| 1 |
|
|
1 |
|
|
| 2 |
|
// Copyright (c) 2020 Krystian Stasiowski (sdkrystian@gmail.com)
|
2 |
|
// Copyright (c) 2020 Krystian Stasiowski (sdkrystian@gmail.com)
|
| 3 |
|
|
3 |
|
|
| 4 |
|
// Distributed under the Boost Software License, Version 1.0. (See accompanying
|
4 |
|
// Distributed under the Boost Software License, Version 1.0. (See accompanying
|
| 5 |
|
// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
|
5 |
|
// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
|
| 6 |
|
|
6 |
|
|
| 7 |
|
// Official repository: https://github.com/boostorg/json
|
7 |
|
// Official repository: https://github.com/boostorg/json
|
| 8 |
|
|
8 |
|
|
| 9 |
|
|
9 |
|
|
| 10 |
|
#ifndef BOOST_JSON_DETAIL_UTF8_HPP
|
10 |
|
#ifndef BOOST_JSON_DETAIL_UTF8_HPP
|
| 11 |
|
#define BOOST_JSON_DETAIL_UTF8_HPP
|
11 |
|
#define BOOST_JSON_DETAIL_UTF8_HPP
|
| 12 |
|
|
12 |
|
|
| 13 |
|
#include <boost/endian/conversion.hpp>
|
13 |
|
#include <boost/endian/conversion.hpp>
|
| 14 |
|
#include <boost/json/detail/config.hpp>
|
14 |
|
#include <boost/json/detail/config.hpp>
|
| 15 |
|
|
15 |
|
|
| 16 |
|
|
16 |
|
|
| 17 |
|
|
17 |
|
|
| 18 |
|
|
18 |
|
|
| 19 |
|
|
19 |
|
|
| 20 |
|
|
20 |
|
|
| 21 |
|
|
21 |
|
|
| 22 |
|
|
22 |
|
|
| 23 |
|
|
23 |
|
|
| 24 |
|
|
24 |
|
|
| 25 |
|
|
25 |
|
|
| 26 |
|
load_little_endian(void const* p)
|
26 |
|
load_little_endian(void const* p)
|
| 27 |
|
|
27 |
|
|
| 28 |
|
|
28 |
|
|
| 29 |
|
|
29 |
|
|
| 30 |
|
endian::little_to_native_inplace(v);
|
30 |
|
endian::little_to_native_inplace(v);
|
| 31 |
|
|
31 |
|
|
| 32 |
|
|
32 |
|
|
| 33 |
|
|
33 |
|
|
| 34 |
|
|
34 |
|
|
| 35 |
|
|
35 |
|
|
| 36 |
|
|
36 |
|
|
| 37 |
|
|
37 |
|
|
| 38 |
|
|
38 |
|
|
| 39 |
|
// 0x102 = 2 bytes, second byte [80, BF]
|
39 |
|
// 0x102 = 2 bytes, second byte [80, BF]
|
| 40 |
|
// 0x203 = 3 bytes, second byte [A0, BF]
|
40 |
|
// 0x203 = 3 bytes, second byte [A0, BF]
|
| 41 |
|
// 0x303 = 3 bytes, second byte [80, BF]
|
41 |
|
// 0x303 = 3 bytes, second byte [80, BF]
|
| 42 |
|
// 0x403 = 3 bytes, second byte [80, 9F]
|
42 |
|
// 0x403 = 3 bytes, second byte [80, 9F]
|
| 43 |
|
// 0x504 = 4 bytes, second byte [90, BF]
|
43 |
|
// 0x504 = 4 bytes, second byte [90, BF]
|
| 44 |
|
// 0x604 = 4 bytes, second byte [80, BF]
|
44 |
|
// 0x604 = 4 bytes, second byte [80, BF]
|
| 45 |
|
// 0x704 = 4 bytes, second byte [80, 8F]
|
45 |
|
// 0x704 = 4 bytes, second byte [80, 8F]
|
| 46 |
|
static constexpr uint16_t first[128]
|
46 |
|
static constexpr uint16_t first[128]
|
| 47 |
|
|
47 |
|
|
| 48 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
48 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 49 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
49 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 50 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
50 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 51 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
51 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 52 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
52 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 53 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
53 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 54 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
54 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 55 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
55 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 56 |
|
|
56 |
|
|
| 57 |
|
0x000, 0x000, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
57 |
|
0x000, 0x000, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
| 58 |
|
0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
58 |
|
0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
| 59 |
|
0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
59 |
|
0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
| 60 |
|
0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
60 |
|
0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102, 0x102,
|
| 61 |
|
0x203, 0x303, 0x303, 0x303, 0x303, 0x303, 0x303, 0x303,
|
61 |
|
0x203, 0x303, 0x303, 0x303, 0x303, 0x303, 0x303, 0x303,
|
| 62 |
|
0x303, 0x303, 0x303, 0x303, 0x303, 0x403, 0x303, 0x303,
|
62 |
|
0x303, 0x303, 0x303, 0x303, 0x303, 0x403, 0x303, 0x303,
|
| 63 |
|
0x504, 0x604, 0x604, 0x604, 0x704, 0x000, 0x000, 0x000,
|
63 |
|
0x504, 0x604, 0x604, 0x604, 0x704, 0x000, 0x000, 0x000,
|
| 64 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
64 |
|
0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000, 0x000,
|
| 65 |
|
|
65 |
|
|
| 66 |
|
return first[static_cast<unsigned char>(c & 0x7F)];
|
66 |
|
return first[static_cast<unsigned char>(c & 0x7F)];
|
| 67 |
|
|
67 |
|
|
| 68 |
|
|
68 |
|
|
| 69 |
|
|
69 |
|
|
| 70 |
|
|
70 |
|
|
| 71 |
|
is_valid_utf8(const char* p, uint16_t first)
|
71 |
|
is_valid_utf8(const char* p, uint16_t first)
|
| 72 |
|
|
72 |
|
|
| 73 |
|
|
73 |
|
|
| 74 |
|
|
74 |
|
|
| 75 |
|
|
75 |
|
|
| 76 |
|
|
76 |
|
|
| 77 |
|
|
77 |
|
|
| 78 |
|
|
78 |
|
|
| 79 |
|
// 2 bytes, second byte [80, BF]
|
79 |
|
// 2 bytes, second byte [80, BF]
|
| 80 |
|
|
80 |
|
|
| 81 |
|
v = load_little_endian<2>(p);
|
81 |
|
v = load_little_endian<2>(p);
|
| 82 |
|
return (v & 0xC000) == 0x8000;
|
82 |
|
return (v & 0xC000) == 0x8000;
|
| 83 |
|
|
83 |
|
|
| 84 |
|
// 3 bytes, second byte [A0, BF]
|
84 |
|
// 3 bytes, second byte [A0, BF]
|
| 85 |
|
|
85 |
|
|
| 86 |
|
v = load_little_endian<3>(p);
|
86 |
|
v = load_little_endian<3>(p);
|
| 87 |
|
return (v & 0xC0E000) == 0x80A000;
|
87 |
|
return (v & 0xC0E000) == 0x80A000;
|
| 88 |
|
|
88 |
|
|
| 89 |
|
// 3 bytes, second byte [80, BF]
|
89 |
|
// 3 bytes, second byte [80, BF]
|
| 90 |
|
|
90 |
|
|
| 91 |
|
v = load_little_endian<3>(p);
|
91 |
|
v = load_little_endian<3>(p);
|
| 92 |
|
return (v & 0xC0C000) == 0x808000;
|
92 |
|
return (v & 0xC0C000) == 0x808000;
|
| 93 |
|
|
93 |
|
|
| 94 |
|
// 3 bytes, second byte [80, 9F]
|
94 |
|
// 3 bytes, second byte [80, 9F]
|
| 95 |
|
|
95 |
|
|
| 96 |
|
v = load_little_endian<3>(p);
|
96 |
|
v = load_little_endian<3>(p);
|
| 97 |
|
return (v & 0xC0E000) == 0x808000;
|
97 |
|
return (v & 0xC0E000) == 0x808000;
|
| 98 |
|
|
98 |
|
|
| 99 |
|
// 4 bytes, second byte [90, BF]
|
99 |
|
// 4 bytes, second byte [90, BF]
|
| 100 |
|
|
100 |
|
|
| 101 |
|
v = load_little_endian<4>(p);
|
101 |
|
v = load_little_endian<4>(p);
|
| 102 |
|
return (v & 0xC0C0FF00) + 0x7F7F7000 <= 0x2F00;
|
102 |
|
return (v & 0xC0C0FF00) + 0x7F7F7000 <= 0x2F00;
|
| 103 |
|
|
103 |
|
|
| 104 |
|
// 4 bytes, second byte [80, BF]
|
104 |
|
// 4 bytes, second byte [80, BF]
|
| 105 |
|
|
105 |
|
|
| 106 |
|
v = load_little_endian<4>(p);
|
106 |
|
v = load_little_endian<4>(p);
|
| 107 |
|
return (v & 0xC0C0C000) == 0x80808000;
|
107 |
|
return (v & 0xC0C0C000) == 0x80808000;
|
| 108 |
|
|
108 |
|
|
| 109 |
|
// 4 bytes, second byte [80, 8F]
|
109 |
|
// 4 bytes, second byte [80, 8F]
|
| 110 |
|
|
110 |
|
|
| 111 |
|
v = load_little_endian<4>(p);
|
111 |
|
v = load_little_endian<4>(p);
|
| 112 |
|
return (v & 0xC0C0F000) == 0x80808000;
|
112 |
|
return (v & 0xC0C0F000) == 0x80808000;
|
| 113 |
|
|
113 |
|
|
| 114 |
|
|
114 |
|
|
| 115 |
|
|
115 |
|
|
| 116 |
|
|
116 |
|
|
| 117 |
|
|
117 |
|
|
| 118 |
|
|
118 |
|
|
| 119 |
|
|
119 |
|
|
| 120 |
|
|
120 |
|
|
| 121 |
|
|
121 |
|
|
| 122 |
|
|
122 |
|
|
| 123 |
|
|
123 |
|
|
| 124 |
|
|
124 |
|
|
| 125 |
|
|
125 |
|
|
| 126 |
|
std::size_t remain) noexcept
|
126 |
|
std::size_t remain) noexcept
|
| 127 |
|
|
127 |
|
|
| 128 |
|
first_ = classify_utf8(*p );
|
128 |
|
first_ = classify_utf8(*p );
|
| 129 |
|
|
129 |
|
|
| 130 |
|
|
130 |
|
|
| 131 |
|
|
131 |
|
|
| 132 |
|
size_ = static_cast<uint8_t>(remain);
|
132 |
|
size_ = static_cast<uint8_t>(remain);
|
| 133 |
|
std::memcpy(seq_, p, size_);
|
133 |
|
std::memcpy(seq_, p, size_);
|
| 134 |
|
|
134 |
|
|
| 135 |
|
|
135 |
|
|
| 136 |
|
|
136 |
|
|
| 137 |
|
|
137 |
|
|
| 138 |
|
|
138 |
|
|
| 139 |
|
|
139 |
|
|
| 140 |
|
|
140 |
|
|
| 141 |
|
|
141 |
|
|
| 142 |
|
|
142 |
|
|
| 143 |
|
complete() const noexcept
|
143 |
|
complete() const noexcept
|
| 144 |
|
|
144 |
|
|
| 145 |
|
return size_ >= length();
|
145 |
|
return size_ >= length();
|
| 146 |
|
|
146 |
|
|
| 147 |
|
|
147 |
|
|
| 148 |
|
// returns true if complete
|
148 |
|
// returns true if complete
|
| 149 |
|
|
149 |
|
|
| 150 |
|
|
150 |
|
|
| 151 |
|
|
151 |
|
|
| 152 |
|
std::size_t remain) noexcept
|
152 |
|
std::size_t remain) noexcept
|
| 153 |
|
|
153 |
|
|
| 154 |
|
if(BOOST_JSON_UNLIKELY(needed() == 0))
|
154 |
|
if(BOOST_JSON_UNLIKELY(needed() == 0))
|
| 155 |
|
|
155 |
|
|
| 156 |
|
if(BOOST_JSON_LIKELY(remain >= needed()))
|
156 |
|
if(BOOST_JSON_LIKELY(remain >= needed()))
|
| 157 |
|
|
157 |
|
|
| 158 |
|
|
158 |
|
|
| 159 |
|
seq_ + size_, p, needed());
|
159 |
|
seq_ + size_, p, needed());
|
| 160 |
|
|
160 |
|
|
| 161 |
|
|
161 |
|
|
| 162 |
|
|
162 |
|
|
| 163 |
|
if(BOOST_JSON_LIKELY(remain > 0))
|
163 |
|
if(BOOST_JSON_LIKELY(remain > 0))
|
| 164 |
|
|
164 |
|
|
| 165 |
|
std::memcpy(seq_ + size_, p, remain);
|
165 |
|
std::memcpy(seq_ + size_, p, remain);
|
| 166 |
|
size_ += static_cast<uint8_t>(remain);
|
166 |
|
size_ += static_cast<uint8_t>(remain);
|
| 167 |
|
|
167 |
|
|
| 168 |
|
|
168 |
|
|
| 169 |
|
|
169 |
|
|
| 170 |
|
|
170 |
|
|
| 171 |
|
|
171 |
|
|
| 172 |
|
|
172 |
|
|
| 173 |
|
|
173 |
|
|
| 174 |
|
|
174 |
|
|
| 175 |
|
|
175 |
|
|
| 176 |
|
|
176 |
|
|
| 177 |
|
|
177 |
|
|
| 178 |
|
|
178 |
|
|
| 179 |
|
|
179 |
|
|
| 180 |
|
|
180 |
|
|
| 181 |
|
|
181 |
|
|
| 182 |
|
|
182 |
|
|
| 183 |
|
|
183 |
|
|
| 184 |
|
|
184 |
|
|
| 185 |
|
|
185 |
|
|
| 186 |
|
BOOST_ASSERT(size_ >= length());
|
186 |
|
BOOST_ASSERT(size_ >= length());
|
| 187 |
|
return is_valid_utf8(seq_, first_);
|
187 |
|
return is_valid_utf8(seq_, first_);
|
| 188 |
|
|
188 |
|
|
| 189 |
|
|
189 |
|
|
| 190 |
|
|
190 |
|
|
| 191 |
|
|
191 |
|
|
| 192 |
|
|
192 |
|
|
| 193 |
|
|
193 |
|
|
| 194 |
|
|
194 |
|
|
| 195 |
|
|
195 |
|
|