| 1 |
// Copyright 2006 Nemanja Trifunovic |
| 2 |
|
| 3 |
/* |
| 4 |
Permission is hereby granted, free of charge, to any person or organization |
| 5 |
obtaining a copy of the software and accompanying documentation covered by |
| 6 |
this license (the "Software") to use, reproduce, display, distribute, |
| 7 |
execute, and transmit the Software, and to prepare derivative works of the |
| 8 |
Software, and to permit third-parties to whom the Software is furnished to |
| 9 |
do so, all subject to the following: |
| 10 |
|
| 11 |
The copyright notices in the Software and this entire statement, including |
| 12 |
the above license grant, this restriction and the following disclaimer, |
| 13 |
must be included in all copies of the Software, in whole or in part, and |
| 14 |
all derivative works of the Software, unless such copies or derivative |
| 15 |
works are solely in the form of machine-executable object code generated by |
| 16 |
a source language processor. |
| 17 |
|
| 18 |
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| 19 |
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 20 |
FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT |
| 21 |
SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE |
| 22 |
FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE, |
| 23 |
ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER |
| 24 |
DEALINGS IN THE SOFTWARE. |
| 25 |
*/ |
| 26 |
|
| 27 |
|
| 28 |
#ifndef UTF8_FOR_CPP_CORE_H_2675DCD0_9480_4c0c_B92A_CC14C027B731 |
| 29 |
#define UTF8_FOR_CPP_CORE_H_2675DCD0_9480_4c0c_B92A_CC14C027B731 |
| 30 |
|
| 31 |
#include <iterator> |
| 32 |
|
| 33 |
// Determine the C++ standard version. |
| 34 |
// If the user defines UTF_CPP_CPLUSPLUS, use that. |
| 35 |
// Otherwise, trust the unreliable predefined macro __cplusplus |
| 36 |
|
| 37 |
#if !defined UTF_CPP_CPLUSPLUS |
| 38 |
#define UTF_CPP_CPLUSPLUS __cplusplus |
| 39 |
#endif |
| 40 |
|
| 41 |
#if UTF_CPP_CPLUSPLUS >= 201103L // C++ 11 or later |
| 42 |
#define UTF_CPP_OVERRIDE override |
| 43 |
#define UTF_CPP_NOEXCEPT noexcept |
| 44 |
#else // C++ 98/03 |
| 45 |
#define UTF_CPP_OVERRIDE |
| 46 |
#define UTF_CPP_NOEXCEPT throw() |
| 47 |
#endif // C++ 11 or later |
| 48 |
|
| 49 |
|
| 50 |
namespace utf8 |
| 51 |
{ |
| 52 |
// The typedefs for 8-bit, 16-bit and 32-bit unsigned integers |
| 53 |
// You may need to change them to match your system. |
| 54 |
// These typedefs have the same names as ones from cstdint, or boost/cstdint |
| 55 |
typedef unsigned char uint8_t; |
| 56 |
typedef unsigned short uint16_t; |
| 57 |
typedef unsigned int uint32_t; |
| 58 |
|
| 59 |
// Helper code - not intended to be directly called by the library users. May be changed at any time |
| 60 |
namespace internal |
| 61 |
{ |
| 62 |
// Unicode constants |
| 63 |
// Leading (high) surrogates: 0xd800 - 0xdbff |
| 64 |
// Trailing (low) surrogates: 0xdc00 - 0xdfff |
| 65 |
const uint16_t LEAD_SURROGATE_MIN = 0xd800u; |
| 66 |
const uint16_t LEAD_SURROGATE_MAX = 0xdbffu; |
| 67 |
const uint16_t TRAIL_SURROGATE_MIN = 0xdc00u; |
| 68 |
const uint16_t TRAIL_SURROGATE_MAX = 0xdfffu; |
| 69 |
const uint16_t LEAD_OFFSET = 0xd7c0u; // LEAD_SURROGATE_MIN - (0x10000 >> 10) |
| 70 |
const uint32_t SURROGATE_OFFSET = 0xfca02400u; // 0x10000u - (LEAD_SURROGATE_MIN << 10) - TRAIL_SURROGATE_MIN |
| 71 |
|
| 72 |
// Maximum valid value for a Unicode code point |
| 73 |
const uint32_t CODE_POINT_MAX = 0x0010ffffu; |
| 74 |
|
| 75 |
template<typename octet_type> |
| 76 |
inline uint8_t mask8(octet_type oc) |
| 77 |
{ |
| 78 |
return static_cast<uint8_t>(0xff & oc); |
| 79 |
} |
| 80 |
template<typename u16_type> |
| 81 |
inline uint16_t mask16(u16_type oc) |
| 82 |
{ |
| 83 |
return static_cast<uint16_t>(0xffff & oc); |
| 84 |
} |
| 85 |
template<typename octet_type> |
| 86 |
inline bool is_trail(octet_type oc) |
| 87 |
{ |
| 88 |
return ((utf8::internal::mask8(oc) >> 6) == 0x2); |
| 89 |
} |
| 90 |
|
| 91 |
template <typename u16> |
| 92 |
inline bool is_lead_surrogate(u16 cp) |
| 93 |
{ |
| 94 |
return (cp >= LEAD_SURROGATE_MIN && cp <= LEAD_SURROGATE_MAX); |
| 95 |
} |
| 96 |
|
| 97 |
template <typename u16> |
| 98 |
inline bool is_trail_surrogate(u16 cp) |
| 99 |
{ |
| 100 |
return (cp >= TRAIL_SURROGATE_MIN && cp <= TRAIL_SURROGATE_MAX); |
| 101 |
} |
| 102 |
|
| 103 |
template <typename u16> |
| 104 |
inline bool is_surrogate(u16 cp) |
| 105 |
{ |
| 106 |
return (cp >= LEAD_SURROGATE_MIN && cp <= TRAIL_SURROGATE_MAX); |
| 107 |
} |
| 108 |
|
| 109 |
template <typename u32> |
| 110 |
inline bool is_code_point_valid(u32 cp) |
| 111 |
{ |
| 112 |
return (cp <= CODE_POINT_MAX && !utf8::internal::is_surrogate(cp)); |
| 113 |
} |
| 114 |
|
| 115 |
template <typename octet_iterator> |
| 116 |
inline typename std::iterator_traits<octet_iterator>::difference_type |
| 117 |
sequence_length(octet_iterator lead_it) |
| 118 |
{ |
| 119 |
uint8_t lead = utf8::internal::mask8(*lead_it); |
| 120 |
if (lead < 0x80) |
| 121 |
return 1; |
| 122 |
else if ((lead >> 5) == 0x6) |
| 123 |
return 2; |
| 124 |
else if ((lead >> 4) == 0xe) |
| 125 |
return 3; |
| 126 |
else if ((lead >> 3) == 0x1e) |
| 127 |
return 4; |
| 128 |
else |
| 129 |
return 0; |
| 130 |
} |
| 131 |
|
| 132 |
template <typename octet_difference_type> |
| 133 |
inline bool is_overlong_sequence(uint32_t cp, octet_difference_type length) |
| 134 |
{ |
| 135 |
if (cp < 0x80) { |
| 136 |
if (length != 1) |
| 137 |
return true; |
| 138 |
} |
| 139 |
else if (cp < 0x800) { |
| 140 |
if (length != 2) |
| 141 |
return true; |
| 142 |
} |
| 143 |
else if (cp < 0x10000) { |
| 144 |
if (length != 3) |
| 145 |
return true; |
| 146 |
} |
| 147 |
|
| 148 |
return false; |
| 149 |
} |
| 150 |
|
| 151 |
enum utf_error {UTF8_OK, NOT_ENOUGH_ROOM, INVALID_LEAD, INCOMPLETE_SEQUENCE, OVERLONG_SEQUENCE, INVALID_CODE_POINT}; |
| 152 |
|
| 153 |
/// Helper for get_sequence_x |
| 154 |
template <typename octet_iterator> |
| 155 |
utf_error increase_safely(octet_iterator& it, octet_iterator end) |
| 156 |
{ |
| 157 |
if (++it == end) |
| 158 |
return NOT_ENOUGH_ROOM; |
| 159 |
|
| 160 |
if (!utf8::internal::is_trail(*it)) |
| 161 |
return INCOMPLETE_SEQUENCE; |
| 162 |
|
| 163 |
return UTF8_OK; |
| 164 |
} |
| 165 |
|
| 166 |
#define UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(IT, END) {utf_error ret = increase_safely(IT, END); if (ret != UTF8_OK) return ret;} |
| 167 |
|
| 168 |
/// get_sequence_x functions decode utf-8 sequences of the length x |
| 169 |
template <typename octet_iterator> |
| 170 |
utf_error get_sequence_1(octet_iterator& it, octet_iterator end, uint32_t& code_point) |
| 171 |
{ |
| 172 |
if (it == end) |
| 173 |
return NOT_ENOUGH_ROOM; |
| 174 |
|
| 175 |
code_point = utf8::internal::mask8(*it); |
| 176 |
|
| 177 |
return UTF8_OK; |
| 178 |
} |
| 179 |
|
| 180 |
template <typename octet_iterator> |
| 181 |
utf_error get_sequence_2(octet_iterator& it, octet_iterator end, uint32_t& code_point) |
| 182 |
{ |
| 183 |
if (it == end) |
| 184 |
return NOT_ENOUGH_ROOM; |
| 185 |
|
| 186 |
code_point = utf8::internal::mask8(*it); |
| 187 |
|
| 188 |
UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) |
| 189 |
|
| 190 |
code_point = ((code_point << 6) & 0x7ff) + ((*it) & 0x3f); |
| 191 |
|
| 192 |
return UTF8_OK; |
| 193 |
} |
| 194 |
|
| 195 |
template <typename octet_iterator> |
| 196 |
utf_error get_sequence_3(octet_iterator& it, octet_iterator end, uint32_t& code_point) |
| 197 |
{ |
| 198 |
if (it == end) |
| 199 |
return NOT_ENOUGH_ROOM; |
| 200 |
|
| 201 |
code_point = utf8::internal::mask8(*it); |
| 202 |
|
| 203 |
UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) |
| 204 |
|
| 205 |
code_point = ((code_point << 12) & 0xffff) + ((utf8::internal::mask8(*it) << 6) & 0xfff); |
| 206 |
|
| 207 |
UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) |
| 208 |
|
| 209 |
code_point += (*it) & 0x3f; |
| 210 |
|
| 211 |
return UTF8_OK; |
| 212 |
} |
| 213 |
|
| 214 |
template <typename octet_iterator> |
| 215 |
utf_error get_sequence_4(octet_iterator& it, octet_iterator end, uint32_t& code_point) |
| 216 |
{ |
| 217 |
if (it == end) |
| 218 |
return NOT_ENOUGH_ROOM; |
| 219 |
|
| 220 |
code_point = utf8::internal::mask8(*it); |
| 221 |
|
| 222 |
UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) |
| 223 |
|
| 224 |
code_point = ((code_point << 18) & 0x1fffff) + ((utf8::internal::mask8(*it) << 12) & 0x3ffff); |
| 225 |
|
| 226 |
UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) |
| 227 |
|
| 228 |
code_point += (utf8::internal::mask8(*it) << 6) & 0xfff; |
| 229 |
|
| 230 |
UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) |
| 231 |
|
| 232 |
code_point += (*it) & 0x3f; |
| 233 |
|
| 234 |
return UTF8_OK; |
| 235 |
} |
| 236 |
|
| 237 |
#undef UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR |
| 238 |
|
| 239 |
template <typename octet_iterator> |
| 240 |
utf_error validate_next(octet_iterator& it, octet_iterator end, uint32_t& code_point) |
| 241 |
{ |
| 242 |
if (it == end) |
| 243 |
return NOT_ENOUGH_ROOM; |
| 244 |
|
| 245 |
// Save the original value of it so we can go back in case of failure |
| 246 |
// Of course, it does not make much sense with i.e. stream iterators |
| 247 |
octet_iterator original_it = it; |
| 248 |
|
| 249 |
uint32_t cp = 0; |
| 250 |
// Determine the sequence length based on the lead octet |
| 251 |
typedef typename std::iterator_traits<octet_iterator>::difference_type octet_difference_type; |
| 252 |
const octet_difference_type length = utf8::internal::sequence_length(it); |
| 253 |
|
| 254 |
// Get trail octets and calculate the code point |
| 255 |
utf_error err = UTF8_OK; |
| 256 |
switch (length) { |
| 257 |
case 0: |
| 258 |
return INVALID_LEAD; |
| 259 |
case 1: |
| 260 |
err = utf8::internal::get_sequence_1(it, end, cp); |
| 261 |
break; |
| 262 |
case 2: |
| 263 |
err = utf8::internal::get_sequence_2(it, end, cp); |
| 264 |
break; |
| 265 |
case 3: |
| 266 |
err = utf8::internal::get_sequence_3(it, end, cp); |
| 267 |
break; |
| 268 |
case 4: |
| 269 |
err = utf8::internal::get_sequence_4(it, end, cp); |
| 270 |
break; |
| 271 |
} |
| 272 |
|
| 273 |
if (err == UTF8_OK) { |
| 274 |
// Decoding succeeded. Now, security checks... |
| 275 |
if (utf8::internal::is_code_point_valid(cp)) { |
| 276 |
if (!utf8::internal::is_overlong_sequence(cp, length)){ |
| 277 |
// Passed! Return here. |
| 278 |
code_point = cp; |
| 279 |
++it; |
| 280 |
return UTF8_OK; |
| 281 |
} |
| 282 |
else |
| 283 |
err = OVERLONG_SEQUENCE; |
| 284 |
} |
| 285 |
else |
| 286 |
err = INVALID_CODE_POINT; |
| 287 |
} |
| 288 |
|
| 289 |
// Failure branch - restore the original value of the iterator |
| 290 |
it = original_it; |
| 291 |
return err; |
| 292 |
} |
| 293 |
|
| 294 |
template <typename octet_iterator> |
| 295 |
inline utf_error validate_next(octet_iterator& it, octet_iterator end) { |
| 296 |
uint32_t ignored; |
| 297 |
return utf8::internal::validate_next(it, end, ignored); |
| 298 |
} |
| 299 |
|
| 300 |
// Internal implementation of both checked and unchecked append() function |
| 301 |
// This function will be invoked by the overloads below, as they will know |
| 302 |
// the octet_type. |
| 303 |
template <typename octet_iterator, typename octet_type> |
| 304 |
octet_iterator append(uint32_t cp, octet_iterator result) { |
| 305 |
if (cp < 0x80) // one octet |
| 306 |
*(result++) = static_cast<octet_type>(cp); |
| 307 |
else if (cp < 0x800) { // two octets |
| 308 |
*(result++) = static_cast<octet_type>((cp >> 6) | 0xc0); |
| 309 |
*(result++) = static_cast<octet_type>((cp & 0x3f) | 0x80); |
| 310 |
} |
| 311 |
else if (cp < 0x10000) { // three octets |
| 312 |
*(result++) = static_cast<octet_type>((cp >> 12) | 0xe0); |
| 313 |
*(result++) = static_cast<octet_type>(((cp >> 6) & 0x3f) | 0x80); |
| 314 |
*(result++) = static_cast<octet_type>((cp & 0x3f) | 0x80); |
| 315 |
} |
| 316 |
else { // four octets |
| 317 |
*(result++) = static_cast<octet_type>((cp >> 18) | 0xf0); |
| 318 |
*(result++) = static_cast<octet_type>(((cp >> 12) & 0x3f)| 0x80); |
| 319 |
*(result++) = static_cast<octet_type>(((cp >> 6) & 0x3f) | 0x80); |
| 320 |
*(result++) = static_cast<octet_type>((cp & 0x3f) | 0x80); |
| 321 |
} |
| 322 |
return result; |
| 323 |
} |
| 324 |
|
| 325 |
// One of the following overloads will be invoked from the API calls |
| 326 |
|
| 327 |
// A simple (but dangerous) case: the caller appends byte(s) to a char array |
| 328 |
inline char* append(uint32_t cp, char* result) { |
| 329 |
return append<char*, char>(cp, result); |
| 330 |
} |
| 331 |
|
| 332 |
// Hopefully, most common case: the caller uses back_inserter |
| 333 |
// i.e. append(cp, std::back_inserter(str)); |
| 334 |
template<typename container_type> |
| 335 |
std::back_insert_iterator<container_type> append |
| 336 |
(uint32_t cp, std::back_insert_iterator<container_type> result) { |
| 337 |
return append<std::back_insert_iterator<container_type>, |
| 338 |
typename container_type::value_type>(cp, result); |
| 339 |
} |
| 340 |
|
| 341 |
// The caller uses some other kind of output operator - not covered above |
| 342 |
// Note that in this case we are not able to determine octet_type |
| 343 |
// so we assume it's uint_8; that can cause a conversion warning if we are wrong. |
| 344 |
template <typename octet_iterator> |
| 345 |
octet_iterator append(uint32_t cp, octet_iterator result) { |
| 346 |
return append<octet_iterator, uint8_t>(cp, result); |
| 347 |
} |
| 348 |
|
| 349 |
} // namespace internal |
| 350 |
|
| 351 |
/// The library API - functions intended to be called by the users |
| 352 |
|
| 353 |
// Byte order mark |
| 354 |
const uint8_t bom[] = {0xef, 0xbb, 0xbf}; |
| 355 |
|
| 356 |
template <typename octet_iterator> |
| 357 |
octet_iterator find_invalid(octet_iterator start, octet_iterator end) |
| 358 |
{ |
| 359 |
octet_iterator result = start; |
| 360 |
while (result != end) { |
| 361 |
utf8::internal::utf_error err_code = utf8::internal::validate_next(result, end); |
| 362 |
if (err_code != internal::UTF8_OK) |
| 363 |
return result; |
| 364 |
} |
| 365 |
return result; |
| 366 |
} |
| 367 |
|
| 368 |
template <typename octet_iterator> |
| 369 |
inline bool is_valid(octet_iterator start, octet_iterator end) |
| 370 |
{ |
| 371 |
return (utf8::find_invalid(start, end) == end); |
| 372 |
} |
| 373 |
|
| 374 |
template <typename octet_iterator> |
| 375 |
inline bool starts_with_bom (octet_iterator it, octet_iterator end) |
| 376 |
{ |
| 377 |
return ( |
| 378 |
((it != end) && (utf8::internal::mask8(*it++)) == bom[0]) && |
| 379 |
((it != end) && (utf8::internal::mask8(*it++)) == bom[1]) && |
| 380 |
((it != end) && (utf8::internal::mask8(*it)) == bom[2]) |
| 381 |
); |
| 382 |
} |
| 383 |
} // namespace utf8 |
| 384 |
|
| 385 |
#endif // header guard |
| 386 |
|
| 387 |
|