| 1 | // © 2016 and later: Unicode, Inc. and others. |
| 2 | // License & terms of use: http://www.unicode.org/copyright.html |
| 3 | /* |
| 4 | ******************************************************************************* |
| 5 | * Copyright (C) 2013-2014, International Business Machines |
| 6 | * Corporation and others. All Rights Reserved. |
| 7 | ******************************************************************************* |
| 8 | * collationsets.cpp |
| 9 | * |
| 10 | * created on: 2013feb09 |
| 11 | * created by: Markus W. Scherer |
| 12 | */ |
| 13 | |
| 14 | #include "unicode/utypes.h" |
| 15 | |
| 16 | #if !UCONFIG_NO_COLLATION |
| 17 | |
| 18 | #include "unicode/ucharstrie.h" |
| 19 | #include "unicode/uniset.h" |
| 20 | #include "unicode/unistr.h" |
| 21 | #include "unicode/ustringtrie.h" |
| 22 | #include "collation.h" |
| 23 | #include "collationdata.h" |
| 24 | #include "collationsets.h" |
| 25 | #include "normalizer2impl.h" |
| 26 | #include "uassert.h" |
| 27 | #include "utf16collationiterator.h" |
| 28 | #include "utrie2.h" |
| 29 | |
| 30 | U_NAMESPACE_BEGIN |
| 31 | |
| 32 | U_CDECL_BEGIN |
| 33 | |
| 34 | static UBool U_CALLCONV |
| 35 | enumTailoredRange(const void *context, UChar32 start, UChar32 end, uint32_t ce32) { |
| 36 | if(ce32 == Collation::FALLBACK_CE32) { |
| 37 | return TRUE; // fallback to base, not tailored |
| 38 | } |
| 39 | TailoredSet *ts = (TailoredSet *)context; |
| 40 | return ts->handleCE32(start, end, ce32); |
| 41 | } |
| 42 | |
| 43 | U_CDECL_END |
| 44 | |
| 45 | void |
| 46 | TailoredSet::forData(const CollationData *d, UErrorCode &ec) { |
| 47 | if(U_FAILURE(ec)) { return; } |
| 48 | errorCode = ec; // Preserve info & warning codes. |
| 49 | data = d; |
| 50 | baseData = d->base; |
| 51 | U_ASSERT(baseData != NULL); |
| 52 | utrie2_enum(data->trie, NULL, enumTailoredRange, this); |
| 53 | ec = errorCode; |
| 54 | } |
| 55 | |
| 56 | UBool |
| 57 | TailoredSet::handleCE32(UChar32 start, UChar32 end, uint32_t ce32) { |
| 58 | U_ASSERT(ce32 != Collation::FALLBACK_CE32); |
| 59 | if(Collation::isSpecialCE32(ce32)) { |
| 60 | ce32 = data->getIndirectCE32(ce32); |
| 61 | if(ce32 == Collation::FALLBACK_CE32) { |
| 62 | return U_SUCCESS(errorCode); |
| 63 | } |
| 64 | } |
| 65 | do { |
| 66 | uint32_t baseCE32 = baseData->getFinalCE32(baseData->getCE32(start)); |
| 67 | // Do not just continue if ce32 == baseCE32 because |
| 68 | // contractions and expansions in different data objects |
| 69 | // normally differ even if they have the same data offsets. |
| 70 | if(Collation::isSelfContainedCE32(ce32) && Collation::isSelfContainedCE32(baseCE32)) { |
| 71 | // fastpath |
| 72 | if(ce32 != baseCE32) { |
| 73 | tailored->add(start); |
| 74 | } |
| 75 | } else { |
| 76 | compare(start, ce32, baseCE32); |
| 77 | } |
| 78 | } while(++start <= end); |
| 79 | return U_SUCCESS(errorCode); |
| 80 | } |
| 81 | |
| 82 | void |
| 83 | TailoredSet::compare(UChar32 c, uint32_t ce32, uint32_t baseCE32) { |
| 84 | if(Collation::isPrefixCE32(ce32)) { |
| 85 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); |
| 86 | ce32 = data->getFinalCE32(CollationData::readCE32(p)); |
| 87 | if(Collation::isPrefixCE32(baseCE32)) { |
| 88 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); |
| 89 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); |
| 90 | comparePrefixes(c, p + 2, q + 2); |
| 91 | } else { |
| 92 | addPrefixes(data, c, p + 2); |
| 93 | } |
| 94 | } else if(Collation::isPrefixCE32(baseCE32)) { |
| 95 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); |
| 96 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); |
| 97 | addPrefixes(baseData, c, q + 2); |
| 98 | } |
| 99 | |
| 100 | if(Collation::isContractionCE32(ce32)) { |
| 101 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); |
| 102 | if((ce32 & Collation::CONTRACT_SINGLE_CP_NO_MATCH) != 0) { |
| 103 | ce32 = Collation::NO_CE32; |
| 104 | } else { |
| 105 | ce32 = data->getFinalCE32(CollationData::readCE32(p)); |
| 106 | } |
| 107 | if(Collation::isContractionCE32(baseCE32)) { |
| 108 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); |
| 109 | if((baseCE32 & Collation::CONTRACT_SINGLE_CP_NO_MATCH) != 0) { |
| 110 | baseCE32 = Collation::NO_CE32; |
| 111 | } else { |
| 112 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); |
| 113 | } |
| 114 | compareContractions(c, p + 2, q + 2); |
| 115 | } else { |
| 116 | addContractions(c, p + 2); |
| 117 | } |
| 118 | } else if(Collation::isContractionCE32(baseCE32)) { |
| 119 | const UChar *q = baseData->contexts + Collation::indexFromCE32(baseCE32); |
| 120 | baseCE32 = baseData->getFinalCE32(CollationData::readCE32(q)); |
| 121 | addContractions(c, q + 2); |
| 122 | } |
| 123 | |
| 124 | int32_t tag; |
| 125 | if(Collation::isSpecialCE32(ce32)) { |
| 126 | tag = Collation::tagFromCE32(ce32); |
| 127 | U_ASSERT(tag != Collation::PREFIX_TAG); |
| 128 | U_ASSERT(tag != Collation::CONTRACTION_TAG); |
| 129 | // Currently, the tailoring data builder does not write offset tags. |
| 130 | // They might be useful for saving space, |
| 131 | // but they would complicate the builder, |
| 132 | // and in tailorings we assume that performance of tailored characters is more important. |
| 133 | U_ASSERT(tag != Collation::OFFSET_TAG); |
| 134 | } else { |
| 135 | tag = -1; |
| 136 | } |
| 137 | int32_t baseTag; |
| 138 | if(Collation::isSpecialCE32(baseCE32)) { |
| 139 | baseTag = Collation::tagFromCE32(baseCE32); |
| 140 | U_ASSERT(baseTag != Collation::PREFIX_TAG); |
| 141 | U_ASSERT(baseTag != Collation::CONTRACTION_TAG); |
| 142 | } else { |
| 143 | baseTag = -1; |
| 144 | } |
| 145 | |
| 146 | // Non-contextual mappings, expansions, etc. |
| 147 | if(baseTag == Collation::OFFSET_TAG) { |
| 148 | // We might be comparing a tailoring CE which is a copy of |
| 149 | // a base offset-tag CE, via the [optimize [set]] syntax |
| 150 | // or when a single-character mapping was copied for tailored contractions. |
| 151 | // Offset tags always result in long-primary CEs, |
| 152 | // with common secondary/tertiary weights. |
| 153 | if(!Collation::isLongPrimaryCE32(ce32)) { |
| 154 | add(c); |
| 155 | return; |
| 156 | } |
| 157 | int64_t dataCE = baseData->ces[Collation::indexFromCE32(baseCE32)]; |
| 158 | uint32_t p = Collation::getThreeBytePrimaryForOffsetData(c, dataCE); |
| 159 | if(Collation::primaryFromLongPrimaryCE32(ce32) != p) { |
| 160 | add(c); |
| 161 | return; |
| 162 | } |
| 163 | } |
| 164 | |
| 165 | if(tag != baseTag) { |
| 166 | add(c); |
| 167 | return; |
| 168 | } |
| 169 | |
| 170 | if(tag == Collation::EXPANSION32_TAG) { |
| 171 | const uint32_t *ce32s = data->ce32s + Collation::indexFromCE32(ce32); |
| 172 | int32_t length = Collation::lengthFromCE32(ce32); |
| 173 | |
| 174 | const uint32_t *baseCE32s = baseData->ce32s + Collation::indexFromCE32(baseCE32); |
| 175 | int32_t baseLength = Collation::lengthFromCE32(baseCE32); |
| 176 | |
| 177 | if(length != baseLength) { |
| 178 | add(c); |
| 179 | return; |
| 180 | } |
| 181 | for(int32_t i = 0; i < length; ++i) { |
| 182 | if(ce32s[i] != baseCE32s[i]) { |
| 183 | add(c); |
| 184 | break; |
| 185 | } |
| 186 | } |
| 187 | } else if(tag == Collation::EXPANSION_TAG) { |
| 188 | const int64_t *ces = data->ces + Collation::indexFromCE32(ce32); |
| 189 | int32_t length = Collation::lengthFromCE32(ce32); |
| 190 | |
| 191 | const int64_t *baseCEs = baseData->ces + Collation::indexFromCE32(baseCE32); |
| 192 | int32_t baseLength = Collation::lengthFromCE32(baseCE32); |
| 193 | |
| 194 | if(length != baseLength) { |
| 195 | add(c); |
| 196 | return; |
| 197 | } |
| 198 | for(int32_t i = 0; i < length; ++i) { |
| 199 | if(ces[i] != baseCEs[i]) { |
| 200 | add(c); |
| 201 | break; |
| 202 | } |
| 203 | } |
| 204 | } else if(tag == Collation::HANGUL_TAG) { |
| 205 | UChar jamos[3]; |
| 206 | int32_t length = Hangul::decompose(c, jamos); |
| 207 | if(tailored->contains(jamos[0]) || tailored->contains(jamos[1]) || |
| 208 | (length == 3 && tailored->contains(jamos[2]))) { |
| 209 | add(c); |
| 210 | } |
| 211 | } else if(ce32 != baseCE32) { |
| 212 | add(c); |
| 213 | } |
| 214 | } |
| 215 | |
| 216 | void |
| 217 | TailoredSet::comparePrefixes(UChar32 c, const UChar *p, const UChar *q) { |
| 218 | // Parallel iteration over prefixes of both tables. |
| 219 | UCharsTrie::Iterator prefixes(p, 0, errorCode); |
| 220 | UCharsTrie::Iterator basePrefixes(q, 0, errorCode); |
| 221 | const UnicodeString *tp = NULL; // Tailoring prefix. |
| 222 | const UnicodeString *bp = NULL; // Base prefix. |
| 223 | // Use a string with a U+FFFF as the limit sentinel. |
| 224 | // U+FFFF is untailorable and will not occur in prefixes. |
| 225 | UnicodeString none((UChar)0xffff); |
| 226 | for(;;) { |
| 227 | if(tp == NULL) { |
| 228 | if(prefixes.next(errorCode)) { |
| 229 | tp = &prefixes.getString(); |
| 230 | } else { |
| 231 | tp = &none; |
| 232 | } |
| 233 | } |
| 234 | if(bp == NULL) { |
| 235 | if(basePrefixes.next(errorCode)) { |
| 236 | bp = &basePrefixes.getString(); |
| 237 | } else { |
| 238 | bp = &none; |
| 239 | } |
| 240 | } |
| 241 | if(tp == &none && bp == &none) { break; } |
| 242 | int32_t cmp = tp->compare(*bp); |
| 243 | if(cmp < 0) { |
| 244 | // tp occurs in the tailoring but not in the base. |
| 245 | addPrefix(data, *tp, c, (uint32_t)prefixes.getValue()); |
| 246 | tp = NULL; |
| 247 | } else if(cmp > 0) { |
| 248 | // bp occurs in the base but not in the tailoring. |
| 249 | addPrefix(baseData, *bp, c, (uint32_t)basePrefixes.getValue()); |
| 250 | bp = NULL; |
| 251 | } else { |
| 252 | setPrefix(*tp); |
| 253 | compare(c, (uint32_t)prefixes.getValue(), (uint32_t)basePrefixes.getValue()); |
| 254 | resetPrefix(); |
| 255 | tp = NULL; |
| 256 | bp = NULL; |
| 257 | } |
| 258 | } |
| 259 | } |
| 260 | |
| 261 | void |
| 262 | TailoredSet::compareContractions(UChar32 c, const UChar *p, const UChar *q) { |
| 263 | // Parallel iteration over suffixes of both tables. |
| 264 | UCharsTrie::Iterator suffixes(p, 0, errorCode); |
| 265 | UCharsTrie::Iterator baseSuffixes(q, 0, errorCode); |
| 266 | const UnicodeString *ts = NULL; // Tailoring suffix. |
| 267 | const UnicodeString *bs = NULL; // Base suffix. |
| 268 | // Use a string with two U+FFFF as the limit sentinel. |
| 269 | // U+FFFF is untailorable and will not occur in contractions except maybe |
| 270 | // as a single suffix character for a root-collator boundary contraction. |
| 271 | UnicodeString none((UChar)0xffff); |
| 272 | none.append((UChar)0xffff); |
| 273 | for(;;) { |
| 274 | if(ts == NULL) { |
| 275 | if(suffixes.next(errorCode)) { |
| 276 | ts = &suffixes.getString(); |
| 277 | } else { |
| 278 | ts = &none; |
| 279 | } |
| 280 | } |
| 281 | if(bs == NULL) { |
| 282 | if(baseSuffixes.next(errorCode)) { |
| 283 | bs = &baseSuffixes.getString(); |
| 284 | } else { |
| 285 | bs = &none; |
| 286 | } |
| 287 | } |
| 288 | if(ts == &none && bs == &none) { break; } |
| 289 | int32_t cmp = ts->compare(*bs); |
| 290 | if(cmp < 0) { |
| 291 | // ts occurs in the tailoring but not in the base. |
| 292 | addSuffix(c, *ts); |
| 293 | ts = NULL; |
| 294 | } else if(cmp > 0) { |
| 295 | // bs occurs in the base but not in the tailoring. |
| 296 | addSuffix(c, *bs); |
| 297 | bs = NULL; |
| 298 | } else { |
| 299 | suffix = ts; |
| 300 | compare(c, (uint32_t)suffixes.getValue(), (uint32_t)baseSuffixes.getValue()); |
| 301 | suffix = NULL; |
| 302 | ts = NULL; |
| 303 | bs = NULL; |
| 304 | } |
| 305 | } |
| 306 | } |
| 307 | |
| 308 | void |
| 309 | TailoredSet::addPrefixes(const CollationData *d, UChar32 c, const UChar *p) { |
| 310 | UCharsTrie::Iterator prefixes(p, 0, errorCode); |
| 311 | while(prefixes.next(errorCode)) { |
| 312 | addPrefix(d, prefixes.getString(), c, (uint32_t)prefixes.getValue()); |
| 313 | } |
| 314 | } |
| 315 | |
| 316 | void |
| 317 | TailoredSet::addPrefix(const CollationData *d, const UnicodeString &pfx, UChar32 c, uint32_t ce32) { |
| 318 | setPrefix(pfx); |
| 319 | ce32 = d->getFinalCE32(ce32); |
| 320 | if(Collation::isContractionCE32(ce32)) { |
| 321 | const UChar *p = d->contexts + Collation::indexFromCE32(ce32); |
| 322 | addContractions(c, p + 2); |
| 323 | } |
| 324 | tailored->add(UnicodeString(unreversedPrefix).append(c)); |
| 325 | resetPrefix(); |
| 326 | } |
| 327 | |
| 328 | void |
| 329 | TailoredSet::addContractions(UChar32 c, const UChar *p) { |
| 330 | UCharsTrie::Iterator suffixes(p, 0, errorCode); |
| 331 | while(suffixes.next(errorCode)) { |
| 332 | addSuffix(c, suffixes.getString()); |
| 333 | } |
| 334 | } |
| 335 | |
| 336 | void |
| 337 | TailoredSet::addSuffix(UChar32 c, const UnicodeString &sfx) { |
| 338 | tailored->add(UnicodeString(unreversedPrefix).append(c).append(sfx)); |
| 339 | } |
| 340 | |
| 341 | void |
| 342 | TailoredSet::add(UChar32 c) { |
| 343 | if(unreversedPrefix.isEmpty() && suffix == NULL) { |
| 344 | tailored->add(c); |
| 345 | } else { |
| 346 | UnicodeString s(unreversedPrefix); |
| 347 | s.append(c); |
| 348 | if(suffix != NULL) { |
| 349 | s.append(*suffix); |
| 350 | } |
| 351 | tailored->add(s); |
| 352 | } |
| 353 | } |
| 354 | |
| 355 | ContractionsAndExpansions::CESink::~CESink() {} |
| 356 | |
| 357 | U_CDECL_BEGIN |
| 358 | |
| 359 | static UBool U_CALLCONV |
| 360 | enumCnERange(const void *context, UChar32 start, UChar32 end, uint32_t ce32) { |
| 361 | ContractionsAndExpansions *cne = (ContractionsAndExpansions *)context; |
| 362 | if(cne->checkTailored == 0) { |
| 363 | // There is no tailoring. |
| 364 | // No need to collect nor check the tailored set. |
| 365 | } else if(cne->checkTailored < 0) { |
| 366 | // Collect the set of code points with mappings in the tailoring data. |
| 367 | if(ce32 == Collation::FALLBACK_CE32) { |
| 368 | return TRUE; // fallback to base, not tailored |
| 369 | } else { |
| 370 | cne->tailored.add(start, end); |
| 371 | } |
| 372 | // checkTailored > 0: Exclude tailored ranges from the base data enumeration. |
| 373 | } else if(start == end) { |
| 374 | if(cne->tailored.contains(start)) { |
| 375 | return TRUE; |
| 376 | } |
| 377 | } else if(cne->tailored.containsSome(start, end)) { |
| 378 | cne->ranges.set(start, end).removeAll(cne->tailored); |
| 379 | int32_t count = cne->ranges.getRangeCount(); |
| 380 | for(int32_t i = 0; i < count; ++i) { |
| 381 | cne->handleCE32(cne->ranges.getRangeStart(i), cne->ranges.getRangeEnd(i), ce32); |
| 382 | } |
| 383 | return U_SUCCESS(cne->errorCode); |
| 384 | } |
| 385 | cne->handleCE32(start, end, ce32); |
| 386 | return U_SUCCESS(cne->errorCode); |
| 387 | } |
| 388 | |
| 389 | U_CDECL_END |
| 390 | |
| 391 | void |
| 392 | ContractionsAndExpansions::forData(const CollationData *d, UErrorCode &ec) { |
| 393 | if(U_FAILURE(ec)) { return; } |
| 394 | errorCode = ec; // Preserve info & warning codes. |
| 395 | // Add all from the data, can be tailoring or base. |
| 396 | if(d->base != NULL) { |
| 397 | checkTailored = -1; |
| 398 | } |
| 399 | data = d; |
| 400 | utrie2_enum(data->trie, NULL, enumCnERange, this); |
| 401 | if(d->base == NULL || U_FAILURE(errorCode)) { |
| 402 | ec = errorCode; |
| 403 | return; |
| 404 | } |
| 405 | // Add all from the base data but only for un-tailored code points. |
| 406 | tailored.freeze(); |
| 407 | checkTailored = 1; |
| 408 | data = d->base; |
| 409 | utrie2_enum(data->trie, NULL, enumCnERange, this); |
| 410 | ec = errorCode; |
| 411 | } |
| 412 | |
| 413 | void |
| 414 | ContractionsAndExpansions::forCodePoint(const CollationData *d, UChar32 c, UErrorCode &ec) { |
| 415 | if(U_FAILURE(ec)) { return; } |
| 416 | errorCode = ec; // Preserve info & warning codes. |
| 417 | uint32_t ce32 = d->getCE32(c); |
| 418 | if(ce32 == Collation::FALLBACK_CE32) { |
| 419 | d = d->base; |
| 420 | ce32 = d->getCE32(c); |
| 421 | } |
| 422 | data = d; |
| 423 | handleCE32(c, c, ce32); |
| 424 | ec = errorCode; |
| 425 | } |
| 426 | |
| 427 | void |
| 428 | ContractionsAndExpansions::handleCE32(UChar32 start, UChar32 end, uint32_t ce32) { |
| 429 | for(;;) { |
| 430 | if((ce32 & 0xff) < Collation::SPECIAL_CE32_LOW_BYTE) { |
| 431 | // !isSpecialCE32() |
| 432 | if(sink != NULL) { |
| 433 | sink->handleCE(Collation::ceFromSimpleCE32(ce32)); |
| 434 | } |
| 435 | return; |
| 436 | } |
| 437 | switch(Collation::tagFromCE32(ce32)) { |
| 438 | case Collation::FALLBACK_TAG: |
| 439 | return; |
| 440 | case Collation::RESERVED_TAG_3: |
| 441 | case Collation::BUILDER_DATA_TAG: |
| 442 | case Collation::LEAD_SURROGATE_TAG: |
| 443 | if(U_SUCCESS(errorCode)) { errorCode = U_INTERNAL_PROGRAM_ERROR; } |
| 444 | return; |
| 445 | case Collation::LONG_PRIMARY_TAG: |
| 446 | if(sink != NULL) { |
| 447 | sink->handleCE(Collation::ceFromLongPrimaryCE32(ce32)); |
| 448 | } |
| 449 | return; |
| 450 | case Collation::LONG_SECONDARY_TAG: |
| 451 | if(sink != NULL) { |
| 452 | sink->handleCE(Collation::ceFromLongSecondaryCE32(ce32)); |
| 453 | } |
| 454 | return; |
| 455 | case Collation::LATIN_EXPANSION_TAG: |
| 456 | if(sink != NULL) { |
| 457 | ces[0] = Collation::latinCE0FromCE32(ce32); |
| 458 | ces[1] = Collation::latinCE1FromCE32(ce32); |
| 459 | sink->handleExpansion(ces, 2); |
| 460 | } |
| 461 | // Optimization: If we have a prefix, |
| 462 | // then the relevant strings have been added already. |
| 463 | if(unreversedPrefix.isEmpty()) { |
| 464 | addExpansions(start, end); |
| 465 | } |
| 466 | return; |
| 467 | case Collation::EXPANSION32_TAG: |
| 468 | if(sink != NULL) { |
| 469 | const uint32_t *ce32s = data->ce32s + Collation::indexFromCE32(ce32); |
| 470 | int32_t length = Collation::lengthFromCE32(ce32); |
| 471 | for(int32_t i = 0; i < length; ++i) { |
| 472 | ces[i] = Collation::ceFromCE32(*ce32s++); |
| 473 | } |
| 474 | sink->handleExpansion(ces, length); |
| 475 | } |
| 476 | // Optimization: If we have a prefix, |
| 477 | // then the relevant strings have been added already. |
| 478 | if(unreversedPrefix.isEmpty()) { |
| 479 | addExpansions(start, end); |
| 480 | } |
| 481 | return; |
| 482 | case Collation::EXPANSION_TAG: |
| 483 | if(sink != NULL) { |
| 484 | int32_t length = Collation::lengthFromCE32(ce32); |
| 485 | sink->handleExpansion(data->ces + Collation::indexFromCE32(ce32), length); |
| 486 | } |
| 487 | // Optimization: If we have a prefix, |
| 488 | // then the relevant strings have been added already. |
| 489 | if(unreversedPrefix.isEmpty()) { |
| 490 | addExpansions(start, end); |
| 491 | } |
| 492 | return; |
| 493 | case Collation::PREFIX_TAG: |
| 494 | handlePrefixes(start, end, ce32); |
| 495 | return; |
| 496 | case Collation::CONTRACTION_TAG: |
| 497 | handleContractions(start, end, ce32); |
| 498 | return; |
| 499 | case Collation::DIGIT_TAG: |
| 500 | // Fetch the non-numeric-collation CE32 and continue. |
| 501 | ce32 = data->ce32s[Collation::indexFromCE32(ce32)]; |
| 502 | break; |
| 503 | case Collation::U0000_TAG: |
| 504 | U_ASSERT(start == 0 && end == 0); |
| 505 | // Fetch the normal ce32 for U+0000 and continue. |
| 506 | ce32 = data->ce32s[0]; |
| 507 | break; |
| 508 | case Collation::HANGUL_TAG: |
| 509 | if(sink != NULL) { |
| 510 | // TODO: This should be optimized, |
| 511 | // especially if [start..end] is the complete Hangul range. (assert that) |
| 512 | UTF16CollationIterator iter(data, FALSE, NULL, NULL, NULL); |
| 513 | UChar hangul[1] = { 0 }; |
| 514 | for(UChar32 c = start; c <= end; ++c) { |
| 515 | hangul[0] = (UChar)c; |
| 516 | iter.setText(hangul, hangul + 1); |
| 517 | int32_t length = iter.fetchCEs(errorCode); |
| 518 | if(U_FAILURE(errorCode)) { return; } |
| 519 | // Ignore the terminating non-CE. |
| 520 | U_ASSERT(length >= 2 && iter.getCE(length - 1) == Collation::NO_CE); |
| 521 | sink->handleExpansion(iter.getCEs(), length - 1); |
| 522 | } |
| 523 | } |
| 524 | // Optimization: If we have a prefix, |
| 525 | // then the relevant strings have been added already. |
| 526 | if(unreversedPrefix.isEmpty()) { |
| 527 | addExpansions(start, end); |
| 528 | } |
| 529 | return; |
| 530 | case Collation::OFFSET_TAG: |
| 531 | // Currently no need to send offset CEs to the sink. |
| 532 | return; |
| 533 | case Collation::IMPLICIT_TAG: |
| 534 | // Currently no need to send implicit CEs to the sink. |
| 535 | return; |
| 536 | } |
| 537 | } |
| 538 | } |
| 539 | |
| 540 | void |
| 541 | ContractionsAndExpansions::handlePrefixes( |
| 542 | UChar32 start, UChar32 end, uint32_t ce32) { |
| 543 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); |
| 544 | ce32 = CollationData::readCE32(p); // Default if no prefix match. |
| 545 | handleCE32(start, end, ce32); |
| 546 | if(!addPrefixes) { return; } |
| 547 | UCharsTrie::Iterator prefixes(p + 2, 0, errorCode); |
| 548 | while(prefixes.next(errorCode)) { |
| 549 | setPrefix(prefixes.getString()); |
| 550 | // Prefix/pre-context mappings are special kinds of contractions |
| 551 | // that always yield expansions. |
| 552 | addStrings(start, end, contractions); |
| 553 | addStrings(start, end, expansions); |
| 554 | handleCE32(start, end, (uint32_t)prefixes.getValue()); |
| 555 | } |
| 556 | resetPrefix(); |
| 557 | } |
| 558 | |
| 559 | void |
| 560 | ContractionsAndExpansions::handleContractions( |
| 561 | UChar32 start, UChar32 end, uint32_t ce32) { |
| 562 | const UChar *p = data->contexts + Collation::indexFromCE32(ce32); |
| 563 | if((ce32 & Collation::CONTRACT_SINGLE_CP_NO_MATCH) != 0) { |
| 564 | // No match on the single code point. |
| 565 | // We are underneath a prefix, and the default mapping is just |
| 566 | // a fallback to the mappings for a shorter prefix. |
| 567 | U_ASSERT(!unreversedPrefix.isEmpty()); |
| 568 | } else { |
| 569 | ce32 = CollationData::readCE32(p); // Default if no suffix match. |
| 570 | U_ASSERT(!Collation::isContractionCE32(ce32)); |
| 571 | handleCE32(start, end, ce32); |
| 572 | } |
| 573 | UCharsTrie::Iterator suffixes(p + 2, 0, errorCode); |
| 574 | while(suffixes.next(errorCode)) { |
| 575 | suffix = &suffixes.getString(); |
| 576 | addStrings(start, end, contractions); |
| 577 | if(!unreversedPrefix.isEmpty()) { |
| 578 | addStrings(start, end, expansions); |
| 579 | } |
| 580 | handleCE32(start, end, (uint32_t)suffixes.getValue()); |
| 581 | } |
| 582 | suffix = NULL; |
| 583 | } |
| 584 | |
| 585 | void |
| 586 | ContractionsAndExpansions::addExpansions(UChar32 start, UChar32 end) { |
| 587 | if(unreversedPrefix.isEmpty() && suffix == NULL) { |
| 588 | if(expansions != NULL) { |
| 589 | expansions->add(start, end); |
| 590 | } |
| 591 | } else { |
| 592 | addStrings(start, end, expansions); |
| 593 | } |
| 594 | } |
| 595 | |
| 596 | void |
| 597 | ContractionsAndExpansions::addStrings(UChar32 start, UChar32 end, UnicodeSet *set) { |
| 598 | if(set == NULL) { return; } |
| 599 | UnicodeString s(unreversedPrefix); |
| 600 | do { |
| 601 | s.append(start); |
| 602 | if(suffix != NULL) { |
| 603 | s.append(*suffix); |
| 604 | } |
| 605 | set->add(s); |
| 606 | s.truncate(unreversedPrefix.length()); |
| 607 | } while(++start <= end); |
| 608 | } |
| 609 | |
| 610 | U_NAMESPACE_END |
| 611 | |
| 612 | #endif // !UCONFIG_NO_COLLATION |
| 613 | |