1/* 2******************************************************************************* 3* 4* Copyright (C) 1999-2004, International Business Machines 5* Corporation and others. All Rights Reserved. 6* 7******************************************************************************* 8* file name: utf.h 9* encoding: US-ASCII 10* tab size: 8 (not used) 11* indentation:4 12* 13* created on: 1999sep09 14* created by: Markus W. Scherer 15*/ 16 17#ifndef BASE_THIRD_PARTY_ICU_ICU_UTF_H_ 18#define BASE_THIRD_PARTY_ICU_ICU_UTF_H_ 19 20#include <stdint.h> 21 22namespace base_icu { 23 24typedef int32_t UChar32; 25typedef uint16_t UChar; 26typedef int8_t UBool; 27 28// General --------------------------------------------------------------------- 29// from utf.h 30 31/** 32 * This value is intended for sentinel values for APIs that 33 * (take or) return single code points (UChar32). 34 * It is outside of the Unicode code point range 0..0x10ffff. 35 * 36 * For example, a "done" or "error" value in a new API 37 * could be indicated with CBU_SENTINEL. 38 * 39 * ICU APIs designed before ICU 2.4 usually define service-specific "done" 40 * values, mostly 0xffff. 41 * Those may need to be distinguished from 42 * actual U+ffff text contents by calling functions like 43 * CharacterIterator::hasNext() or UnicodeString::length(). 44 * 45 * @return -1 46 * @see UChar32 47 * @stable ICU 2.4 48 */ 49#define CBU_SENTINEL (-1) 50 51/** 52 * Is this code point a Unicode noncharacter? 53 * @param c 32-bit code point 54 * @return TRUE or FALSE 55 * @stable ICU 2.4 56 */ 57#define CBU_IS_UNICODE_NONCHAR(c) \ 58 ((c) >= 0xfdd0 && ((uint32_t)(c) <= 0xfdef || ((c)&0xfffe) == 0xfffe) && \ 59 (uint32_t)(c) <= 0x10ffff) 60 61/** 62 * Is c a Unicode code point value (0..U+10ffff) 63 * that can be assigned a character? 64 * 65 * Code points that are not characters include: 66 * - single surrogate code points (U+d800..U+dfff, 2048 code points) 67 * - the last two code points on each plane (U+__fffe and U+__ffff, 34 code points) 68 * - U+fdd0..U+fdef (new with Unicode 3.1, 32 code points) 69 * - the highest Unicode code point value is U+10ffff 70 * 71 * This means that all code points below U+d800 are character code points, 72 * and that boundary is tested first for performance. 73 * 74 * @param c 32-bit code point 75 * @return TRUE or FALSE 76 * @stable ICU 2.4 77 */ 78#define CBU_IS_UNICODE_CHAR(c) \ 79 ((uint32_t)(c) < 0xd800 || \ 80 ((uint32_t)(c) > 0xdfff && (uint32_t)(c) <= 0x10ffff && \ 81 !CBU_IS_UNICODE_NONCHAR(c))) 82 83/** 84 * Is this code point a surrogate (U+d800..U+dfff)? 85 * @param c 32-bit code point 86 * @return TRUE or FALSE 87 * @stable ICU 2.4 88 */ 89#define CBU_IS_SURROGATE(c) (((c)&0xfffff800)==0xd800) 90 91/** 92 * Assuming c is a surrogate code point (U_IS_SURROGATE(c)), 93 * is it a lead surrogate? 94 * @param c 32-bit code point 95 * @return TRUE or FALSE 96 * @stable ICU 2.4 97 */ 98#define CBU_IS_SURROGATE_LEAD(c) (((c)&0x400)==0) 99 100 101// UTF-8 macros ---------------------------------------------------------------- 102// from utf8.h 103 104extern const uint8_t utf8_countTrailBytes[256]; 105 106/** 107 * Count the trail bytes for a UTF-8 lead byte. 108 * @internal 109 */ 110#define CBU8_COUNT_TRAIL_BYTES(leadByte) \ 111 (base_icu::utf8_countTrailBytes[(uint8_t)leadByte]) 112 113/** 114 * Mask a UTF-8 lead byte, leave only the lower bits that form part of the code point value. 115 * @internal 116 */ 117#define CBU8_MASK_LEAD_BYTE(leadByte, countTrailBytes) ((leadByte)&=(1<<(6-(countTrailBytes)))-1) 118 119/** 120 * Does this code unit (byte) encode a code point by itself (US-ASCII 0..0x7f)? 121 * @param c 8-bit code unit (byte) 122 * @return TRUE or FALSE 123 * @stable ICU 2.4 124 */ 125#define CBU8_IS_SINGLE(c) (((c)&0x80)==0) 126 127/** 128 * Is this code unit (byte) a UTF-8 lead byte? 129 * @param c 8-bit code unit (byte) 130 * @return TRUE or FALSE 131 * @stable ICU 2.4 132 */ 133#define CBU8_IS_LEAD(c) ((uint8_t)((c)-0xc0) < 0x3e) 134 135/** 136 * Is this code unit (byte) a UTF-8 trail byte? 137 * @param c 8-bit code unit (byte) 138 * @return TRUE or FALSE 139 * @stable ICU 2.4 140 */ 141#define CBU8_IS_TRAIL(c) (((c)&0xc0)==0x80) 142 143/** 144 * How many code units (bytes) are used for the UTF-8 encoding 145 * of this Unicode code point? 146 * @param c 32-bit code point 147 * @return 1..4, or 0 if c is a surrogate or not a Unicode code point 148 * @stable ICU 2.4 149 */ 150#define CBU8_LENGTH(c) \ 151 ((uint32_t)(c) <= 0x7f \ 152 ? 1 \ 153 : ((uint32_t)(c) <= 0x7ff \ 154 ? 2 \ 155 : ((uint32_t)(c) <= 0xd7ff \ 156 ? 3 \ 157 : ((uint32_t)(c) <= 0xdfff || (uint32_t)(c) > 0x10ffff \ 158 ? 0 \ 159 : ((uint32_t)(c) <= 0xffff ? 3 : 4))))) 160 161/** 162 * The maximum number of UTF-8 code units (bytes) per Unicode code point (U+0000..U+10ffff). 163 * @return 4 164 * @stable ICU 2.4 165 */ 166#define CBU8_MAX_LENGTH 4 167 168/** 169 * Function for handling "next code point" with error-checking. 170 * @internal 171 */ 172UChar32 utf8_nextCharSafeBody(const uint8_t* s, 173 int32_t* pi, 174 int32_t length, 175 UChar32 c, 176 UBool strict); 177 178/** 179 * Get a code point from a string at a code point boundary offset, 180 * and advance the offset to the next code point boundary. 181 * (Post-incrementing forward iteration.) 182 * "Safe" macro, checks for illegal sequences and for string boundaries. 183 * 184 * The offset may point to the lead byte of a multi-byte sequence, 185 * in which case the macro will read the whole sequence. 186 * If the offset points to a trail byte or an illegal UTF-8 sequence, then 187 * c is set to a negative value. 188 * 189 * @param s const uint8_t * string 190 * @param i string offset, i<length 191 * @param length string length 192 * @param c output UChar32 variable, set to <0 in case of an error 193 * @see CBU8_NEXT_UNSAFE 194 * @stable ICU 2.4 195 */ 196#define CBU8_NEXT(s, i, length, c) \ 197 { \ 198 (c) = (s)[(i)++]; \ 199 if (((uint8_t)(c)) >= 0x80) { \ 200 if (CBU8_IS_LEAD(c)) { \ 201 (c) = base_icu::utf8_nextCharSafeBody((const uint8_t*)s, &(i), \ 202 (int32_t)(length), c, -1); \ 203 } else { \ 204 (c) = CBU_SENTINEL; \ 205 } \ 206 } \ 207 } 208 209/** 210 * Append a code point to a string, overwriting 1 to 4 bytes. 211 * The offset points to the current end of the string contents 212 * and is advanced (post-increment). 213 * "Unsafe" macro, assumes a valid code point and sufficient space in the 214 * string. 215 * Otherwise, the result is undefined. 216 * 217 * @param s const uint8_t * string buffer 218 * @param i string offset 219 * @param c code point to append 220 * @see CBU8_APPEND 221 * @stable ICU 2.4 222 */ 223#define CBU8_APPEND_UNSAFE(s, i, c) \ 224 { \ 225 if ((uint32_t)(c) <= 0x7f) { \ 226 (s)[(i)++] = (uint8_t)(c); \ 227 } else { \ 228 if ((uint32_t)(c) <= 0x7ff) { \ 229 (s)[(i)++] = (uint8_t)(((c) >> 6) | 0xc0); \ 230 } else { \ 231 if ((uint32_t)(c) <= 0xffff) { \ 232 (s)[(i)++] = (uint8_t)(((c) >> 12) | 0xe0); \ 233 } else { \ 234 (s)[(i)++] = (uint8_t)(((c) >> 18) | 0xf0); \ 235 (s)[(i)++] = (uint8_t)((((c) >> 12) & 0x3f) | 0x80); \ 236 } \ 237 (s)[(i)++] = (uint8_t)((((c) >> 6) & 0x3f) | 0x80); \ 238 } \ 239 (s)[(i)++] = (uint8_t)(((c)&0x3f) | 0x80); \ 240 } \ 241 } 242 243// UTF-16 macros --------------------------------------------------------------- 244// from utf16.h 245 246/** 247 * Does this code unit alone encode a code point (BMP, not a surrogate)? 248 * @param c 16-bit code unit 249 * @return TRUE or FALSE 250 * @stable ICU 2.4 251 */ 252#define CBU16_IS_SINGLE(c) !CBU_IS_SURROGATE(c) 253 254/** 255 * Is this code unit a lead surrogate (U+d800..U+dbff)? 256 * @param c 16-bit code unit 257 * @return TRUE or FALSE 258 * @stable ICU 2.4 259 */ 260#define CBU16_IS_LEAD(c) (((c)&0xfffffc00)==0xd800) 261 262/** 263 * Is this code unit a trail surrogate (U+dc00..U+dfff)? 264 * @param c 16-bit code unit 265 * @return TRUE or FALSE 266 * @stable ICU 2.4 267 */ 268#define CBU16_IS_TRAIL(c) (((c)&0xfffffc00)==0xdc00) 269 270/** 271 * Is this code unit a surrogate (U+d800..U+dfff)? 272 * @param c 16-bit code unit 273 * @return TRUE or FALSE 274 * @stable ICU 2.4 275 */ 276#define CBU16_IS_SURROGATE(c) CBU_IS_SURROGATE(c) 277 278/** 279 * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)), 280 * is it a lead surrogate? 281 * @param c 16-bit code unit 282 * @return TRUE or FALSE 283 * @stable ICU 2.4 284 */ 285#define CBU16_IS_SURROGATE_LEAD(c) (((c)&0x400)==0) 286 287/** 288 * Helper constant for CBU16_GET_SUPPLEMENTARY. 289 * @internal 290 */ 291#define CBU16_SURROGATE_OFFSET ((0xd800<<10UL)+0xdc00-0x10000) 292 293/** 294 * Get a supplementary code point value (U+10000..U+10ffff) 295 * from its lead and trail surrogates. 296 * The result is undefined if the input values are not 297 * lead and trail surrogates. 298 * 299 * @param lead lead surrogate (U+d800..U+dbff) 300 * @param trail trail surrogate (U+dc00..U+dfff) 301 * @return supplementary code point (U+10000..U+10ffff) 302 * @stable ICU 2.4 303 */ 304#define CBU16_GET_SUPPLEMENTARY(lead, trail) \ 305 (((base_icu::UChar32)(lead)<<10UL)+(base_icu::UChar32)(trail)-CBU16_SURROGATE_OFFSET) 306 307 308/** 309 * Get the lead surrogate (0xd800..0xdbff) for a 310 * supplementary code point (0x10000..0x10ffff). 311 * @param supplementary 32-bit code point (U+10000..U+10ffff) 312 * @return lead surrogate (U+d800..U+dbff) for supplementary 313 * @stable ICU 2.4 314 */ 315#define CBU16_LEAD(supplementary) \ 316 (base_icu::UChar)(((supplementary)>>10)+0xd7c0) 317 318/** 319 * Get the trail surrogate (0xdc00..0xdfff) for a 320 * supplementary code point (0x10000..0x10ffff). 321 * @param supplementary 32-bit code point (U+10000..U+10ffff) 322 * @return trail surrogate (U+dc00..U+dfff) for supplementary 323 * @stable ICU 2.4 324 */ 325#define CBU16_TRAIL(supplementary) \ 326 (base_icu::UChar)(((supplementary)&0x3ff)|0xdc00) 327 328/** 329 * How many 16-bit code units are used to encode this Unicode code point? (1 or 2) 330 * The result is not defined if c is not a Unicode code point (U+0000..U+10ffff). 331 * @param c 32-bit code point 332 * @return 1 or 2 333 * @stable ICU 2.4 334 */ 335#define CBU16_LENGTH(c) ((uint32_t)(c) <= 0xffff ? 1 : 2) 336 337/** 338 * The maximum number of 16-bit code units per Unicode code point (U+0000..U+10ffff). 339 * @return 2 340 * @stable ICU 2.4 341 */ 342#define CBU16_MAX_LENGTH 2 343 344/** 345 * Get a code point from a string at a code point boundary offset, 346 * and advance the offset to the next code point boundary. 347 * (Post-incrementing forward iteration.) 348 * "Safe" macro, handles unpaired surrogates and checks for string boundaries. 349 * 350 * The offset may point to the lead surrogate unit 351 * for a supplementary code point, in which case the macro will read 352 * the following trail surrogate as well. 353 * If the offset points to a trail surrogate or 354 * to a single, unpaired lead surrogate, then that itself 355 * will be returned as the code point. 356 * 357 * @param s const UChar * string 358 * @param i string offset, i<length 359 * @param length string length 360 * @param c output UChar32 variable 361 * @stable ICU 2.4 362 */ 363#define CBU16_NEXT(s, i, length, c) \ 364 { \ 365 (c) = (s)[(i)++]; \ 366 if (CBU16_IS_LEAD(c)) { \ 367 uint16_t __c2; \ 368 if ((i) < (length) && CBU16_IS_TRAIL(__c2 = (s)[(i)])) { \ 369 ++(i); \ 370 (c) = CBU16_GET_SUPPLEMENTARY((c), __c2); \ 371 } \ 372 } \ 373 } 374 375/** 376 * Append a code point to a string, overwriting 1 or 2 code units. 377 * The offset points to the current end of the string contents 378 * and is advanced (post-increment). 379 * "Unsafe" macro, assumes a valid code point and sufficient space in the string. 380 * Otherwise, the result is undefined. 381 * 382 * @param s const UChar * string buffer 383 * @param i string offset 384 * @param c code point to append 385 * @see CBU16_APPEND 386 * @stable ICU 2.4 387 */ 388#define CBU16_APPEND_UNSAFE(s, i, c) \ 389 { \ 390 if ((uint32_t)(c) <= 0xffff) { \ 391 (s)[(i)++] = (uint16_t)(c); \ 392 } else { \ 393 (s)[(i)++] = (uint16_t)(((c) >> 10) + 0xd7c0); \ 394 (s)[(i)++] = (uint16_t)(((c)&0x3ff) | 0xdc00); \ 395 } \ 396 } 397 398} // namesapce base_icu 399 400#endif // BASE_THIRD_PARTY_ICU_ICU_UTF_H_ 401