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 namespace utf8 34 { 35 // The typedefs for 8-bit, 16-bit and 32-bit unsigned integers 36 // You may need to change them to match your system. 37 // These typedefs have the same names as ones from cstdint, or boost/cstdint 38 typedef unsigned char uint8_t; 39 typedef unsigned short uint16_t; 40 typedef unsigned int uint32_t; 41 42 // Helper code - not intended to be directly called by the library users. May be changed at any time 43 namespace internal 44 { 45 // Unicode constants 46 // Leading (high) surrogates: 0xd800 - 0xdbff 47 // Trailing (low) surrogates: 0xdc00 - 0xdfff 48 const uint16_t LEAD_SURROGATE_MIN = 0xd800u; 49 const uint16_t LEAD_SURROGATE_MAX = 0xdbffu; 50 const uint16_t TRAIL_SURROGATE_MIN = 0xdc00u; 51 const uint16_t TRAIL_SURROGATE_MAX = 0xdfffu; 52 const uint16_t LEAD_OFFSET = LEAD_SURROGATE_MIN - (0x10000 >> 10); 53 const uint32_t SURROGATE_OFFSET = 0x10000u - (LEAD_SURROGATE_MIN << 10) - TRAIL_SURROGATE_MIN; 54 55 // Maximum valid value for a Unicode code point 56 const uint32_t CODE_POINT_MAX = 0x0010ffffu; 57 58 template<typename octet_type> mask8(octet_type oc)59 inline uint8_t mask8(octet_type oc) 60 { 61 return static_cast<uint8_t>(0xff & oc); 62 } 63 template<typename u16_type> mask16(u16_type oc)64 inline uint16_t mask16(u16_type oc) 65 { 66 return static_cast<uint16_t>(0xffff & oc); 67 } 68 template<typename octet_type> is_trail(octet_type oc)69 inline bool is_trail(octet_type oc) 70 { 71 return ((utf8::internal::mask8(oc) >> 6) == 0x2); 72 } 73 74 template <typename u16> is_lead_surrogate(u16 cp)75 inline bool is_lead_surrogate(u16 cp) 76 { 77 return (cp >= LEAD_SURROGATE_MIN && cp <= LEAD_SURROGATE_MAX); 78 } 79 80 template <typename u16> is_trail_surrogate(u16 cp)81 inline bool is_trail_surrogate(u16 cp) 82 { 83 return (cp >= TRAIL_SURROGATE_MIN && cp <= TRAIL_SURROGATE_MAX); 84 } 85 86 template <typename u16> is_surrogate(u16 cp)87 inline bool is_surrogate(u16 cp) 88 { 89 return (cp >= LEAD_SURROGATE_MIN && cp <= TRAIL_SURROGATE_MAX); 90 } 91 92 template <typename u32> is_code_point_valid(u32 cp)93 inline bool is_code_point_valid(u32 cp) 94 { 95 return (cp <= CODE_POINT_MAX && !utf8::internal::is_surrogate(cp)); 96 } 97 98 template <typename octet_iterator> 99 inline typename std::iterator_traits<octet_iterator>::difference_type sequence_length(octet_iterator lead_it)100 sequence_length(octet_iterator lead_it) 101 { 102 uint8_t lead = utf8::internal::mask8(*lead_it); 103 if (lead < 0x80) 104 return 1; 105 else if ((lead >> 5) == 0x6) 106 return 2; 107 else if ((lead >> 4) == 0xe) 108 return 3; 109 else if ((lead >> 3) == 0x1e) 110 return 4; 111 else 112 return 0; 113 } 114 115 template <typename octet_difference_type> is_overlong_sequence(uint32_t cp,octet_difference_type length)116 inline bool is_overlong_sequence(uint32_t cp, octet_difference_type length) 117 { 118 if (cp < 0x80) { 119 if (length != 1) 120 return true; 121 } 122 else if (cp < 0x800) { 123 if (length != 2) 124 return true; 125 } 126 else if (cp < 0x10000) { 127 if (length != 3) 128 return true; 129 } 130 131 return false; 132 } 133 134 enum utf_error {UTF8_OK, NOT_ENOUGH_ROOM, INVALID_LEAD, INCOMPLETE_SEQUENCE, OVERLONG_SEQUENCE, INVALID_CODE_POINT}; 135 136 /// Helper for get_sequence_x 137 template <typename octet_iterator> increase_safely(octet_iterator & it,octet_iterator end)138 utf_error increase_safely(octet_iterator& it, octet_iterator end) 139 { 140 if (++it == end) 141 return NOT_ENOUGH_ROOM; 142 143 if (!utf8::internal::is_trail(*it)) 144 return INCOMPLETE_SEQUENCE; 145 146 return UTF8_OK; 147 } 148 149 #define UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(IT, END) {utf_error ret = increase_safely(IT, END); if (ret != UTF8_OK) return ret;} 150 151 /// get_sequence_x functions decode utf-8 sequences of the length x 152 template <typename octet_iterator> get_sequence_1(octet_iterator & it,octet_iterator end,uint32_t & code_point)153 utf_error get_sequence_1(octet_iterator& it, octet_iterator end, uint32_t& code_point) 154 { 155 if (it == end) 156 return NOT_ENOUGH_ROOM; 157 158 code_point = utf8::internal::mask8(*it); 159 160 return UTF8_OK; 161 } 162 163 template <typename octet_iterator> get_sequence_2(octet_iterator & it,octet_iterator end,uint32_t & code_point)164 utf_error get_sequence_2(octet_iterator& it, octet_iterator end, uint32_t& code_point) 165 { 166 if (it == end) 167 return NOT_ENOUGH_ROOM; 168 169 code_point = utf8::internal::mask8(*it); 170 171 UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) 172 173 code_point = ((code_point << 6) & 0x7ff) + ((*it) & 0x3f); 174 175 return UTF8_OK; 176 } 177 178 template <typename octet_iterator> get_sequence_3(octet_iterator & it,octet_iterator end,uint32_t & code_point)179 utf_error get_sequence_3(octet_iterator& it, octet_iterator end, uint32_t& code_point) 180 { 181 if (it == end) 182 return NOT_ENOUGH_ROOM; 183 184 code_point = utf8::internal::mask8(*it); 185 186 UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) 187 188 code_point = ((code_point << 12) & 0xffff) + ((utf8::internal::mask8(*it) << 6) & 0xfff); 189 190 UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) 191 192 code_point += (*it) & 0x3f; 193 194 return UTF8_OK; 195 } 196 197 template <typename octet_iterator> get_sequence_4(octet_iterator & it,octet_iterator end,uint32_t & code_point)198 utf_error get_sequence_4(octet_iterator& it, octet_iterator end, uint32_t& code_point) 199 { 200 if (it == end) 201 return NOT_ENOUGH_ROOM; 202 203 code_point = utf8::internal::mask8(*it); 204 205 UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) 206 207 code_point = ((code_point << 18) & 0x1fffff) + ((utf8::internal::mask8(*it) << 12) & 0x3ffff); 208 209 UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) 210 211 code_point += (utf8::internal::mask8(*it) << 6) & 0xfff; 212 213 UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR(it, end) 214 215 code_point += (*it) & 0x3f; 216 217 return UTF8_OK; 218 } 219 220 #undef UTF8_CPP_INCREASE_AND_RETURN_ON_ERROR 221 222 template <typename octet_iterator> validate_next(octet_iterator & it,octet_iterator end,uint32_t & code_point)223 utf_error validate_next(octet_iterator& it, octet_iterator end, uint32_t& code_point) 224 { 225 if (it == end) 226 return NOT_ENOUGH_ROOM; 227 228 // Save the original value of it so we can go back in case of failure 229 // Of course, it does not make much sense with i.e. stream iterators 230 octet_iterator original_it = it; 231 232 uint32_t cp = 0; 233 // Determine the sequence length based on the lead octet 234 typedef typename std::iterator_traits<octet_iterator>::difference_type octet_difference_type; 235 const octet_difference_type length = utf8::internal::sequence_length(it); 236 237 // Get trail octets and calculate the code point 238 utf_error err = UTF8_OK; 239 switch (length) { 240 case 0: 241 return INVALID_LEAD; 242 case 1: 243 err = utf8::internal::get_sequence_1(it, end, cp); 244 break; 245 case 2: 246 err = utf8::internal::get_sequence_2(it, end, cp); 247 break; 248 case 3: 249 err = utf8::internal::get_sequence_3(it, end, cp); 250 break; 251 case 4: 252 err = utf8::internal::get_sequence_4(it, end, cp); 253 break; 254 } 255 256 if (err == UTF8_OK) { 257 // Decoding succeeded. Now, security checks... 258 if (utf8::internal::is_code_point_valid(cp)) { 259 if (!utf8::internal::is_overlong_sequence(cp, length)){ 260 // Passed! Return here. 261 code_point = cp; 262 ++it; 263 return UTF8_OK; 264 } 265 else 266 err = OVERLONG_SEQUENCE; 267 } 268 else 269 err = INVALID_CODE_POINT; 270 } 271 272 // Failure branch - restore the original value of the iterator 273 it = original_it; 274 return err; 275 } 276 277 template <typename octet_iterator> validate_next(octet_iterator & it,octet_iterator end)278 inline utf_error validate_next(octet_iterator& it, octet_iterator end) { 279 uint32_t ignored; 280 return utf8::internal::validate_next(it, end, ignored); 281 } 282 283 } // namespace internal 284 285 /// The library API - functions intended to be called by the users 286 287 // Byte order mark 288 const uint8_t bom[] = {0xef, 0xbb, 0xbf}; 289 290 template <typename octet_iterator> find_invalid(octet_iterator start,octet_iterator end)291 octet_iterator find_invalid(octet_iterator start, octet_iterator end) 292 { 293 octet_iterator result = start; 294 while (result != end) { 295 utf8::internal::utf_error err_code = utf8::internal::validate_next(result, end); 296 if (err_code != internal::UTF8_OK) 297 return result; 298 } 299 return result; 300 } 301 302 template <typename octet_iterator> is_valid(octet_iterator start,octet_iterator end)303 inline bool is_valid(octet_iterator start, octet_iterator end) 304 { 305 return (utf8::find_invalid(start, end) == end); 306 } 307 308 template <typename octet_iterator> starts_with_bom(octet_iterator it,octet_iterator end)309 inline bool starts_with_bom (octet_iterator it, octet_iterator end) 310 { 311 return ( 312 ((it != end) && (utf8::internal::mask8(*it++)) == bom[0]) && 313 ((it != end) && (utf8::internal::mask8(*it++)) == bom[1]) && 314 ((it != end) && (utf8::internal::mask8(*it)) == bom[2]) 315 ); 316 } 317 318 //Deprecated in release 2.3 319 template <typename octet_iterator> is_bom(octet_iterator it)320 inline bool is_bom (octet_iterator it) 321 { 322 return ( 323 (utf8::internal::mask8(*it++)) == bom[0] && 324 (utf8::internal::mask8(*it++)) == bom[1] && 325 (utf8::internal::mask8(*it)) == bom[2] 326 ); 327 } 328 } // namespace utf8 329 330 #endif // header guard 331 332 333