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 "base/basictypes.h" 21 22namespace base_icu { 23 24typedef uint32 UChar32; 25typedef int8 UBool; 26 27// General --------------------------------------------------------------------- 28// from utf.h 29 30/** 31 * This value is intended for sentinel values for APIs that 32 * (take or) return single code points (UChar32). 33 * It is outside of the Unicode code point range 0..0x10ffff. 34 * 35 * For example, a "done" or "error" value in a new API 36 * could be indicated with CBU_SENTINEL. 37 * 38 * ICU APIs designed before ICU 2.4 usually define service-specific "done" 39 * values, mostly 0xffff. 40 * Those may need to be distinguished from 41 * actual U+ffff text contents by calling functions like 42 * CharacterIterator::hasNext() or UnicodeString::length(). 43 * 44 * @return -1 45 * @see UChar32 46 * @stable ICU 2.4 47 */ 48#define CBU_SENTINEL (-1) 49 50/** 51 * Is this code point a Unicode noncharacter? 52 * @param c 32-bit code point 53 * @return TRUE or FALSE 54 * @stable ICU 2.4 55 */ 56#define CBU_IS_UNICODE_NONCHAR(c) \ 57 ((c)>=0xfdd0 && \ 58 ((uint32)(c)<=0xfdef || ((c)&0xfffe)==0xfffe) && \ 59 (uint32)(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)(c)<0xd800 || \ 80 ((uint32)(c)>0xdfff && \ 81 (uint32)(c)<=0x10ffff && \ 82 !CBU_IS_UNICODE_NONCHAR(c))) 83 84/** 85 * Is this code point a surrogate (U+d800..U+dfff)? 86 * @param c 32-bit code point 87 * @return TRUE or FALSE 88 * @stable ICU 2.4 89 */ 90#define CBU_IS_SURROGATE(c) (((c)&0xfffff800)==0xd800) 91 92/** 93 * Assuming c is a surrogate code point (U_IS_SURROGATE(c)), 94 * is it a lead surrogate? 95 * @param c 32-bit code point 96 * @return TRUE or FALSE 97 * @stable ICU 2.4 98 */ 99#define CBU_IS_SURROGATE_LEAD(c) (((c)&0x400)==0) 100 101 102// UTF-8 macros ---------------------------------------------------------------- 103// from utf8.h 104 105extern const uint8 utf8_countTrailBytes[256]; 106 107/** 108 * Count the trail bytes for a UTF-8 lead byte. 109 * @internal 110 */ 111#define CBU8_COUNT_TRAIL_BYTES(leadByte) (base_icu::utf8_countTrailBytes[(uint8)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)((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)(c)<=0x7f ? 1 : \ 152 ((uint32)(c)<=0x7ff ? 2 : \ 153 ((uint32)(c)<=0xd7ff ? 3 : \ 154 ((uint32)(c)<=0xdfff || (uint32)(c)>0x10ffff ? 0 : \ 155 ((uint32)(c)<=0xffff ? 3 : 4)\ 156 ) \ 157 ) \ 158 ) \ 159 ) 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 *s, int32 *pi, int32 length, UChar32 c, UBool strict); 173 174/** 175 * Get a code point from a string at a code point boundary offset, 176 * and advance the offset to the next code point boundary. 177 * (Post-incrementing forward iteration.) 178 * "Safe" macro, checks for illegal sequences and for string boundaries. 179 * 180 * The offset may point to the lead byte of a multi-byte sequence, 181 * in which case the macro will read the whole sequence. 182 * If the offset points to a trail byte or an illegal UTF-8 sequence, then 183 * c is set to a negative value. 184 * 185 * @param s const uint8 * string 186 * @param i string offset, i<length 187 * @param length string length 188 * @param c output UChar32 variable, set to <0 in case of an error 189 * @see CBU8_NEXT_UNSAFE 190 * @stable ICU 2.4 191 */ 192#define CBU8_NEXT(s, i, length, c) { \ 193 (c)=(s)[(i)++]; \ 194 if(((uint8)(c))>=0x80) { \ 195 if(CBU8_IS_LEAD(c)) { \ 196 (c)=base_icu::utf8_nextCharSafeBody((const uint8 *)s, &(i), (int32)(length), c, -1); \ 197 } else { \ 198 (c)=CBU_SENTINEL; \ 199 } \ 200 } \ 201} 202 203/** 204 * Append a code point to a string, overwriting 1 to 4 bytes. 205 * The offset points to the current end of the string contents 206 * and is advanced (post-increment). 207 * "Unsafe" macro, assumes a valid code point and sufficient space in the string. 208 * Otherwise, the result is undefined. 209 * 210 * @param s const uint8 * string buffer 211 * @param i string offset 212 * @param c code point to append 213 * @see CBU8_APPEND 214 * @stable ICU 2.4 215 */ 216#define CBU8_APPEND_UNSAFE(s, i, c) { \ 217 if((uint32)(c)<=0x7f) { \ 218 (s)[(i)++]=(uint8)(c); \ 219 } else { \ 220 if((uint32)(c)<=0x7ff) { \ 221 (s)[(i)++]=(uint8)(((c)>>6)|0xc0); \ 222 } else { \ 223 if((uint32)(c)<=0xffff) { \ 224 (s)[(i)++]=(uint8)(((c)>>12)|0xe0); \ 225 } else { \ 226 (s)[(i)++]=(uint8)(((c)>>18)|0xf0); \ 227 (s)[(i)++]=(uint8)((((c)>>12)&0x3f)|0x80); \ 228 } \ 229 (s)[(i)++]=(uint8)((((c)>>6)&0x3f)|0x80); \ 230 } \ 231 (s)[(i)++]=(uint8)(((c)&0x3f)|0x80); \ 232 } \ 233} 234 235// UTF-16 macros --------------------------------------------------------------- 236// from utf16.h 237 238/** 239 * Does this code unit alone encode a code point (BMP, not a surrogate)? 240 * @param c 16-bit code unit 241 * @return TRUE or FALSE 242 * @stable ICU 2.4 243 */ 244#define CBU16_IS_SINGLE(c) !CBU_IS_SURROGATE(c) 245 246/** 247 * Is this code unit a lead surrogate (U+d800..U+dbff)? 248 * @param c 16-bit code unit 249 * @return TRUE or FALSE 250 * @stable ICU 2.4 251 */ 252#define CBU16_IS_LEAD(c) (((c)&0xfffffc00)==0xd800) 253 254/** 255 * Is this code unit a trail surrogate (U+dc00..U+dfff)? 256 * @param c 16-bit code unit 257 * @return TRUE or FALSE 258 * @stable ICU 2.4 259 */ 260#define CBU16_IS_TRAIL(c) (((c)&0xfffffc00)==0xdc00) 261 262/** 263 * Is this code unit a surrogate (U+d800..U+dfff)? 264 * @param c 16-bit code unit 265 * @return TRUE or FALSE 266 * @stable ICU 2.4 267 */ 268#define CBU16_IS_SURROGATE(c) CBU_IS_SURROGATE(c) 269 270/** 271 * Assuming c is a surrogate code point (U16_IS_SURROGATE(c)), 272 * is it a lead surrogate? 273 * @param c 16-bit code unit 274 * @return TRUE or FALSE 275 * @stable ICU 2.4 276 */ 277#define CBU16_IS_SURROGATE_LEAD(c) (((c)&0x400)==0) 278 279/** 280 * Helper constant for CBU16_GET_SUPPLEMENTARY. 281 * @internal 282 */ 283#define CBU16_SURROGATE_OFFSET ((0xd800<<10UL)+0xdc00-0x10000) 284 285/** 286 * Get a supplementary code point value (U+10000..U+10ffff) 287 * from its lead and trail surrogates. 288 * The result is undefined if the input values are not 289 * lead and trail surrogates. 290 * 291 * @param lead lead surrogate (U+d800..U+dbff) 292 * @param trail trail surrogate (U+dc00..U+dfff) 293 * @return supplementary code point (U+10000..U+10ffff) 294 * @stable ICU 2.4 295 */ 296#define CBU16_GET_SUPPLEMENTARY(lead, trail) \ 297 (((base_icu::UChar32)(lead)<<10UL)+(base_icu::UChar32)(trail)-CBU16_SURROGATE_OFFSET) 298 299 300/** 301 * Get the lead surrogate (0xd800..0xdbff) for a 302 * supplementary code point (0x10000..0x10ffff). 303 * @param supplementary 32-bit code point (U+10000..U+10ffff) 304 * @return lead surrogate (U+d800..U+dbff) for supplementary 305 * @stable ICU 2.4 306 */ 307#define CBU16_LEAD(supplementary) (UChar)(((supplementary)>>10)+0xd7c0) 308 309/** 310 * Get the trail surrogate (0xdc00..0xdfff) for a 311 * supplementary code point (0x10000..0x10ffff). 312 * @param supplementary 32-bit code point (U+10000..U+10ffff) 313 * @return trail surrogate (U+dc00..U+dfff) for supplementary 314 * @stable ICU 2.4 315 */ 316#define CBU16_TRAIL(supplementary) (UChar)(((supplementary)&0x3ff)|0xdc00) 317 318/** 319 * How many 16-bit code units are used to encode this Unicode code point? (1 or 2) 320 * The result is not defined if c is not a Unicode code point (U+0000..U+10ffff). 321 * @param c 32-bit code point 322 * @return 1 or 2 323 * @stable ICU 2.4 324 */ 325#define CBU16_LENGTH(c) ((uint32)(c)<=0xffff ? 1 : 2) 326 327/** 328 * The maximum number of 16-bit code units per Unicode code point (U+0000..U+10ffff). 329 * @return 2 330 * @stable ICU 2.4 331 */ 332#define CBU16_MAX_LENGTH 2 333 334/** 335 * Get a code point from a string at a code point boundary offset, 336 * and advance the offset to the next code point boundary. 337 * (Post-incrementing forward iteration.) 338 * "Safe" macro, handles unpaired surrogates and checks for string boundaries. 339 * 340 * The offset may point to the lead surrogate unit 341 * for a supplementary code point, in which case the macro will read 342 * the following trail surrogate as well. 343 * If the offset points to a trail surrogate or 344 * to a single, unpaired lead surrogate, then that itself 345 * will be returned as the code point. 346 * 347 * @param s const UChar * string 348 * @param i string offset, i<length 349 * @param length string length 350 * @param c output UChar32 variable 351 * @stable ICU 2.4 352 */ 353#define CBU16_NEXT(s, i, length, c) { \ 354 (c)=(s)[(i)++]; \ 355 if(CBU16_IS_LEAD(c)) { \ 356 uint16 __c2; \ 357 if((i)<(length) && CBU16_IS_TRAIL(__c2=(s)[(i)])) { \ 358 ++(i); \ 359 (c)=CBU16_GET_SUPPLEMENTARY((c), __c2); \ 360 } \ 361 } \ 362} 363 364/** 365 * Append a code point to a string, overwriting 1 or 2 code units. 366 * The offset points to the current end of the string contents 367 * and is advanced (post-increment). 368 * "Unsafe" macro, assumes a valid code point and sufficient space in the string. 369 * Otherwise, the result is undefined. 370 * 371 * @param s const UChar * string buffer 372 * @param i string offset 373 * @param c code point to append 374 * @see CBU16_APPEND 375 * @stable ICU 2.4 376 */ 377#define CBU16_APPEND_UNSAFE(s, i, c) { \ 378 if((uint32)(c)<=0xffff) { \ 379 (s)[(i)++]=(uint16)(c); \ 380 } else { \ 381 (s)[(i)++]=(uint16)(((c)>>10)+0xd7c0); \ 382 (s)[(i)++]=(uint16)(((c)&0x3ff)|0xdc00); \ 383 } \ 384} 385 386} // namesapce base_icu 387 388#endif // BASE_THIRD_PARTY_ICU_ICU_UTF_H_ 389