1 // © 2016 and later: Unicode, Inc. and others. 2 // License & terms of use: http://www.unicode.org/copyright.html 3 /* 4 ****************************************************************************** 5 * 6 * Copyright (C) 1999-2015, International Business Machines 7 * Corporation and others. All Rights Reserved. 8 * 9 ****************************************************************************** 10 */ 11 12 #ifndef BASE_THIRD_PARTY_ICU_ICU_UTF_H_ 13 #define BASE_THIRD_PARTY_ICU_ICU_UTF_H_ 14 15 #include <stdint.h> 16 17 namespace base_icu { 18 19 // source/common/unicode/umachine.h 20 21 /** The ICU boolean type @stable ICU 2.0 */ 22 typedef int8_t UBool; 23 24 /** 25 * Define UChar32 as a type for single Unicode code points. 26 * UChar32 is a signed 32-bit integer (same as int32_t). 27 * 28 * The Unicode code point range is 0..0x10ffff. 29 * All other values (negative or >=0x110000) are illegal as Unicode code points. 30 * They may be used as sentinel values to indicate "done", "error" 31 * or similar non-code point conditions. 32 * 33 * Before ICU 2.4 (Jitterbug 2146), UChar32 was defined 34 * to be wchar_t if that is 32 bits wide (wchar_t may be signed or unsigned) 35 * or else to be uint32_t. 36 * That is, the definition of UChar32 was platform-dependent. 37 * 38 * @see U_SENTINEL 39 * @stable ICU 2.4 40 */ 41 typedef int32_t UChar32; 42 43 /** 44 * This value is intended for sentinel values for APIs that 45 * (take or) return single code points (UChar32). 46 * It is outside of the Unicode code point range 0..0x10ffff. 47 * 48 * For example, a "done" or "error" value in a new API 49 * could be indicated with U_SENTINEL. 50 * 51 * ICU APIs designed before ICU 2.4 usually define service-specific "done" 52 * values, mostly 0xffff. 53 * Those may need to be distinguished from 54 * actual U+ffff text contents by calling functions like 55 * CharacterIterator::hasNext() or UnicodeString::length(). 56 * 57 * @return -1 58 * @see UChar32 59 * @stable ICU 2.4 60 */ 61 #define CBU_SENTINEL (-1) 62 63 // source/common/unicode/utf.h 64 65 /** 66 * Is this code point a Unicode noncharacter? 67 * @param c 32-bit code point 68 * @return TRUE or FALSE 69 * @stable ICU 2.4 70 */ 71 #define CBU_IS_UNICODE_NONCHAR(c) \ 72 ((c) >= 0xfdd0 && ((c) <= 0xfdef || ((c)&0xfffe) == 0xfffe) && \ 73 (c) <= 0x10ffff) 74 75 /** 76 * Is c a Unicode code point value (0..U+10ffff) 77 * that can be assigned a character? 78 * 79 * Code points that are not characters include: 80 * - single surrogate code points (U+d800..U+dfff, 2048 code points) 81 * - the last two code points on each plane (U+__fffe and U+__ffff, 34 code 82 * points) - U+fdd0..U+fdef (new with Unicode 3.1, 32 code points) - the highest 83 * Unicode code point value is U+10ffff 84 * 85 * This means that all code points below U+d800 are character code points, 86 * and that boundary is tested first for performance. 87 * 88 * @param c 32-bit code point 89 * @return TRUE or FALSE 90 * @stable ICU 2.4 91 */ 92 #define CBU_IS_UNICODE_CHAR(c) \ 93 ((uint32_t)(c) < 0xd800 || \ 94 (0xdfff < (c) && (c) <= 0x10ffff && !CBU_IS_UNICODE_NONCHAR(c))) 95 96 /** 97 * Is this code point a surrogate (U+d800..U+dfff)? 98 * @param c 32-bit code point 99 * @return TRUE or FALSE 100 * @stable ICU 2.4 101 */ 102 #define CBU_IS_SURROGATE(c) (((c)&0xfffff800) == 0xd800) 103 104 /** 105 * Assuming c is a surrogate code point (U_IS_SURROGATE(c)), 106 * is it a lead surrogate? 107 * @param c 32-bit code point 108 * @return TRUE or FALSE 109 * @stable ICU 2.4 110 */ 111 #define CBU_IS_SURROGATE_LEAD(c) (((c)&0x400) == 0) 112 113 // source/common/unicode/utf8.h 114 115 /** 116 * Internal bit vector for 3-byte UTF-8 validity check, for use in 117 * U8_IS_VALID_LEAD3_AND_T1. Each bit indicates whether one lead byte + first 118 * trail byte pair starts a valid sequence. Lead byte E0..EF bits 3..0 are used 119 * as byte index, first trail byte bits 7..5 are used as bit index into that 120 * byte. 121 * @see U8_IS_VALID_LEAD3_AND_T1 122 * @internal 123 */ 124 #define CBU8_LEAD3_T1_BITS \ 125 "\x20\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x10\x30\x30" 126 127 /** 128 * Internal 3-byte UTF-8 validity check. 129 * Non-zero if lead byte E0..EF and first trail byte 00..FF start a valid 130 * sequence. 131 * @internal 132 */ 133 #define CBU8_IS_VALID_LEAD3_AND_T1(lead, t1) \ 134 (CBU8_LEAD3_T1_BITS[(lead)&0xf] & (1 << ((uint8_t)(t1) >> 5))) 135 136 /** 137 * Internal bit vector for 4-byte UTF-8 validity check, for use in 138 * U8_IS_VALID_LEAD4_AND_T1. Each bit indicates whether one lead byte + first 139 * trail byte pair starts a valid sequence. First trail byte bits 7..4 are used 140 * as byte index, lead byte F0..F4 bits 2..0 are used as bit index into that 141 * byte. 142 * @see U8_IS_VALID_LEAD4_AND_T1 143 * @internal 144 */ 145 #define CBU8_LEAD4_T1_BITS \ 146 "\x00\x00\x00\x00\x00\x00\x00\x00\x1E\x0F\x0F\x0F\x00\x00\x00\x00" 147 148 /** 149 * Internal 4-byte UTF-8 validity check. 150 * Non-zero if lead byte F0..F4 and first trail byte 00..FF start a valid 151 * sequence. 152 * @internal 153 */ 154 #define CBU8_IS_VALID_LEAD4_AND_T1(lead, t1) \ 155 (CBU8_LEAD4_T1_BITS[(uint8_t)(t1) >> 4] & (1 << ((lead)&7))) 156 157 /** 158 * Function for handling "next code point" with error-checking. 159 * 160 * This is internal since it is not meant to be called directly by external clie 161 nts; 162 * however it is U_STABLE (not U_INTERNAL) since it is called by public macros i 163 n this 164 * file and thus must remain stable, and should not be hidden when other interna 165 l 166 * functions are hidden (otherwise public macros would fail to compile). 167 * @internal 168 */ 169 UChar32 utf8_nextCharSafeBody(const uint8_t* s, 170 int32_t* pi, 171 int32_t length, 172 ::base_icu::UChar32 c, 173 ::base_icu::UBool strict); 174 175 /** 176 * Does this code unit (byte) encode a code point by itself (US-ASCII 0..0x7f)? 177 * @param c 8-bit code unit (byte) 178 * @return TRUE or FALSE 179 * @stable ICU 2.4 180 */ 181 #define CBU8_IS_SINGLE(c) (((c)&0x80) == 0) 182 183 /** 184 * Is this code unit (byte) a UTF-8 lead byte? (0xC2..0xF4) 185 * @param c 8-bit code unit (byte) 186 * @return TRUE or FALSE 187 * @stable ICU 2.4 188 */ 189 #define CBU8_IS_LEAD(c) ((uint8_t)((c)-0xc2) <= 0x32) 190 191 /** 192 * Is this code unit (byte) a UTF-8 trail byte? (0x80..0xBF) 193 * @param c 8-bit code unit (byte) 194 * @return TRUE or FALSE 195 * @stable ICU 2.4 196 */ 197 #define CBU8_IS_TRAIL(c) ((int8_t)(c) < -0x40) 198 199 /** 200 * How many code units (bytes) are used for the UTF-8 encoding 201 * of this Unicode code point? 202 * @param c 32-bit code point 203 * @return 1..4, or 0 if c is a surrogate or not a Unicode code point 204 * @stable ICU 2.4 205 */ 206 #define CBU8_LENGTH(c) \ 207 ((uint32_t)(c) <= 0x7f \ 208 ? 1 \ 209 : ((uint32_t)(c) <= 0x7ff \ 210 ? 2 \ 211 : ((uint32_t)(c) <= 0xd7ff \ 212 ? 3 \ 213 : ((uint32_t)(c) <= 0xdfff || (uint32_t)(c) > 0x10ffff \ 214 ? 0 \ 215 : ((uint32_t)(c) <= 0xffff ? 3 : 4))))) 216 217 /** 218 * The maximum number of UTF-8 code units (bytes) per Unicode code point 219 * (U+0000..U+10ffff). 220 * @return 4 221 * @stable ICU 2.4 222 */ 223 #define CBU8_MAX_LENGTH 4 224 225 /** 226 * Get a code point from a string at a code point boundary offset, 227 * and advance the offset to the next code point boundary. 228 * (Post-incrementing forward iteration.) 229 * "Safe" macro, checks for illegal sequences and for string boundaries. 230 * 231 * The length can be negative for a NUL-terminated string. 232 * 233 * The offset may point to the lead byte of a multi-byte sequence, 234 * in which case the macro will read the whole sequence. 235 * If the offset points to a trail byte or an illegal UTF-8 sequence, then 236 * c is set to a negative value. 237 * 238 * @param s const uint8_t * string 239 * @param i int32_t string offset, must be i<length 240 * @param length int32_t string length 241 * @param c output UChar32 variable, set to <0 in case of an error 242 * @see U8_NEXT_UNSAFE 243 * @stable ICU 2.4 244 */ 245 #define CBU8_NEXT(s, i, length, c) \ 246 { \ 247 (c) = (uint8_t)(s)[(i)++]; \ 248 if (!CBU8_IS_SINGLE(c)) { \ 249 uint8_t __t1, __t2; \ 250 if (/* handle U+0800..U+FFFF inline */ \ 251 (0xe0 <= (c) && (c) < 0xf0) && \ 252 (((i) + 1) < (length) || (length) < 0) && \ 253 CBU8_IS_VALID_LEAD3_AND_T1((c), __t1 = (s)[i]) && \ 254 (__t2 = (s)[(i) + 1] - 0x80) <= 0x3f) { \ 255 (c) = (((c)&0xf) << 12) | ((__t1 & 0x3f) << 6) | __t2; \ 256 (i) += 2; \ 257 } else if (/* handle U+0080..U+07FF inline */ \ 258 ((c) < 0xe0 && (c) >= 0xc2) && ((i) != (length)) && \ 259 (__t1 = (s)[i] - 0x80) <= 0x3f) { \ 260 (c) = (((c)&0x1f) << 6) | __t1; \ 261 ++(i); \ 262 } else { \ 263 /* function call for "complicated" and error cases */ \ 264 (c) = ::base_icu::utf8_nextCharSafeBody((const uint8_t*)s, &(i), \ 265 (length), c, -1); \ 266 } \ 267 } \ 268 } 269 270 /** 271 * Append a code point to a string, overwriting 1 to 4 bytes. 272 * The offset points to the current end of the string contents 273 * and is advanced (post-increment). 274 * "Unsafe" macro, assumes a valid code point and sufficient space in the 275 * string. Otherwise, the result is undefined. 276 * 277 * @param s const uint8_t * string buffer 278 * @param i string offset 279 * @param c code point to append 280 * @see U8_APPEND 281 * @stable ICU 2.4 282 */ 283 #define CBU8_APPEND_UNSAFE(s, i, c) \ 284 { \ 285 if ((uint32_t)(c) <= 0x7f) { \ 286 (s)[(i)++] = (uint8_t)(c); \ 287 } else { \ 288 if ((uint32_t)(c) <= 0x7ff) { \ 289 (s)[(i)++] = (uint8_t)(((c) >> 6) | 0xc0); \ 290 } else { \ 291 if ((uint32_t)(c) <= 0xffff) { \ 292 (s)[(i)++] = (uint8_t)(((c) >> 12) | 0xe0); \ 293 } else { \ 294 (s)[(i)++] = (uint8_t)(((c) >> 18) | 0xf0); \ 295 (s)[(i)++] = (uint8_t)((((c) >> 12) & 0x3f) | 0x80); \ 296 } \ 297 (s)[(i)++] = (uint8_t)((((c) >> 6) & 0x3f) | 0x80); \ 298 } \ 299 (s)[(i)++] = (uint8_t)(((c)&0x3f) | 0x80); \ 300 } \ 301 } 302 303 // source/common/unicode/utf16.h 304 305 /** 306 * Does this code unit alone encode a code point (BMP, not a surrogate)? 307 * @param c 16-bit code unit 308 * @return TRUE or FALSE 309 * @stable ICU 2.4 310 */ 311 #define CBU16_IS_SINGLE(c) !CBU_IS_SURROGATE(c) 312 313 /** 314 * Is this code unit a lead surrogate (U+d800..U+dbff)? 315 * @param c 16-bit code unit 316 * @return TRUE or FALSE 317 * @stable ICU 2.4 318 */ 319 #define CBU16_IS_LEAD(c) (((c)&0xfffffc00) == 0xd800) 320 321 /** 322 * Is this code unit a trail surrogate (U+dc00..U+dfff)? 323 * @param c 16-bit code unit 324 * @return TRUE or FALSE 325 * @stable ICU 2.4 326 */ 327 #define CBU16_IS_TRAIL(c) (((c)&0xfffffc00) == 0xdc00) 328 329 /** 330 * Is this code unit a surrogate (U+d800..U+dfff)? 331 * @param c 16-bit code unit 332 * @return TRUE or FALSE 333 * @stable ICU 2.4 334 */ 335 #define CBU16_IS_SURROGATE(c) CBU_IS_SURROGATE(c) 336 337 /** 338 * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)), 339 * is it a lead surrogate? 340 * @param c 16-bit code unit 341 * @return TRUE or FALSE 342 * @stable ICU 2.4 343 */ 344 #define CBU16_IS_SURROGATE_LEAD(c) (((c)&0x400) == 0) 345 346 /** 347 * Helper constant for U16_GET_SUPPLEMENTARY. 348 * @internal 349 */ 350 #define CBU16_SURROGATE_OFFSET ((0xd800 << 10UL) + 0xdc00 - 0x10000) 351 352 /** 353 * Get a supplementary code point value (U+10000..U+10ffff) 354 * from its lead and trail surrogates. 355 * The result is undefined if the input values are not 356 * lead and trail surrogates. 357 * 358 * @param lead lead surrogate (U+d800..U+dbff) 359 * @param trail trail surrogate (U+dc00..U+dfff) 360 * @return supplementary code point (U+10000..U+10ffff) 361 * @stable ICU 2.4 362 */ 363 #define CBU16_GET_SUPPLEMENTARY(lead, trail) \ 364 (((::base_icu::UChar32)(lead) << 10UL) + \ 365 (::base_icu::UChar32)(trail)-CBU16_SURROGATE_OFFSET) 366 367 /** 368 * Get the lead surrogate (0xd800..0xdbff) for a 369 * supplementary code point (0x10000..0x10ffff). 370 * @param supplementary 32-bit code point (U+10000..U+10ffff) 371 * @return lead surrogate (U+d800..U+dbff) for supplementary 372 * @stable ICU 2.4 373 */ 374 #define CBU16_LEAD(supplementary) \ 375 (::base_icu::UChar)(((supplementary) >> 10) + 0xd7c0) 376 377 /** 378 * Get the trail surrogate (0xdc00..0xdfff) for a 379 * supplementary code point (0x10000..0x10ffff). 380 * @param supplementary 32-bit code point (U+10000..U+10ffff) 381 * @return trail surrogate (U+dc00..U+dfff) for supplementary 382 * @stable ICU 2.4 383 */ 384 #define CBU16_TRAIL(supplementary) \ 385 (::base_icu::UChar)(((supplementary)&0x3ff) | 0xdc00) 386 387 /** 388 * How many 16-bit code units are used to encode this Unicode code point? (1 or 389 * 2) The result is not defined if c is not a Unicode code point 390 * (U+0000..U+10ffff). 391 * @param c 32-bit code point 392 * @return 1 or 2 393 * @stable ICU 2.4 394 */ 395 #define CBU16_LENGTH(c) ((uint32_t)(c) <= 0xffff ? 1 : 2) 396 397 /** 398 * The maximum number of 16-bit code units per Unicode code point 399 * (U+0000..U+10ffff). 400 * @return 2 401 * @stable ICU 2.4 402 */ 403 #define CBU16_MAX_LENGTH 2 404 405 /** 406 * Get a code point from a string at a code point boundary offset, 407 * and advance the offset to the next code point boundary. 408 * (Post-incrementing forward iteration.) 409 * "Safe" macro, handles unpaired surrogates and checks for string boundaries. 410 * 411 * The length can be negative for a NUL-terminated string. 412 * 413 * The offset may point to the lead surrogate unit 414 * for a supplementary code point, in which case the macro will read 415 * the following trail surrogate as well. 416 * If the offset points to a trail surrogate or 417 * to a single, unpaired lead surrogate, then c is set to that unpaired 418 * surrogate. 419 * 420 * @param s const UChar * string 421 * @param i string offset, must be i<length 422 * @param length string length 423 * @param c output UChar32 variable 424 * @see U16_NEXT_UNSAFE 425 * @stable ICU 2.4 426 */ 427 #define CBU16_NEXT(s, i, length, c) \ 428 { \ 429 (c) = (s)[(i)++]; \ 430 if (CBU16_IS_LEAD(c)) { \ 431 uint16_t __c2; \ 432 if ((i) != (length) && CBU16_IS_TRAIL(__c2 = (s)[(i)])) { \ 433 ++(i); \ 434 (c) = CBU16_GET_SUPPLEMENTARY((c), __c2); \ 435 } \ 436 } \ 437 } 438 439 /** 440 * Append a code point to a string, overwriting 1 or 2 code units. 441 * The offset points to the current end of the string contents 442 * and is advanced (post-increment). 443 * "Unsafe" macro, assumes a valid code point and sufficient space in the 444 * string. Otherwise, the result is undefined. 445 * 446 * @param s const UChar * string buffer 447 * @param i string offset 448 * @param c code point to append 449 * @see U16_APPEND 450 * @stable ICU 2.4 451 */ 452 #define CBU16_APPEND_UNSAFE(s, i, c) \ 453 { \ 454 if ((uint32_t)(c) <= 0xffff) { \ 455 (s)[(i)++] = (uint16_t)(c); \ 456 } else { \ 457 (s)[(i)++] = (uint16_t)(((c) >> 10) + 0xd7c0); \ 458 (s)[(i)++] = (uint16_t)(((c)&0x3ff) | 0xdc00); \ 459 } \ 460 } 461 462 } // namespace base_icu 463 464 #endif // BASE_THIRD_PARTY_ICU_ICU_UTF_H_ 465