jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame^] | 1 | /* |
| 2 | ******************************************************************************* |
| 3 | * |
| 4 | * Copyright (C) 2002-2012, International Business Machines |
| 5 | * Corporation and others. All Rights Reserved. |
| 6 | * |
| 7 | ******************************************************************************* |
| 8 | * file name: uiter.cpp |
| 9 | * encoding: US-ASCII |
| 10 | * tab size: 8 (not used) |
| 11 | * indentation:4 |
| 12 | * |
| 13 | * created on: 2002jan18 |
| 14 | * created by: Markus W. Scherer |
| 15 | */ |
| 16 | |
| 17 | #include "unicode/utypes.h" |
| 18 | #include "unicode/ustring.h" |
| 19 | #include "unicode/chariter.h" |
| 20 | #include "unicode/rep.h" |
| 21 | #include "unicode/uiter.h" |
| 22 | #include "unicode/utf.h" |
| 23 | #include "unicode/utf8.h" |
| 24 | #include "unicode/utf16.h" |
| 25 | #include "cstring.h" |
| 26 | |
| 27 | U_NAMESPACE_USE |
| 28 | |
| 29 | #define IS_EVEN(n) (((n)&1)==0) |
| 30 | #define IS_POINTER_EVEN(p) IS_EVEN((size_t)p) |
| 31 | |
| 32 | U_CDECL_BEGIN |
| 33 | |
| 34 | /* No-Op UCharIterator implementation for illegal input --------------------- */ |
| 35 | |
| 36 | static int32_t U_CALLCONV |
| 37 | noopGetIndex(UCharIterator * /*iter*/, UCharIteratorOrigin /*origin*/) { |
| 38 | return 0; |
| 39 | } |
| 40 | |
| 41 | static int32_t U_CALLCONV |
| 42 | noopMove(UCharIterator * /*iter*/, int32_t /*delta*/, UCharIteratorOrigin /*origin*/) { |
| 43 | return 0; |
| 44 | } |
| 45 | |
| 46 | static UBool U_CALLCONV |
| 47 | noopHasNext(UCharIterator * /*iter*/) { |
| 48 | return FALSE; |
| 49 | } |
| 50 | |
| 51 | static UChar32 U_CALLCONV |
| 52 | noopCurrent(UCharIterator * /*iter*/) { |
| 53 | return U_SENTINEL; |
| 54 | } |
| 55 | |
| 56 | static uint32_t U_CALLCONV |
| 57 | noopGetState(const UCharIterator * /*iter*/) { |
| 58 | return UITER_NO_STATE; |
| 59 | } |
| 60 | |
| 61 | static void U_CALLCONV |
| 62 | noopSetState(UCharIterator * /*iter*/, uint32_t /*state*/, UErrorCode *pErrorCode) { |
| 63 | *pErrorCode=U_UNSUPPORTED_ERROR; |
| 64 | } |
| 65 | |
| 66 | static const UCharIterator noopIterator={ |
| 67 | 0, 0, 0, 0, 0, 0, |
| 68 | noopGetIndex, |
| 69 | noopMove, |
| 70 | noopHasNext, |
| 71 | noopHasNext, |
| 72 | noopCurrent, |
| 73 | noopCurrent, |
| 74 | noopCurrent, |
| 75 | NULL, |
| 76 | noopGetState, |
| 77 | noopSetState |
| 78 | }; |
| 79 | |
| 80 | /* UCharIterator implementation for simple strings -------------------------- */ |
| 81 | |
| 82 | /* |
| 83 | * This is an implementation of a code unit (UChar) iterator |
| 84 | * for UChar * strings. |
| 85 | * |
| 86 | * The UCharIterator.context field holds a pointer to the string. |
| 87 | */ |
| 88 | |
| 89 | static int32_t U_CALLCONV |
| 90 | stringIteratorGetIndex(UCharIterator *iter, UCharIteratorOrigin origin) { |
| 91 | switch(origin) { |
| 92 | case UITER_ZERO: |
| 93 | return 0; |
| 94 | case UITER_START: |
| 95 | return iter->start; |
| 96 | case UITER_CURRENT: |
| 97 | return iter->index; |
| 98 | case UITER_LIMIT: |
| 99 | return iter->limit; |
| 100 | case UITER_LENGTH: |
| 101 | return iter->length; |
| 102 | default: |
| 103 | /* not a valid origin */ |
| 104 | /* Should never get here! */ |
| 105 | return -1; |
| 106 | } |
| 107 | } |
| 108 | |
| 109 | static int32_t U_CALLCONV |
| 110 | stringIteratorMove(UCharIterator *iter, int32_t delta, UCharIteratorOrigin origin) { |
| 111 | int32_t pos; |
| 112 | |
| 113 | switch(origin) { |
| 114 | case UITER_ZERO: |
| 115 | pos=delta; |
| 116 | break; |
| 117 | case UITER_START: |
| 118 | pos=iter->start+delta; |
| 119 | break; |
| 120 | case UITER_CURRENT: |
| 121 | pos=iter->index+delta; |
| 122 | break; |
| 123 | case UITER_LIMIT: |
| 124 | pos=iter->limit+delta; |
| 125 | break; |
| 126 | case UITER_LENGTH: |
| 127 | pos=iter->length+delta; |
| 128 | break; |
| 129 | default: |
| 130 | return -1; /* Error */ |
| 131 | } |
| 132 | |
| 133 | if(pos<iter->start) { |
| 134 | pos=iter->start; |
| 135 | } else if(pos>iter->limit) { |
| 136 | pos=iter->limit; |
| 137 | } |
| 138 | |
| 139 | return iter->index=pos; |
| 140 | } |
| 141 | |
| 142 | static UBool U_CALLCONV |
| 143 | stringIteratorHasNext(UCharIterator *iter) { |
| 144 | return iter->index<iter->limit; |
| 145 | } |
| 146 | |
| 147 | static UBool U_CALLCONV |
| 148 | stringIteratorHasPrevious(UCharIterator *iter) { |
| 149 | return iter->index>iter->start; |
| 150 | } |
| 151 | |
| 152 | static UChar32 U_CALLCONV |
| 153 | stringIteratorCurrent(UCharIterator *iter) { |
| 154 | if(iter->index<iter->limit) { |
| 155 | return ((const UChar *)(iter->context))[iter->index]; |
| 156 | } else { |
| 157 | return U_SENTINEL; |
| 158 | } |
| 159 | } |
| 160 | |
| 161 | static UChar32 U_CALLCONV |
| 162 | stringIteratorNext(UCharIterator *iter) { |
| 163 | if(iter->index<iter->limit) { |
| 164 | return ((const UChar *)(iter->context))[iter->index++]; |
| 165 | } else { |
| 166 | return U_SENTINEL; |
| 167 | } |
| 168 | } |
| 169 | |
| 170 | static UChar32 U_CALLCONV |
| 171 | stringIteratorPrevious(UCharIterator *iter) { |
| 172 | if(iter->index>iter->start) { |
| 173 | return ((const UChar *)(iter->context))[--iter->index]; |
| 174 | } else { |
| 175 | return U_SENTINEL; |
| 176 | } |
| 177 | } |
| 178 | |
| 179 | static uint32_t U_CALLCONV |
| 180 | stringIteratorGetState(const UCharIterator *iter) { |
| 181 | return (uint32_t)iter->index; |
| 182 | } |
| 183 | |
| 184 | static void U_CALLCONV |
| 185 | stringIteratorSetState(UCharIterator *iter, uint32_t state, UErrorCode *pErrorCode) { |
| 186 | if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) { |
| 187 | /* do nothing */ |
| 188 | } else if(iter==NULL) { |
| 189 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
| 190 | } else if((int32_t)state<iter->start || iter->limit<(int32_t)state) { |
| 191 | *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
| 192 | } else { |
| 193 | iter->index=(int32_t)state; |
| 194 | } |
| 195 | } |
| 196 | |
| 197 | static const UCharIterator stringIterator={ |
| 198 | 0, 0, 0, 0, 0, 0, |
| 199 | stringIteratorGetIndex, |
| 200 | stringIteratorMove, |
| 201 | stringIteratorHasNext, |
| 202 | stringIteratorHasPrevious, |
| 203 | stringIteratorCurrent, |
| 204 | stringIteratorNext, |
| 205 | stringIteratorPrevious, |
| 206 | NULL, |
| 207 | stringIteratorGetState, |
| 208 | stringIteratorSetState |
| 209 | }; |
| 210 | |
| 211 | U_CAPI void U_EXPORT2 |
| 212 | uiter_setString(UCharIterator *iter, const UChar *s, int32_t length) { |
| 213 | if(iter!=0) { |
| 214 | if(s!=0 && length>=-1) { |
| 215 | *iter=stringIterator; |
| 216 | iter->context=s; |
| 217 | if(length>=0) { |
| 218 | iter->length=length; |
| 219 | } else { |
| 220 | iter->length=u_strlen(s); |
| 221 | } |
| 222 | iter->limit=iter->length; |
| 223 | } else { |
| 224 | *iter=noopIterator; |
| 225 | } |
| 226 | } |
| 227 | } |
| 228 | |
| 229 | /* UCharIterator implementation for UTF-16BE strings ------------------------ */ |
| 230 | |
| 231 | /* |
| 232 | * This is an implementation of a code unit (UChar) iterator |
| 233 | * for UTF-16BE strings, i.e., strings in byte-vectors where |
| 234 | * each UChar is stored as a big-endian pair of bytes. |
| 235 | * |
| 236 | * The UCharIterator.context field holds a pointer to the string. |
| 237 | * Everything works just like with a normal UChar iterator (uiter_setString), |
| 238 | * except that UChars are assembled from byte pairs. |
| 239 | */ |
| 240 | |
| 241 | /* internal helper function */ |
| 242 | static inline UChar32 |
| 243 | utf16BEIteratorGet(UCharIterator *iter, int32_t index) { |
| 244 | const uint8_t *p=(const uint8_t *)iter->context; |
| 245 | return ((UChar)p[2*index]<<8)|(UChar)p[2*index+1]; |
| 246 | } |
| 247 | |
| 248 | static UChar32 U_CALLCONV |
| 249 | utf16BEIteratorCurrent(UCharIterator *iter) { |
| 250 | int32_t index; |
| 251 | |
| 252 | if((index=iter->index)<iter->limit) { |
| 253 | return utf16BEIteratorGet(iter, index); |
| 254 | } else { |
| 255 | return U_SENTINEL; |
| 256 | } |
| 257 | } |
| 258 | |
| 259 | static UChar32 U_CALLCONV |
| 260 | utf16BEIteratorNext(UCharIterator *iter) { |
| 261 | int32_t index; |
| 262 | |
| 263 | if((index=iter->index)<iter->limit) { |
| 264 | iter->index=index+1; |
| 265 | return utf16BEIteratorGet(iter, index); |
| 266 | } else { |
| 267 | return U_SENTINEL; |
| 268 | } |
| 269 | } |
| 270 | |
| 271 | static UChar32 U_CALLCONV |
| 272 | utf16BEIteratorPrevious(UCharIterator *iter) { |
| 273 | int32_t index; |
| 274 | |
| 275 | if((index=iter->index)>iter->start) { |
| 276 | iter->index=--index; |
| 277 | return utf16BEIteratorGet(iter, index); |
| 278 | } else { |
| 279 | return U_SENTINEL; |
| 280 | } |
| 281 | } |
| 282 | |
| 283 | static const UCharIterator utf16BEIterator={ |
| 284 | 0, 0, 0, 0, 0, 0, |
| 285 | stringIteratorGetIndex, |
| 286 | stringIteratorMove, |
| 287 | stringIteratorHasNext, |
| 288 | stringIteratorHasPrevious, |
| 289 | utf16BEIteratorCurrent, |
| 290 | utf16BEIteratorNext, |
| 291 | utf16BEIteratorPrevious, |
| 292 | NULL, |
| 293 | stringIteratorGetState, |
| 294 | stringIteratorSetState |
| 295 | }; |
| 296 | |
| 297 | /* |
| 298 | * Count the number of UChars in a UTF-16BE string before a terminating UChar NUL, |
| 299 | * i.e., before a pair of 0 bytes where the first 0 byte is at an even |
| 300 | * offset from s. |
| 301 | */ |
| 302 | static int32_t |
| 303 | utf16BE_strlen(const char *s) { |
| 304 | if(IS_POINTER_EVEN(s)) { |
| 305 | /* |
| 306 | * even-aligned, call u_strlen(s) |
| 307 | * we are probably on a little-endian machine, but searching for UChar NUL |
| 308 | * does not care about endianness |
| 309 | */ |
| 310 | return u_strlen((const UChar *)s); |
| 311 | } else { |
| 312 | /* odd-aligned, search for pair of 0 bytes */ |
| 313 | const char *p=s; |
| 314 | |
| 315 | while(!(*p==0 && p[1]==0)) { |
| 316 | p+=2; |
| 317 | } |
| 318 | return (int32_t)((p-s)/2); |
| 319 | } |
| 320 | } |
| 321 | |
| 322 | U_CAPI void U_EXPORT2 |
| 323 | uiter_setUTF16BE(UCharIterator *iter, const char *s, int32_t length) { |
| 324 | if(iter!=NULL) { |
| 325 | /* allow only even-length strings (the input length counts bytes) */ |
| 326 | if(s!=NULL && (length==-1 || (length>=0 && IS_EVEN(length)))) { |
| 327 | /* length/=2, except that >>=1 also works for -1 (-1/2==0, -1>>1==-1) */ |
| 328 | length>>=1; |
| 329 | |
| 330 | if(U_IS_BIG_ENDIAN && IS_POINTER_EVEN(s)) { |
| 331 | /* big-endian machine and 2-aligned UTF-16BE string: use normal UChar iterator */ |
| 332 | uiter_setString(iter, (const UChar *)s, length); |
| 333 | return; |
| 334 | } |
| 335 | |
| 336 | *iter=utf16BEIterator; |
| 337 | iter->context=s; |
| 338 | if(length>=0) { |
| 339 | iter->length=length; |
| 340 | } else { |
| 341 | iter->length=utf16BE_strlen(s); |
| 342 | } |
| 343 | iter->limit=iter->length; |
| 344 | } else { |
| 345 | *iter=noopIterator; |
| 346 | } |
| 347 | } |
| 348 | } |
| 349 | |
| 350 | /* UCharIterator wrapper around CharacterIterator --------------------------- */ |
| 351 | |
| 352 | /* |
| 353 | * This is wrapper code around a C++ CharacterIterator to |
| 354 | * look like a C UCharIterator. |
| 355 | * |
| 356 | * The UCharIterator.context field holds a pointer to the CharacterIterator. |
| 357 | */ |
| 358 | |
| 359 | static int32_t U_CALLCONV |
| 360 | characterIteratorGetIndex(UCharIterator *iter, UCharIteratorOrigin origin) { |
| 361 | switch(origin) { |
| 362 | case UITER_ZERO: |
| 363 | return 0; |
| 364 | case UITER_START: |
| 365 | return ((CharacterIterator *)(iter->context))->startIndex(); |
| 366 | case UITER_CURRENT: |
| 367 | return ((CharacterIterator *)(iter->context))->getIndex(); |
| 368 | case UITER_LIMIT: |
| 369 | return ((CharacterIterator *)(iter->context))->endIndex(); |
| 370 | case UITER_LENGTH: |
| 371 | return ((CharacterIterator *)(iter->context))->getLength(); |
| 372 | default: |
| 373 | /* not a valid origin */ |
| 374 | /* Should never get here! */ |
| 375 | return -1; |
| 376 | } |
| 377 | } |
| 378 | |
| 379 | static int32_t U_CALLCONV |
| 380 | characterIteratorMove(UCharIterator *iter, int32_t delta, UCharIteratorOrigin origin) { |
| 381 | switch(origin) { |
| 382 | case UITER_ZERO: |
| 383 | ((CharacterIterator *)(iter->context))->setIndex(delta); |
| 384 | return ((CharacterIterator *)(iter->context))->getIndex(); |
| 385 | case UITER_START: |
| 386 | case UITER_CURRENT: |
| 387 | case UITER_LIMIT: |
| 388 | return ((CharacterIterator *)(iter->context))->move(delta, (CharacterIterator::EOrigin)origin); |
| 389 | case UITER_LENGTH: |
| 390 | ((CharacterIterator *)(iter->context))->setIndex(((CharacterIterator *)(iter->context))->getLength()+delta); |
| 391 | return ((CharacterIterator *)(iter->context))->getIndex(); |
| 392 | default: |
| 393 | /* not a valid origin */ |
| 394 | /* Should never get here! */ |
| 395 | return -1; |
| 396 | } |
| 397 | } |
| 398 | |
| 399 | static UBool U_CALLCONV |
| 400 | characterIteratorHasNext(UCharIterator *iter) { |
| 401 | return ((CharacterIterator *)(iter->context))->hasNext(); |
| 402 | } |
| 403 | |
| 404 | static UBool U_CALLCONV |
| 405 | characterIteratorHasPrevious(UCharIterator *iter) { |
| 406 | return ((CharacterIterator *)(iter->context))->hasPrevious(); |
| 407 | } |
| 408 | |
| 409 | static UChar32 U_CALLCONV |
| 410 | characterIteratorCurrent(UCharIterator *iter) { |
| 411 | UChar32 c; |
| 412 | |
| 413 | c=((CharacterIterator *)(iter->context))->current(); |
| 414 | if(c!=0xffff || ((CharacterIterator *)(iter->context))->hasNext()) { |
| 415 | return c; |
| 416 | } else { |
| 417 | return U_SENTINEL; |
| 418 | } |
| 419 | } |
| 420 | |
| 421 | static UChar32 U_CALLCONV |
| 422 | characterIteratorNext(UCharIterator *iter) { |
| 423 | if(((CharacterIterator *)(iter->context))->hasNext()) { |
| 424 | return ((CharacterIterator *)(iter->context))->nextPostInc(); |
| 425 | } else { |
| 426 | return U_SENTINEL; |
| 427 | } |
| 428 | } |
| 429 | |
| 430 | static UChar32 U_CALLCONV |
| 431 | characterIteratorPrevious(UCharIterator *iter) { |
| 432 | if(((CharacterIterator *)(iter->context))->hasPrevious()) { |
| 433 | return ((CharacterIterator *)(iter->context))->previous(); |
| 434 | } else { |
| 435 | return U_SENTINEL; |
| 436 | } |
| 437 | } |
| 438 | |
| 439 | static uint32_t U_CALLCONV |
| 440 | characterIteratorGetState(const UCharIterator *iter) { |
| 441 | return ((CharacterIterator *)(iter->context))->getIndex(); |
| 442 | } |
| 443 | |
| 444 | static void U_CALLCONV |
| 445 | characterIteratorSetState(UCharIterator *iter, uint32_t state, UErrorCode *pErrorCode) { |
| 446 | if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) { |
| 447 | /* do nothing */ |
| 448 | } else if(iter==NULL || iter->context==NULL) { |
| 449 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
| 450 | } else if((int32_t)state<((CharacterIterator *)(iter->context))->startIndex() || ((CharacterIterator *)(iter->context))->endIndex()<(int32_t)state) { |
| 451 | *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
| 452 | } else { |
| 453 | ((CharacterIterator *)(iter->context))->setIndex((int32_t)state); |
| 454 | } |
| 455 | } |
| 456 | |
| 457 | static const UCharIterator characterIteratorWrapper={ |
| 458 | 0, 0, 0, 0, 0, 0, |
| 459 | characterIteratorGetIndex, |
| 460 | characterIteratorMove, |
| 461 | characterIteratorHasNext, |
| 462 | characterIteratorHasPrevious, |
| 463 | characterIteratorCurrent, |
| 464 | characterIteratorNext, |
| 465 | characterIteratorPrevious, |
| 466 | NULL, |
| 467 | characterIteratorGetState, |
| 468 | characterIteratorSetState |
| 469 | }; |
| 470 | |
| 471 | U_CAPI void U_EXPORT2 |
| 472 | uiter_setCharacterIterator(UCharIterator *iter, CharacterIterator *charIter) { |
| 473 | if(iter!=0) { |
| 474 | if(charIter!=0) { |
| 475 | *iter=characterIteratorWrapper; |
| 476 | iter->context=charIter; |
| 477 | } else { |
| 478 | *iter=noopIterator; |
| 479 | } |
| 480 | } |
| 481 | } |
| 482 | |
| 483 | /* UCharIterator wrapper around Replaceable --------------------------------- */ |
| 484 | |
| 485 | /* |
| 486 | * This is an implementation of a code unit (UChar) iterator |
| 487 | * based on a Replaceable object. |
| 488 | * |
| 489 | * The UCharIterator.context field holds a pointer to the Replaceable. |
| 490 | * UCharIterator.length and UCharIterator.index hold Replaceable.length() |
| 491 | * and the iteration index. |
| 492 | */ |
| 493 | |
| 494 | static UChar32 U_CALLCONV |
| 495 | replaceableIteratorCurrent(UCharIterator *iter) { |
| 496 | if(iter->index<iter->limit) { |
| 497 | return ((Replaceable *)(iter->context))->charAt(iter->index); |
| 498 | } else { |
| 499 | return U_SENTINEL; |
| 500 | } |
| 501 | } |
| 502 | |
| 503 | static UChar32 U_CALLCONV |
| 504 | replaceableIteratorNext(UCharIterator *iter) { |
| 505 | if(iter->index<iter->limit) { |
| 506 | return ((Replaceable *)(iter->context))->charAt(iter->index++); |
| 507 | } else { |
| 508 | return U_SENTINEL; |
| 509 | } |
| 510 | } |
| 511 | |
| 512 | static UChar32 U_CALLCONV |
| 513 | replaceableIteratorPrevious(UCharIterator *iter) { |
| 514 | if(iter->index>iter->start) { |
| 515 | return ((Replaceable *)(iter->context))->charAt(--iter->index); |
| 516 | } else { |
| 517 | return U_SENTINEL; |
| 518 | } |
| 519 | } |
| 520 | |
| 521 | static const UCharIterator replaceableIterator={ |
| 522 | 0, 0, 0, 0, 0, 0, |
| 523 | stringIteratorGetIndex, |
| 524 | stringIteratorMove, |
| 525 | stringIteratorHasNext, |
| 526 | stringIteratorHasPrevious, |
| 527 | replaceableIteratorCurrent, |
| 528 | replaceableIteratorNext, |
| 529 | replaceableIteratorPrevious, |
| 530 | NULL, |
| 531 | stringIteratorGetState, |
| 532 | stringIteratorSetState |
| 533 | }; |
| 534 | |
| 535 | U_CAPI void U_EXPORT2 |
| 536 | uiter_setReplaceable(UCharIterator *iter, const Replaceable *rep) { |
| 537 | if(iter!=0) { |
| 538 | if(rep!=0) { |
| 539 | *iter=replaceableIterator; |
| 540 | iter->context=rep; |
| 541 | iter->limit=iter->length=rep->length(); |
| 542 | } else { |
| 543 | *iter=noopIterator; |
| 544 | } |
| 545 | } |
| 546 | } |
| 547 | |
| 548 | /* UCharIterator implementation for UTF-8 strings --------------------------- */ |
| 549 | |
| 550 | /* |
| 551 | * Possible, probably necessary only for an implementation for arbitrary |
| 552 | * converters: |
| 553 | * Maintain a buffer (ring buffer?) for a piece of converted 16-bit text. |
| 554 | * This would require to turn reservedFn into a close function and |
| 555 | * to introduce a uiter_close(iter). |
| 556 | */ |
| 557 | |
| 558 | #define UITER_CNV_CAPACITY 16 |
| 559 | |
| 560 | /* |
| 561 | * Minimal implementation: |
| 562 | * Maintain a single-UChar buffer for an additional surrogate. |
| 563 | * The caller must not modify start and limit because they are used internally. |
| 564 | * |
| 565 | * Use UCharIterator fields as follows: |
| 566 | * context pointer to UTF-8 string |
| 567 | * length UTF-16 length of the string; -1 until lazy evaluation |
| 568 | * start current UTF-8 index |
| 569 | * index current UTF-16 index; may be -1="unknown" after setState() |
| 570 | * limit UTF-8 length of the string |
| 571 | * reservedField supplementary code point |
| 572 | * |
| 573 | * Since UCharIterator delivers 16-bit code units, the iteration can be |
| 574 | * currently in the middle of the byte sequence for a supplementary code point. |
| 575 | * In this case, reservedField will contain that code point and start will |
| 576 | * point to after the corresponding byte sequence. The UTF-16 index will be |
| 577 | * one less than what it would otherwise be corresponding to the UTF-8 index. |
| 578 | * Otherwise, reservedField will be 0. |
| 579 | */ |
| 580 | |
| 581 | /* |
| 582 | * Possible optimization for NUL-terminated UTF-8 and UTF-16 strings: |
| 583 | * Add implementations that do not call strlen() for iteration but check for NUL. |
| 584 | */ |
| 585 | |
| 586 | static int32_t U_CALLCONV |
| 587 | utf8IteratorGetIndex(UCharIterator *iter, UCharIteratorOrigin origin) { |
| 588 | switch(origin) { |
| 589 | case UITER_ZERO: |
| 590 | case UITER_START: |
| 591 | return 0; |
| 592 | case UITER_CURRENT: |
| 593 | if(iter->index<0) { |
| 594 | /* the current UTF-16 index is unknown after setState(), count from the beginning */ |
| 595 | const uint8_t *s; |
| 596 | UChar32 c; |
| 597 | int32_t i, limit, index; |
| 598 | |
| 599 | s=(const uint8_t *)iter->context; |
| 600 | i=index=0; |
| 601 | limit=iter->start; /* count up to the UTF-8 index */ |
| 602 | while(i<limit) { |
| 603 | U8_NEXT_OR_FFFD(s, i, limit, c); |
| 604 | index+=U16_LENGTH(c); |
| 605 | } |
| 606 | |
| 607 | iter->start=i; /* just in case setState() did not get us to a code point boundary */ |
| 608 | if(i==iter->limit) { |
| 609 | iter->length=index; /* in case it was <0 or wrong */ |
| 610 | } |
| 611 | if(iter->reservedField!=0) { |
| 612 | --index; /* we are in the middle of a supplementary code point */ |
| 613 | } |
| 614 | iter->index=index; |
| 615 | } |
| 616 | return iter->index; |
| 617 | case UITER_LIMIT: |
| 618 | case UITER_LENGTH: |
| 619 | if(iter->length<0) { |
| 620 | const uint8_t *s; |
| 621 | UChar32 c; |
| 622 | int32_t i, limit, length; |
| 623 | |
| 624 | s=(const uint8_t *)iter->context; |
| 625 | if(iter->index<0) { |
| 626 | /* |
| 627 | * the current UTF-16 index is unknown after setState(), |
| 628 | * we must first count from the beginning to here |
| 629 | */ |
| 630 | i=length=0; |
| 631 | limit=iter->start; |
| 632 | |
| 633 | /* count from the beginning to the current index */ |
| 634 | while(i<limit) { |
| 635 | U8_NEXT_OR_FFFD(s, i, limit, c); |
| 636 | length+=U16_LENGTH(c); |
| 637 | } |
| 638 | |
| 639 | /* assume i==limit==iter->start, set the UTF-16 index */ |
| 640 | iter->start=i; /* just in case setState() did not get us to a code point boundary */ |
| 641 | iter->index= iter->reservedField!=0 ? length-1 : length; |
| 642 | } else { |
| 643 | i=iter->start; |
| 644 | length=iter->index; |
| 645 | if(iter->reservedField!=0) { |
| 646 | ++length; |
| 647 | } |
| 648 | } |
| 649 | |
| 650 | /* count from the current index to the end */ |
| 651 | limit=iter->limit; |
| 652 | while(i<limit) { |
| 653 | U8_NEXT_OR_FFFD(s, i, limit, c); |
| 654 | length+=U16_LENGTH(c); |
| 655 | } |
| 656 | iter->length=length; |
| 657 | } |
| 658 | return iter->length; |
| 659 | default: |
| 660 | /* not a valid origin */ |
| 661 | /* Should never get here! */ |
| 662 | return -1; |
| 663 | } |
| 664 | } |
| 665 | |
| 666 | static int32_t U_CALLCONV |
| 667 | utf8IteratorMove(UCharIterator *iter, int32_t delta, UCharIteratorOrigin origin) { |
| 668 | const uint8_t *s; |
| 669 | UChar32 c; |
| 670 | int32_t pos; /* requested UTF-16 index */ |
| 671 | int32_t i; /* UTF-8 index */ |
| 672 | UBool havePos; |
| 673 | |
| 674 | /* calculate the requested UTF-16 index */ |
| 675 | switch(origin) { |
| 676 | case UITER_ZERO: |
| 677 | case UITER_START: |
| 678 | pos=delta; |
| 679 | havePos=TRUE; |
| 680 | /* iter->index<0 (unknown) is possible */ |
| 681 | break; |
| 682 | case UITER_CURRENT: |
| 683 | if(iter->index>=0) { |
| 684 | pos=iter->index+delta; |
| 685 | havePos=TRUE; |
| 686 | } else { |
| 687 | /* the current UTF-16 index is unknown after setState(), use only delta */ |
| 688 | pos=0; |
| 689 | havePos=FALSE; |
| 690 | } |
| 691 | break; |
| 692 | case UITER_LIMIT: |
| 693 | case UITER_LENGTH: |
| 694 | if(iter->length>=0) { |
| 695 | pos=iter->length+delta; |
| 696 | havePos=TRUE; |
| 697 | } else { |
| 698 | /* pin to the end, avoid counting the length */ |
| 699 | iter->index=-1; |
| 700 | iter->start=iter->limit; |
| 701 | iter->reservedField=0; |
| 702 | if(delta>=0) { |
| 703 | return UITER_UNKNOWN_INDEX; |
| 704 | } else { |
| 705 | /* the current UTF-16 index is unknown, use only delta */ |
| 706 | pos=0; |
| 707 | havePos=FALSE; |
| 708 | } |
| 709 | } |
| 710 | break; |
| 711 | default: |
| 712 | return -1; /* Error */ |
| 713 | } |
| 714 | |
| 715 | if(havePos) { |
| 716 | /* shortcuts: pinning to the edges of the string */ |
| 717 | if(pos<=0) { |
| 718 | iter->index=iter->start=iter->reservedField=0; |
| 719 | return 0; |
| 720 | } else if(iter->length>=0 && pos>=iter->length) { |
| 721 | iter->index=iter->length; |
| 722 | iter->start=iter->limit; |
| 723 | iter->reservedField=0; |
| 724 | return iter->index; |
| 725 | } |
| 726 | |
| 727 | /* minimize the number of U8_NEXT/PREV operations */ |
| 728 | if(iter->index<0 || pos<iter->index/2) { |
| 729 | /* go forward from the start instead of backward from the current index */ |
| 730 | iter->index=iter->start=iter->reservedField=0; |
| 731 | } else if(iter->length>=0 && (iter->length-pos)<(pos-iter->index)) { |
| 732 | /* |
| 733 | * if we have the UTF-16 index and length and the new position is |
| 734 | * closer to the end than the current index, |
| 735 | * then go backward from the end instead of forward from the current index |
| 736 | */ |
| 737 | iter->index=iter->length; |
| 738 | iter->start=iter->limit; |
| 739 | iter->reservedField=0; |
| 740 | } |
| 741 | |
| 742 | delta=pos-iter->index; |
| 743 | if(delta==0) { |
| 744 | return iter->index; /* nothing to do */ |
| 745 | } |
| 746 | } else { |
| 747 | /* move relative to unknown UTF-16 index */ |
| 748 | if(delta==0) { |
| 749 | return UITER_UNKNOWN_INDEX; /* nothing to do */ |
| 750 | } else if(-delta>=iter->start) { |
| 751 | /* moving backwards by more UChars than there are UTF-8 bytes, pin to 0 */ |
| 752 | iter->index=iter->start=iter->reservedField=0; |
| 753 | return 0; |
| 754 | } else if(delta>=(iter->limit-iter->start)) { |
| 755 | /* moving forward by more UChars than the remaining UTF-8 bytes, pin to the end */ |
| 756 | iter->index=iter->length; /* may or may not be <0 (unknown) */ |
| 757 | iter->start=iter->limit; |
| 758 | iter->reservedField=0; |
| 759 | return iter->index>=0 ? iter->index : (int32_t)UITER_UNKNOWN_INDEX; |
| 760 | } |
| 761 | } |
| 762 | |
| 763 | /* delta!=0 */ |
| 764 | |
| 765 | /* move towards the requested position, pin to the edges of the string */ |
| 766 | s=(const uint8_t *)iter->context; |
| 767 | pos=iter->index; /* could be <0 (unknown) */ |
| 768 | i=iter->start; |
| 769 | if(delta>0) { |
| 770 | /* go forward */ |
| 771 | int32_t limit=iter->limit; |
| 772 | if(iter->reservedField!=0) { |
| 773 | iter->reservedField=0; |
| 774 | ++pos; |
| 775 | --delta; |
| 776 | } |
| 777 | while(delta>0 && i<limit) { |
| 778 | U8_NEXT_OR_FFFD(s, i, limit, c); |
| 779 | if(c<=0xffff) { |
| 780 | ++pos; |
| 781 | --delta; |
| 782 | } else if(delta>=2) { |
| 783 | pos+=2; |
| 784 | delta-=2; |
| 785 | } else /* delta==1 */ { |
| 786 | /* stop in the middle of a supplementary code point */ |
| 787 | iter->reservedField=c; |
| 788 | ++pos; |
| 789 | break; /* delta=0; */ |
| 790 | } |
| 791 | } |
| 792 | if(i==limit) { |
| 793 | if(iter->length<0 && iter->index>=0) { |
| 794 | iter->length= iter->reservedField==0 ? pos : pos+1; |
| 795 | } else if(iter->index<0 && iter->length>=0) { |
| 796 | iter->index= iter->reservedField==0 ? iter->length : iter->length-1; |
| 797 | } |
| 798 | } |
| 799 | } else /* delta<0 */ { |
| 800 | /* go backward */ |
| 801 | if(iter->reservedField!=0) { |
| 802 | iter->reservedField=0; |
| 803 | i-=4; /* we stayed behind the supplementary code point; go before it now */ |
| 804 | --pos; |
| 805 | ++delta; |
| 806 | } |
| 807 | while(delta<0 && i>0) { |
| 808 | U8_PREV_OR_FFFD(s, 0, i, c); |
| 809 | if(c<=0xffff) { |
| 810 | --pos; |
| 811 | ++delta; |
| 812 | } else if(delta<=-2) { |
| 813 | pos-=2; |
| 814 | delta+=2; |
| 815 | } else /* delta==-1 */ { |
| 816 | /* stop in the middle of a supplementary code point */ |
| 817 | i+=4; /* back to behind this supplementary code point for consistent state */ |
| 818 | iter->reservedField=c; |
| 819 | --pos; |
| 820 | break; /* delta=0; */ |
| 821 | } |
| 822 | } |
| 823 | } |
| 824 | |
| 825 | iter->start=i; |
| 826 | if(iter->index>=0) { |
| 827 | return iter->index=pos; |
| 828 | } else { |
| 829 | /* we started with index<0 (unknown) so pos is bogus */ |
| 830 | if(i<=1) { |
| 831 | return iter->index=i; /* reached the beginning */ |
| 832 | } else { |
| 833 | /* we still don't know the UTF-16 index */ |
| 834 | return UITER_UNKNOWN_INDEX; |
| 835 | } |
| 836 | } |
| 837 | } |
| 838 | |
| 839 | static UBool U_CALLCONV |
| 840 | utf8IteratorHasNext(UCharIterator *iter) { |
| 841 | return iter->start<iter->limit || iter->reservedField!=0; |
| 842 | } |
| 843 | |
| 844 | static UBool U_CALLCONV |
| 845 | utf8IteratorHasPrevious(UCharIterator *iter) { |
| 846 | return iter->start>0; |
| 847 | } |
| 848 | |
| 849 | static UChar32 U_CALLCONV |
| 850 | utf8IteratorCurrent(UCharIterator *iter) { |
| 851 | if(iter->reservedField!=0) { |
| 852 | return U16_TRAIL(iter->reservedField); |
| 853 | } else if(iter->start<iter->limit) { |
| 854 | const uint8_t *s=(const uint8_t *)iter->context; |
| 855 | UChar32 c; |
| 856 | int32_t i=iter->start; |
| 857 | |
| 858 | U8_NEXT_OR_FFFD(s, i, iter->limit, c); |
| 859 | if(c<=0xffff) { |
| 860 | return c; |
| 861 | } else { |
| 862 | return U16_LEAD(c); |
| 863 | } |
| 864 | } else { |
| 865 | return U_SENTINEL; |
| 866 | } |
| 867 | } |
| 868 | |
| 869 | static UChar32 U_CALLCONV |
| 870 | utf8IteratorNext(UCharIterator *iter) { |
| 871 | int32_t index; |
| 872 | |
| 873 | if(iter->reservedField!=0) { |
| 874 | UChar trail=U16_TRAIL(iter->reservedField); |
| 875 | iter->reservedField=0; |
| 876 | if((index=iter->index)>=0) { |
| 877 | iter->index=index+1; |
| 878 | } |
| 879 | return trail; |
| 880 | } else if(iter->start<iter->limit) { |
| 881 | const uint8_t *s=(const uint8_t *)iter->context; |
| 882 | UChar32 c; |
| 883 | |
| 884 | U8_NEXT_OR_FFFD(s, iter->start, iter->limit, c); |
| 885 | if((index=iter->index)>=0) { |
| 886 | iter->index=++index; |
| 887 | if(iter->length<0 && iter->start==iter->limit) { |
| 888 | iter->length= c<=0xffff ? index : index+1; |
| 889 | } |
| 890 | } else if(iter->start==iter->limit && iter->length>=0) { |
| 891 | iter->index= c<=0xffff ? iter->length : iter->length-1; |
| 892 | } |
| 893 | if(c<=0xffff) { |
| 894 | return c; |
| 895 | } else { |
| 896 | iter->reservedField=c; |
| 897 | return U16_LEAD(c); |
| 898 | } |
| 899 | } else { |
| 900 | return U_SENTINEL; |
| 901 | } |
| 902 | } |
| 903 | |
| 904 | static UChar32 U_CALLCONV |
| 905 | utf8IteratorPrevious(UCharIterator *iter) { |
| 906 | int32_t index; |
| 907 | |
| 908 | if(iter->reservedField!=0) { |
| 909 | UChar lead=U16_LEAD(iter->reservedField); |
| 910 | iter->reservedField=0; |
| 911 | iter->start-=4; /* we stayed behind the supplementary code point; go before it now */ |
| 912 | if((index=iter->index)>0) { |
| 913 | iter->index=index-1; |
| 914 | } |
| 915 | return lead; |
| 916 | } else if(iter->start>0) { |
| 917 | const uint8_t *s=(const uint8_t *)iter->context; |
| 918 | UChar32 c; |
| 919 | |
| 920 | U8_PREV_OR_FFFD(s, 0, iter->start, c); |
| 921 | if((index=iter->index)>0) { |
| 922 | iter->index=index-1; |
| 923 | } else if(iter->start<=1) { |
| 924 | iter->index= c<=0xffff ? iter->start : iter->start+1; |
| 925 | } |
| 926 | if(c<=0xffff) { |
| 927 | return c; |
| 928 | } else { |
| 929 | iter->start+=4; /* back to behind this supplementary code point for consistent state */ |
| 930 | iter->reservedField=c; |
| 931 | return U16_TRAIL(c); |
| 932 | } |
| 933 | } else { |
| 934 | return U_SENTINEL; |
| 935 | } |
| 936 | } |
| 937 | |
| 938 | static uint32_t U_CALLCONV |
| 939 | utf8IteratorGetState(const UCharIterator *iter) { |
| 940 | uint32_t state=(uint32_t)(iter->start<<1); |
| 941 | if(iter->reservedField!=0) { |
| 942 | state|=1; |
| 943 | } |
| 944 | return state; |
| 945 | } |
| 946 | |
| 947 | static void U_CALLCONV |
| 948 | utf8IteratorSetState(UCharIterator *iter, |
| 949 | uint32_t state, |
| 950 | UErrorCode *pErrorCode) |
| 951 | { |
| 952 | if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) { |
| 953 | /* do nothing */ |
| 954 | } else if(iter==NULL) { |
| 955 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
| 956 | } else if(state==utf8IteratorGetState(iter)) { |
| 957 | /* setting to the current state: no-op */ |
| 958 | } else { |
| 959 | int32_t index=(int32_t)(state>>1); /* UTF-8 index */ |
| 960 | state&=1; /* 1 if in surrogate pair, must be index>=4 */ |
| 961 | |
| 962 | if((state==0 ? index<0 : index<4) || iter->limit<index) { |
| 963 | *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
| 964 | } else { |
| 965 | iter->start=index; /* restore UTF-8 byte index */ |
| 966 | if(index<=1) { |
| 967 | iter->index=index; |
| 968 | } else { |
| 969 | iter->index=-1; /* unknown UTF-16 index */ |
| 970 | } |
| 971 | if(state==0) { |
| 972 | iter->reservedField=0; |
| 973 | } else { |
| 974 | /* verified index>=4 above */ |
| 975 | UChar32 c; |
| 976 | U8_PREV_OR_FFFD((const uint8_t *)iter->context, 0, index, c); |
| 977 | if(c<=0xffff) { |
| 978 | *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
| 979 | } else { |
| 980 | iter->reservedField=c; |
| 981 | } |
| 982 | } |
| 983 | } |
| 984 | } |
| 985 | } |
| 986 | |
| 987 | static const UCharIterator utf8Iterator={ |
| 988 | 0, 0, 0, 0, 0, 0, |
| 989 | utf8IteratorGetIndex, |
| 990 | utf8IteratorMove, |
| 991 | utf8IteratorHasNext, |
| 992 | utf8IteratorHasPrevious, |
| 993 | utf8IteratorCurrent, |
| 994 | utf8IteratorNext, |
| 995 | utf8IteratorPrevious, |
| 996 | NULL, |
| 997 | utf8IteratorGetState, |
| 998 | utf8IteratorSetState |
| 999 | }; |
| 1000 | |
| 1001 | U_CAPI void U_EXPORT2 |
| 1002 | uiter_setUTF8(UCharIterator *iter, const char *s, int32_t length) { |
| 1003 | if(iter!=0) { |
| 1004 | if(s!=0 && length>=-1) { |
| 1005 | *iter=utf8Iterator; |
| 1006 | iter->context=s; |
| 1007 | if(length>=0) { |
| 1008 | iter->limit=length; |
| 1009 | } else { |
| 1010 | iter->limit=(int32_t)uprv_strlen(s); |
| 1011 | } |
| 1012 | iter->length= iter->limit<=1 ? iter->limit : -1; |
| 1013 | } else { |
| 1014 | *iter=noopIterator; |
| 1015 | } |
| 1016 | } |
| 1017 | } |
| 1018 | |
| 1019 | /* Helper functions --------------------------------------------------------- */ |
| 1020 | |
| 1021 | U_CAPI UChar32 U_EXPORT2 |
| 1022 | uiter_current32(UCharIterator *iter) { |
| 1023 | UChar32 c, c2; |
| 1024 | |
| 1025 | c=iter->current(iter); |
| 1026 | if(U16_IS_SURROGATE(c)) { |
| 1027 | if(U16_IS_SURROGATE_LEAD(c)) { |
| 1028 | /* |
| 1029 | * go to the next code unit |
| 1030 | * we know that we are not at the limit because c!=U_SENTINEL |
| 1031 | */ |
| 1032 | iter->move(iter, 1, UITER_CURRENT); |
| 1033 | if(U16_IS_TRAIL(c2=iter->current(iter))) { |
| 1034 | c=U16_GET_SUPPLEMENTARY(c, c2); |
| 1035 | } |
| 1036 | |
| 1037 | /* undo index movement */ |
| 1038 | iter->move(iter, -1, UITER_CURRENT); |
| 1039 | } else { |
| 1040 | if(U16_IS_LEAD(c2=iter->previous(iter))) { |
| 1041 | c=U16_GET_SUPPLEMENTARY(c2, c); |
| 1042 | } |
| 1043 | if(c2>=0) { |
| 1044 | /* undo index movement */ |
| 1045 | iter->move(iter, 1, UITER_CURRENT); |
| 1046 | } |
| 1047 | } |
| 1048 | } |
| 1049 | return c; |
| 1050 | } |
| 1051 | |
| 1052 | U_CAPI UChar32 U_EXPORT2 |
| 1053 | uiter_next32(UCharIterator *iter) { |
| 1054 | UChar32 c, c2; |
| 1055 | |
| 1056 | c=iter->next(iter); |
| 1057 | if(U16_IS_LEAD(c)) { |
| 1058 | if(U16_IS_TRAIL(c2=iter->next(iter))) { |
| 1059 | c=U16_GET_SUPPLEMENTARY(c, c2); |
| 1060 | } else if(c2>=0) { |
| 1061 | /* unmatched first surrogate, undo index movement */ |
| 1062 | iter->move(iter, -1, UITER_CURRENT); |
| 1063 | } |
| 1064 | } |
| 1065 | return c; |
| 1066 | } |
| 1067 | |
| 1068 | U_CAPI UChar32 U_EXPORT2 |
| 1069 | uiter_previous32(UCharIterator *iter) { |
| 1070 | UChar32 c, c2; |
| 1071 | |
| 1072 | c=iter->previous(iter); |
| 1073 | if(U16_IS_TRAIL(c)) { |
| 1074 | if(U16_IS_LEAD(c2=iter->previous(iter))) { |
| 1075 | c=U16_GET_SUPPLEMENTARY(c2, c); |
| 1076 | } else if(c2>=0) { |
| 1077 | /* unmatched second surrogate, undo index movement */ |
| 1078 | iter->move(iter, 1, UITER_CURRENT); |
| 1079 | } |
| 1080 | } |
| 1081 | return c; |
| 1082 | } |
| 1083 | |
| 1084 | U_CAPI uint32_t U_EXPORT2 |
| 1085 | uiter_getState(const UCharIterator *iter) { |
| 1086 | if(iter==NULL || iter->getState==NULL) { |
| 1087 | return UITER_NO_STATE; |
| 1088 | } else { |
| 1089 | return iter->getState(iter); |
| 1090 | } |
| 1091 | } |
| 1092 | |
| 1093 | U_CAPI void U_EXPORT2 |
| 1094 | uiter_setState(UCharIterator *iter, uint32_t state, UErrorCode *pErrorCode) { |
| 1095 | if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) { |
| 1096 | /* do nothing */ |
| 1097 | } else if(iter==NULL) { |
| 1098 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
| 1099 | } else if(iter->setState==NULL) { |
| 1100 | *pErrorCode=U_UNSUPPORTED_ERROR; |
| 1101 | } else { |
| 1102 | iter->setState(iter, state, pErrorCode); |
| 1103 | } |
| 1104 | } |
| 1105 | |
| 1106 | U_CDECL_END |