47
|
-
bool
|
|
48
|
-
utf16le_validate (const utf16_t *data, size_t len);
|
|
49
|
-
|
|
50
|
-
size_t
|
|
51
|
-
utf16_length_from_utf8 (const utf8_t *data, size_t len);
|
|
52
|
-
|
|
53
|
-
size_t
|
|
54
|
-
utf16_length_from_utf32 (const utf32_t *data, size_t len);
|
|
55
|
-
|
|
56
|
-
size_t
|
|
57
|
-
utf16_length_from_latin1 (const latin1_t *data, size_t len);
|
|
58
|
-
|
|
59
|
-
size_t
|
|
60
|
-
utf16le_convert_to_utf8 (const utf16_t *data, size_t len, utf8_t *result);
|
|
61
|
-
|
|
62
|
-
size_t
|
|
63
|
-
utf16le_convert_to_utf32 (const utf16_t *data, size_t len, utf32_t *result);
|
|
64
|
-
|
|
65
|
-
size_t
|
|
66
|
-
utf16le_convert_to_latin1 (const utf16_t *data, size_t len, latin1_t *result);
|
|
67
|
-
|
|
68
|
-
/**
|
|
69
|
-
* UTF-32
|
|
70
|
-
*/
|
|
71
|
-
|
|
72
|
-
bool
|
|
73
|
-
utf32_validate (const utf32_t *data, size_t len);
|
|
74
|
-
|
|
75
|
-
size_t
|
|
76
|
-
utf32_length_from_utf8 (const utf8_t *data, size_t len);
|
|
77
|
-
|
|
78
|
-
size_t
|
|
79
|
-
utf32_length_from_utf16le (const utf16_t *data, size_t len);
|
|
80
|
-
|
|
81
|
-
size_t
|
|
82
|
-
utf32_length_from_latin1 (const latin1_t *data, size_t len);
|
|
83
|
-
|
|
84
|
-
size_t
|
|
85
|
-
utf32_convert_to_utf8 (const utf32_t *data, size_t len, utf8_t *result);
|
|
86
|
-
|
|
87
|
-
size_t
|
|
88
|
-
utf32_convert_to_utf16le (const utf32_t *data, size_t len, utf16_t *result);
|
|
89
|
-
|
|
90
|
-
size_t
|
|
91
|
-
utf32_convert_to_latin1 (const utf32_t *data, size_t len, latin1_t *result);
|
|
92
|
-
|
|
93
|
-
/**
|
|
94
|
-
* Latin-1
|
|
95
|
-
*/
|
|
96
|
-
|
|
97
|
-
size_t
|
|
98
|
-
latin1_length_from_utf8 (const utf8_t *data, size_t len);
|
|
99
|
-
|
|
100
|
-
size_t
|
|
101
|
-
latin1_length_from_utf16le (const utf16_t *data, size_t len);
|
|
102
|
-
|
|
103
|
-
size_t
|
|
104
|
-
latin1_length_from_utf32 (const utf32_t *data, size_t len);
|
|
105
|
-
|
|
106
|
-
size_t
|
|
107
|
-
latin1_convert_to_utf8 (const latin1_t *data, size_t len, utf8_t *result);
|
|
108
|
-
|
|
109
|
-
size_t
|
|
110
|
-
latin1_convert_to_utf16le (const latin1_t *data, size_t len, utf16_t *result);
|
|
111
|
-
|
|
112
|
-
size_t
|
|
113
|
-
latin1_convert_to_utf32 (const latin1_t *data, size_t len, utf32_t *result);
|
|
114
|
-
|
|
115
|
-
/**
|
|
116
|
-
* ASCII
|
|
117
|
-
*/
|
|
118
|
-
|
|
119
|
-
bool
|
|
120
|
-
ascii_validate (const ascii_t *data, size_t len);
|
|
121
|
-
|
|
122
|
-
/**
|
|
123
|
-
* Strings
|
|
124
|
-
*/
|
|
125
|
-
|
|
126
|
-
#include <utf/string.h>
|
|
127
|
-
|
|
128
|
-
#ifdef __cplusplus
|
|
129
|
-
}
|
|
130
|
-
#endif
|
|
131
|
-
|
|
132
|
-
#endif // UTF_H
|
|
@@ -1,47 +0,0 @@
|
|
|
1
|
-
#include <stdbool.h>
|
|
2
|
-
#include <stddef.h>
|
|
3
|
-
#include <stdint.h>
|
|
4
|
-
#include <string.h>
|
|
5
|
-
|
|
6
|
-
#include "../../include/utf.h"
|
|
7
|
-
|
|
8
|
-
/**
|
|
9
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
10
|
-
*
|
|
11
|
-
* Copyright 2020 The simdutf authors
|
|
12
|
-
*
|
|
13
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
14
|
-
* you may not use this file except in compliance with the License.
|
|
15
|
-
* You may obtain a copy of the License at
|
|
16
|
-
*
|
|
17
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
18
|
-
*
|
|
19
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
20
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
21
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
22
|
-
* See the License for the specific language governing permissions and
|
|
23
|
-
* limitations under the License.
|
|
24
|
-
*/
|
|
25
|
-
|
|
26
|
-
bool
|
|
27
|
-
ascii_validate (const ascii_t *data, size_t len) {
|
|
28
|
-
uint64_t pos = 0;
|
|
29
|
-
|
|
30
|
-
for (; pos + 16 <= len; pos += 16) {
|
|
31
|
-
uint64_t v1;
|
|
32
|
-
memcpy(&v1, data + pos, sizeof(uint64_t));
|
|
33
|
-
uint64_t v2;
|
|
34
|
-
memcpy(&v2, data + pos + sizeof(uint64_t), sizeof(uint64_t));
|
|
35
|
-
uint64_t v = v1 | v2;
|
|
36
|
-
if ((v & 0x8080808080808080) != 0) {
|
|
37
|
-
return false;
|
|
38
|
-
}
|
|
39
|
-
}
|
|
40
|
-
for (; pos < len; pos++) {
|
|
41
|
-
if (data[pos] >= 0b10000000) {
|
|
42
|
-
return false;
|
|
43
|
-
}
|
|
44
|
-
}
|
|
45
|
-
|
|
46
|
-
return true;
|
|
47
|
-
}
|
|
@@ -1,19 +0,0 @@
|
|
|
1
|
-
#include <stdbool.h>
|
|
2
|
-
#include <stdint.h>
|
|
3
|
-
|
|
4
|
-
#include "endianness.h"
|
|
5
|
-
|
|
6
|
-
extern utf_endianness_t
|
|
7
|
-
utf_endianness (void);
|
|
8
|
-
|
|
9
|
-
extern bool
|
|
10
|
-
utf_is_le (void);
|
|
11
|
-
|
|
12
|
-
extern bool
|
|
13
|
-
utf_is_be (void);
|
|
14
|
-
|
|
15
|
-
extern uint16_t
|
|
16
|
-
utf_swap_uint16 (uint16_t n);
|
|
17
|
-
|
|
18
|
-
extern uint32_t
|
|
19
|
-
utf_swap_uint32 (uint32_t n);
|
|
@@ -1,54 +0,0 @@
|
|
|
1
|
-
#ifndef UTF_ENDIANNESS_H
|
|
2
|
-
#define UTF_ENDIANNESS_H
|
|
3
|
-
|
|
4
|
-
#ifdef __cplusplus
|
|
5
|
-
extern "C" {
|
|
6
|
-
#endif
|
|
7
|
-
|
|
8
|
-
#include <stdbool.h>
|
|
9
|
-
#include <stdint.h>
|
|
10
|
-
|
|
11
|
-
typedef enum {
|
|
12
|
-
utf_le,
|
|
13
|
-
utf_be
|
|
14
|
-
} utf_endianness_t;
|
|
15
|
-
|
|
16
|
-
inline utf_endianness_t
|
|
17
|
-
utf_endianness (void) {
|
|
18
|
-
const union {
|
|
19
|
-
uint8_t u8[2];
|
|
20
|
-
uint16_t u16;
|
|
21
|
-
} byte_order = {{1, 0}};
|
|
22
|
-
|
|
23
|
-
return byte_order.u16 == 1 ? utf_le : utf_be;
|
|
24
|
-
}
|
|
25
|
-
|
|
26
|
-
inline bool
|
|
27
|
-
utf_is_le (void) {
|
|
28
|
-
return utf_endianness() == utf_le;
|
|
29
|
-
}
|
|
30
|
-
|
|
31
|
-
inline bool
|
|
32
|
-
utf_is_be (void) {
|
|
33
|
-
return utf_endianness() == utf_be;
|
|
34
|
-
}
|
|
35
|
-
|
|
36
|
-
inline uint16_t
|
|
37
|
-
utf_swap_uint16 (uint16_t n) {
|
|
38
|
-
return ((n & 0x00ff) << 8) |
|
|
39
|
-
((n & 0xff00) >> 8);
|
|
40
|
-
}
|
|
41
|
-
|
|
42
|
-
inline uint32_t
|
|
43
|
-
utf_swap_uint32 (uint32_t n) {
|
|
44
|
-
return ((n & 0x000000ff) << 24) |
|
|
45
|
-
((n & 0x0000ff00) << 8) |
|
|
46
|
-
((n & 0x00ff0000) >> 8) |
|
|
47
|
-
((n & 0xff000000) >> 24);
|
|
48
|
-
}
|
|
49
|
-
|
|
50
|
-
#ifdef __cplusplus
|
|
51
|
-
}
|
|
52
|
-
#endif
|
|
53
|
-
|
|
54
|
-
#endif // UTF_ENDIANNESS_H
|
|
@@ -1,37 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
#include <stdint.h>
|
|
3
|
-
#include <string.h>
|
|
4
|
-
|
|
5
|
-
#include "../../include/utf.h"
|
|
6
|
-
#include "../endianness.h"
|
|
7
|
-
|
|
8
|
-
/**
|
|
9
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
10
|
-
*
|
|
11
|
-
* Copyright 2020 The simdutf authors
|
|
12
|
-
*
|
|
13
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
14
|
-
* you may not use this file except in compliance with the License.
|
|
15
|
-
* You may obtain a copy of the License at
|
|
16
|
-
*
|
|
17
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
18
|
-
*
|
|
19
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
20
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
21
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
22
|
-
* See the License for the specific language governing permissions and
|
|
23
|
-
* limitations under the License.
|
|
24
|
-
*/
|
|
25
|
-
|
|
26
|
-
size_t
|
|
27
|
-
latin1_convert_to_utf16le (const latin1_t *data, size_t len, utf16_t *result) {
|
|
28
|
-
size_t pos = 0;
|
|
29
|
-
utf16_t *start = result;
|
|
30
|
-
|
|
31
|
-
while (pos < len) {
|
|
32
|
-
*result++ = utf_is_be() ? utf_swap_uint16(data[pos]) : data[pos];
|
|
33
|
-
pos++;
|
|
34
|
-
}
|
|
35
|
-
|
|
36
|
-
return result - start;
|
|
37
|
-
}
|
|
@@ -1,34 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
#include <stdint.h>
|
|
3
|
-
#include <string.h>
|
|
4
|
-
|
|
5
|
-
#include "../../include/utf.h"
|
|
6
|
-
|
|
7
|
-
/**
|
|
8
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
9
|
-
*
|
|
10
|
-
* Copyright 2020 The simdutf authors
|
|
11
|
-
*
|
|
12
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
13
|
-
* you may not use this file except in compliance with the License.
|
|
14
|
-
* You may obtain a copy of the License at
|
|
15
|
-
*
|
|
16
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
17
|
-
*
|
|
18
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
19
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
20
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
21
|
-
* See the License for the specific language governing permissions and
|
|
22
|
-
* limitations under the License.
|
|
23
|
-
*/
|
|
24
|
-
|
|
25
|
-
size_t
|
|
26
|
-
latin1_convert_to_utf32 (const latin1_t *data, size_t len, utf32_t *result) {
|
|
27
|
-
utf32_t *start = result;
|
|
28
|
-
|
|
29
|
-
for (size_t i = 0; i < len; i++) {
|
|
30
|
-
*result++ = data[i];
|
|
31
|
-
}
|
|
32
|
-
|
|
33
|
-
return result - start;
|
|
34
|
-
}
|
|
@@ -1,58 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
#include <stdint.h>
|
|
3
|
-
#include <string.h>
|
|
4
|
-
|
|
5
|
-
#include "../../include/utf.h"
|
|
6
|
-
|
|
7
|
-
/**
|
|
8
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
9
|
-
*
|
|
10
|
-
* Copyright 2020 The simdutf authors
|
|
11
|
-
*
|
|
12
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
13
|
-
* you may not use this file except in compliance with the License.
|
|
14
|
-
* You may obtain a copy of the License at
|
|
15
|
-
*
|
|
16
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
17
|
-
*
|
|
18
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
19
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
20
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
21
|
-
* See the License for the specific language governing permissions and
|
|
22
|
-
* limitations under the License.
|
|
23
|
-
*/
|
|
24
|
-
|
|
25
|
-
size_t
|
|
26
|
-
latin1_convert_to_utf8 (const latin1_t *data, size_t len, utf8_t *result) {
|
|
27
|
-
size_t pos = 0;
|
|
28
|
-
utf8_t *start = result;
|
|
29
|
-
|
|
30
|
-
while (pos < len) {
|
|
31
|
-
if (pos + 16 <= len) {
|
|
32
|
-
uint64_t v1;
|
|
33
|
-
memcpy(&v1, data + pos, sizeof(uint64_t));
|
|
34
|
-
uint64_t v2;
|
|
35
|
-
memcpy(&v2, data + pos + sizeof(uint64_t), sizeof(uint64_t));
|
|
36
|
-
uint64_t v = v1 | v2;
|
|
37
|
-
if ((v & 0x8080808080808080) == 0) {
|
|
38
|
-
size_t final_pos = pos + 16;
|
|
39
|
-
while (pos < final_pos) {
|
|
40
|
-
*result++ = data[pos];
|
|
41
|
-
pos++;
|
|
42
|
-
}
|
|
43
|
-
continue;
|
|
44
|
-
}
|
|
45
|
-
}
|
|
46
|
-
unsigned char byte = data[pos];
|
|
47
|
-
if ((byte & 0x80) == 0) {
|
|
48
|
-
*result++ = byte;
|
|
49
|
-
pos++;
|
|
50
|
-
} else {
|
|
51
|
-
*result++ = (byte >> 6) | 0b11000000;
|
|
52
|
-
*result++ = (byte & 0b111111) | 0b10000000;
|
|
53
|
-
pos++;
|
|
54
|
-
}
|
|
55
|
-
}
|
|
56
|
-
|
|
57
|
-
return result - start;
|
|
58
|
-
}
|
|
@@ -1,26 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
|
|
3
|
-
#include "../../include/utf.h"
|
|
4
|
-
|
|
5
|
-
/**
|
|
6
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
7
|
-
*
|
|
8
|
-
* Copyright 2020 The simdutf authors
|
|
9
|
-
*
|
|
10
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
11
|
-
* you may not use this file except in compliance with the License.
|
|
12
|
-
* You may obtain a copy of the License at
|
|
13
|
-
*
|
|
14
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
15
|
-
*
|
|
16
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
17
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
18
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
19
|
-
* See the License for the specific language governing permissions and
|
|
20
|
-
* limitations under the License.
|
|
21
|
-
*/
|
|
22
|
-
|
|
23
|
-
size_t
|
|
24
|
-
latin1_length_from_utf16le (const utf16_t *data, size_t len) {
|
|
25
|
-
return len;
|
|
26
|
-
}
|
|
@@ -1,26 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
|
|
3
|
-
#include "../../include/utf.h"
|
|
4
|
-
|
|
5
|
-
/**
|
|
6
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
7
|
-
*
|
|
8
|
-
* Copyright 2020 The simdutf authors
|
|
9
|
-
*
|
|
10
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
11
|
-
* you may not use this file except in compliance with the License.
|
|
12
|
-
* You may obtain a copy of the License at
|
|
13
|
-
*
|
|
14
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
15
|
-
*
|
|
16
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
17
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
18
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
19
|
-
* See the License for the specific language governing permissions and
|
|
20
|
-
* limitations under the License.
|
|
21
|
-
*/
|
|
22
|
-
|
|
23
|
-
size_t
|
|
24
|
-
latin1_length_from_utf32 (const utf32_t *data, size_t len) {
|
|
25
|
-
return len;
|
|
26
|
-
}
|
|
@@ -1,34 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
|
|
3
|
-
#include "../../include/utf.h"
|
|
4
|
-
|
|
5
|
-
/**
|
|
6
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
7
|
-
*
|
|
8
|
-
* Copyright 2020 The simdutf authors
|
|
9
|
-
*
|
|
10
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
11
|
-
* you may not use this file except in compliance with the License.
|
|
12
|
-
* You may obtain a copy of the License at
|
|
13
|
-
*
|
|
14
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
15
|
-
*
|
|
16
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
17
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
18
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
19
|
-
* See the License for the specific language governing permissions and
|
|
20
|
-
* limitations under the License.
|
|
21
|
-
*/
|
|
22
|
-
|
|
23
|
-
size_t
|
|
24
|
-
latin1_length_from_utf8 (const utf8_t *data, size_t len) {
|
|
25
|
-
size_t counter = 0;
|
|
26
|
-
|
|
27
|
-
for (size_t i = 0; i < len; i++) {
|
|
28
|
-
if ((int8_t) data[i] > -65) {
|
|
29
|
-
counter++;
|
|
30
|
-
}
|
|
31
|
-
}
|
|
32
|
-
|
|
33
|
-
return counter;
|
|
34
|
-
}
|
|
@@ -1,41 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
#include <stdint.h>
|
|
3
|
-
|
|
4
|
-
#include "../../include/utf.h"
|
|
5
|
-
#include "../endianness.h"
|
|
6
|
-
|
|
7
|
-
/**
|
|
8
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
9
|
-
*
|
|
10
|
-
* Copyright 2020 The simdutf authors
|
|
11
|
-
*
|
|
12
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
13
|
-
* you may not use this file except in compliance with the License.
|
|
14
|
-
* You may obtain a copy of the License at
|
|
15
|
-
*
|
|
16
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
17
|
-
*
|
|
18
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
19
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
20
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
21
|
-
* See the License for the specific language governing permissions and
|
|
22
|
-
* limitations under the License.
|
|
23
|
-
*/
|
|
24
|
-
|
|
25
|
-
size_t
|
|
26
|
-
utf16le_convert_to_latin1 (const utf16_t *data, size_t len, latin1_t *result) {
|
|
27
|
-
size_t pos = 0;
|
|
28
|
-
uint16_t word, overflow = 0;
|
|
29
|
-
latin1_t *start = result;
|
|
30
|
-
|
|
31
|
-
while (pos < len) {
|
|
32
|
-
word = utf_is_be() ? utf_swap_uint16(data[pos]) : data[pos];
|
|
33
|
-
overflow |= word;
|
|
34
|
-
*result++ = word & 0xff;
|
|
35
|
-
pos++;
|
|
36
|
-
}
|
|
37
|
-
|
|
38
|
-
if (overflow & 0xff00) return 0;
|
|
39
|
-
|
|
40
|
-
return result - start;
|
|
41
|
-
}
|
|
@@ -1,56 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
#include <stdint.h>
|
|
3
|
-
|
|
4
|
-
#include "../../include/utf.h"
|
|
5
|
-
#include "../endianness.h"
|
|
6
|
-
|
|
7
|
-
/**
|
|
8
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
9
|
-
*
|
|
10
|
-
* Copyright 2020 The simdutf authors
|
|
11
|
-
*
|
|
12
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
13
|
-
* you may not use this file except in compliance with the License.
|
|
14
|
-
* You may obtain a copy of the License at
|
|
15
|
-
*
|
|
16
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
17
|
-
*
|
|
18
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
19
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
20
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
21
|
-
* See the License for the specific language governing permissions and
|
|
22
|
-
* limitations under the License.
|
|
23
|
-
*/
|
|
24
|
-
|
|
25
|
-
size_t
|
|
26
|
-
utf16le_convert_to_utf32 (const utf16_t *data, size_t len, utf32_t *result) {
|
|
27
|
-
size_t pos = 0;
|
|
28
|
-
uint16_t word, diff;
|
|
29
|
-
utf32_t *start = result;
|
|
30
|
-
|
|
31
|
-
while (pos < len) {
|
|
32
|
-
word = utf_is_be() ? utf_swap_uint16(data[pos]) : data[pos];
|
|
33
|
-
if ((word & 0xf800) != 0xd800) {
|
|
34
|
-
*result++ = word;
|
|
35
|
-
pos++;
|
|
36
|
-
} else {
|
|
37
|
-
diff = word - 0xd800;
|
|
38
|
-
if (diff > 0x3ff) {
|
|
39
|
-
return 0;
|
|
40
|
-
}
|
|
41
|
-
if (pos + 1 >= len) {
|
|
42
|
-
return 0;
|
|
43
|
-
}
|
|
44
|
-
word = utf_is_be() ? utf_swap_uint16(data[pos + 1]) : data[pos + 1];
|
|
45
|
-
uint32_t value = (diff << 10) + (word - 0xdc00) + 0x10000;
|
|
46
|
-
diff = word - 0xdc00;
|
|
47
|
-
if (diff > 0x3ff) {
|
|
48
|
-
return 0;
|
|
49
|
-
}
|
|
50
|
-
*result++ = value;
|
|
51
|
-
pos += 2;
|
|
52
|
-
}
|
|
53
|
-
}
|
|
54
|
-
|
|
55
|
-
return result - start;
|
|
56
|
-
}
|
|
@@ -1,75 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
#include <stdint.h>
|
|
3
|
-
#include <string.h>
|
|
4
|
-
|
|
5
|
-
#include "../../include/utf.h"
|
|
6
|
-
#include "../endianness.h"
|
|
7
|
-
|
|
8
|
-
/**
|
|
9
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
10
|
-
*
|
|
11
|
-
* Copyright 2020 The simdutf authors
|
|
12
|
-
*
|
|
13
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
14
|
-
* you may not use this file except in compliance with the License.
|
|
15
|
-
* You may obtain a copy of the License at
|
|
16
|
-
*
|
|
17
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
18
|
-
*
|
|
19
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
20
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
21
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
22
|
-
* See the License for the specific language governing permissions and
|
|
23
|
-
* limitations under the License.
|
|
24
|
-
*/
|
|
25
|
-
|
|
26
|
-
size_t
|
|
27
|
-
utf16le_convert_to_utf8 (const utf16_t *data, size_t len, utf8_t *result) {
|
|
28
|
-
size_t pos = 0;
|
|
29
|
-
uint16_t word, diff;
|
|
30
|
-
utf8_t *start = result;
|
|
31
|
-
|
|
32
|
-
while (pos < len) {
|
|
33
|
-
if (pos + 4 <= len) {
|
|
34
|
-
uint64_t v;
|
|
35
|
-
memcpy(&v, data + pos, sizeof(uint64_t));
|
|
36
|
-
if (utf_is_be()) v = (v >> 8) | (v << (64 - 8));
|
|
37
|
-
if ((v & 0xff80ff80ff80ff80) == 0) {
|
|
38
|
-
size_t final_pos = pos + 4;
|
|
39
|
-
while (pos < final_pos) {
|
|
40
|
-
*result++ = utf_is_be() ? utf_swap_uint16(data[pos]) : data[pos];
|
|
41
|
-
pos++;
|
|
42
|
-
}
|
|
43
|
-
continue;
|
|
44
|
-
}
|
|
45
|
-
}
|
|
46
|
-
word = utf_is_be() ? utf_swap_uint16(data[pos]) : data[pos];
|
|
47
|
-
if ((word & 0xff80) == 0) {
|
|
48
|
-
*result++ = word;
|
|
49
|
-
pos++;
|
|
50
|
-
} else if ((word & 0xf800) == 0) {
|
|
51
|
-
*result++ = (word >> 6) | 0b11000000;
|
|
52
|
-
*result++ = (word & 0b111111) | 0b10000000;
|
|
53
|
-
pos++;
|
|
54
|
-
} else if ((word & 0xf800) != 0xd800) {
|
|
55
|
-
*result++ = (word >> 12) | 0b11100000;
|
|
56
|
-
*result++ = ((word >> 6) & 0b111111) | 0b10000000;
|
|
57
|
-
*result++ = (word & 0b111111) | 0b10000000;
|
|
58
|
-
pos++;
|
|
59
|
-
} else {
|
|
60
|
-
diff = word - 0xd800;
|
|
61
|
-
if (pos + 1 >= len) {
|
|
62
|
-
return 0;
|
|
63
|
-
}
|
|
64
|
-
word = utf_is_be() ? utf_swap_uint16(data[pos + 1]) : data[pos + 1];
|
|
65
|
-
uint32_t value = (diff << 10) + (word - 0xdc00) + 0x10000;
|
|
66
|
-
*result++ = (value >> 18) | 0b11110000;
|
|
67
|
-
*result++ = ((value >> 12) & 0b111111) | 0b10000000;
|
|
68
|
-
*result++ = ((value >> 6) & 0b111111) | 0b10000000;
|
|
69
|
-
*result++ = (value & 0b111111) | 0b10000000;
|
|
70
|
-
pos += 2;
|
|
71
|
-
}
|
|
72
|
-
}
|
|
73
|
-
|
|
74
|
-
return result - start;
|
|
75
|
-
}
|
|
@@ -1,26 +0,0 @@
|
|
|
1
|
-
#include <stddef.h>
|
|
2
|
-
|
|
3
|
-
#include "../../include/utf.h"
|
|
4
|
-
|
|
5
|
-
/**
|
|
6
|
-
* Modified from https://github.com/simdutf/simdutf
|
|
7
|
-
*
|
|
8
|
-
* Copyright 2020 The simdutf authors
|
|
9
|
-
*
|
|
10
|
-
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
11
|
-
* you may not use this file except in compliance with the License.
|
|
12
|
-
* You may obtain a copy of the License at
|
|
13
|
-
*
|
|
14
|
-
* http://www.apache.org/licenses/LICENSE-2.0
|
|
15
|
-
*
|
|
16
|
-
* Unless required by applicable law or agreed to in writing, software
|
|
17
|
-
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
18
|
-
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
19
|
-
* See the License for the specific language governing permissions and
|
|
20
|
-
* limitations under the License.
|
|
21
|
-
*/
|
|
22
|
-
|
|
23
|
-
size_t
|
|
24
|
-
utf16_length_from_latin1 (const utf8_t *data, size_t len) {
|
|
25
|
-
return len;
|
|
26
|
-
}
|