1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/coleitr.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,488 @@
1.4 +/*
1.5 +*******************************************************************************
1.6 +* Copyright (C) 1996-2011, International Business Machines Corporation and *
1.7 +* others. All Rights Reserved. *
1.8 +*******************************************************************************
1.9 +*/
1.10 +
1.11 +/*
1.12 +* File coleitr.cpp
1.13 +*
1.14 +*
1.15 +*
1.16 +* Created by: Helena Shih
1.17 +*
1.18 +* Modification History:
1.19 +*
1.20 +* Date Name Description
1.21 +*
1.22 +* 6/23/97 helena Adding comments to make code more readable.
1.23 +* 08/03/98 erm Synched with 1.2 version of CollationElementIterator.java
1.24 +* 12/10/99 aliu Ported Thai collation support from Java.
1.25 +* 01/25/01 swquek Modified to a C++ wrapper calling C APIs (ucoliter.h)
1.26 +* 02/19/01 swquek Removed CollationElementsIterator() since it is
1.27 +* private constructor and no calls are made to it
1.28 +*/
1.29 +
1.30 +#include "unicode/utypes.h"
1.31 +
1.32 +#if !UCONFIG_NO_COLLATION
1.33 +
1.34 +#include "unicode/coleitr.h"
1.35 +#include "unicode/ustring.h"
1.36 +#include "ucol_imp.h"
1.37 +#include "uassert.h"
1.38 +#include "cmemory.h"
1.39 +
1.40 +
1.41 +/* Constants --------------------------------------------------------------- */
1.42 +
1.43 +U_NAMESPACE_BEGIN
1.44 +
1.45 +UOBJECT_DEFINE_RTTI_IMPLEMENTATION(CollationElementIterator)
1.46 +
1.47 +/* CollationElementIterator public constructor/destructor ------------------ */
1.48 +
1.49 +CollationElementIterator::CollationElementIterator(
1.50 + const CollationElementIterator& other)
1.51 + : UObject(other), isDataOwned_(TRUE)
1.52 +{
1.53 + UErrorCode status = U_ZERO_ERROR;
1.54 + m_data_ = ucol_openElements(other.m_data_->iteratordata_.coll, NULL, 0,
1.55 + &status);
1.56 +
1.57 + *this = other;
1.58 +}
1.59 +
1.60 +CollationElementIterator::~CollationElementIterator()
1.61 +{
1.62 + if (isDataOwned_) {
1.63 + ucol_closeElements(m_data_);
1.64 + }
1.65 +}
1.66 +
1.67 +/* CollationElementIterator public methods --------------------------------- */
1.68 +
1.69 +int32_t CollationElementIterator::getOffset() const
1.70 +{
1.71 + return ucol_getOffset(m_data_);
1.72 +}
1.73 +
1.74 +/**
1.75 +* Get the ordering priority of the next character in the string.
1.76 +* @return the next character's ordering. Returns NULLORDER if an error has
1.77 +* occured or if the end of string has been reached
1.78 +*/
1.79 +int32_t CollationElementIterator::next(UErrorCode& status)
1.80 +{
1.81 + return ucol_next(m_data_, &status);
1.82 +}
1.83 +
1.84 +UBool CollationElementIterator::operator!=(
1.85 + const CollationElementIterator& other) const
1.86 +{
1.87 + return !(*this == other);
1.88 +}
1.89 +
1.90 +UBool CollationElementIterator::operator==(
1.91 + const CollationElementIterator& that) const
1.92 +{
1.93 + if (this == &that || m_data_ == that.m_data_) {
1.94 + return TRUE;
1.95 + }
1.96 +
1.97 + // option comparison
1.98 + if (m_data_->iteratordata_.coll != that.m_data_->iteratordata_.coll)
1.99 + {
1.100 + return FALSE;
1.101 + }
1.102 +
1.103 + // the constructor and setText always sets a length
1.104 + // and we only compare the string not the contents of the normalization
1.105 + // buffer
1.106 + int thislength = (int)(m_data_->iteratordata_.endp - m_data_->iteratordata_.string);
1.107 + int thatlength = (int)(that.m_data_->iteratordata_.endp - that.m_data_->iteratordata_.string);
1.108 +
1.109 + if (thislength != thatlength) {
1.110 + return FALSE;
1.111 + }
1.112 +
1.113 + if (uprv_memcmp(m_data_->iteratordata_.string,
1.114 + that.m_data_->iteratordata_.string,
1.115 + thislength * U_SIZEOF_UCHAR) != 0) {
1.116 + return FALSE;
1.117 + }
1.118 + if (getOffset() != that.getOffset()) {
1.119 + return FALSE;
1.120 + }
1.121 +
1.122 + // checking normalization buffer
1.123 + if ((m_data_->iteratordata_.flags & UCOL_ITER_HASLEN) == 0) {
1.124 + if ((that.m_data_->iteratordata_.flags & UCOL_ITER_HASLEN) != 0) {
1.125 + return FALSE;
1.126 + }
1.127 + // both are in the normalization buffer
1.128 + if (m_data_->iteratordata_.pos
1.129 + - m_data_->iteratordata_.writableBuffer.getBuffer()
1.130 + != that.m_data_->iteratordata_.pos
1.131 + - that.m_data_->iteratordata_.writableBuffer.getBuffer()) {
1.132 + // not in the same position in the normalization buffer
1.133 + return FALSE;
1.134 + }
1.135 + }
1.136 + else if ((that.m_data_->iteratordata_.flags & UCOL_ITER_HASLEN) == 0) {
1.137 + return FALSE;
1.138 + }
1.139 + // checking ce position
1.140 + return (m_data_->iteratordata_.CEpos - m_data_->iteratordata_.CEs)
1.141 + == (that.m_data_->iteratordata_.CEpos
1.142 + - that.m_data_->iteratordata_.CEs);
1.143 +}
1.144 +
1.145 +/**
1.146 +* Get the ordering priority of the previous collation element in the string.
1.147 +* @param status the error code status.
1.148 +* @return the previous element's ordering. Returns NULLORDER if an error has
1.149 +* occured or if the start of string has been reached.
1.150 +*/
1.151 +int32_t CollationElementIterator::previous(UErrorCode& status)
1.152 +{
1.153 + return ucol_previous(m_data_, &status);
1.154 +}
1.155 +
1.156 +/**
1.157 +* Resets the cursor to the beginning of the string.
1.158 +*/
1.159 +void CollationElementIterator::reset()
1.160 +{
1.161 + ucol_reset(m_data_);
1.162 +}
1.163 +
1.164 +void CollationElementIterator::setOffset(int32_t newOffset,
1.165 + UErrorCode& status)
1.166 +{
1.167 + ucol_setOffset(m_data_, newOffset, &status);
1.168 +}
1.169 +
1.170 +/**
1.171 +* Sets the source to the new source string.
1.172 +*/
1.173 +void CollationElementIterator::setText(const UnicodeString& source,
1.174 + UErrorCode& status)
1.175 +{
1.176 + if (U_FAILURE(status)) {
1.177 + return;
1.178 + }
1.179 +
1.180 + int32_t length = source.length();
1.181 + UChar *string = NULL;
1.182 + if (m_data_->isWritable && m_data_->iteratordata_.string != NULL) {
1.183 + uprv_free((UChar *)m_data_->iteratordata_.string);
1.184 + }
1.185 + m_data_->isWritable = TRUE;
1.186 + if (length > 0) {
1.187 + string = (UChar *)uprv_malloc(U_SIZEOF_UCHAR * length);
1.188 + /* test for NULL */
1.189 + if (string == NULL) {
1.190 + status = U_MEMORY_ALLOCATION_ERROR;
1.191 + return;
1.192 + }
1.193 + u_memcpy(string, source.getBuffer(), length);
1.194 + }
1.195 + else {
1.196 + string = (UChar *)uprv_malloc(U_SIZEOF_UCHAR);
1.197 + /* test for NULL */
1.198 + if (string == NULL) {
1.199 + status = U_MEMORY_ALLOCATION_ERROR;
1.200 + return;
1.201 + }
1.202 + *string = 0;
1.203 + }
1.204 + /* Free offsetBuffer before initializing it. */
1.205 + ucol_freeOffsetBuffer(&(m_data_->iteratordata_));
1.206 + uprv_init_collIterate(m_data_->iteratordata_.coll, string, length,
1.207 + &m_data_->iteratordata_, &status);
1.208 +
1.209 + m_data_->reset_ = TRUE;
1.210 +}
1.211 +
1.212 +// Sets the source to the new character iterator.
1.213 +void CollationElementIterator::setText(CharacterIterator& source,
1.214 + UErrorCode& status)
1.215 +{
1.216 + if (U_FAILURE(status))
1.217 + return;
1.218 +
1.219 + int32_t length = source.getLength();
1.220 + UChar *buffer = NULL;
1.221 +
1.222 + if (length == 0) {
1.223 + buffer = (UChar *)uprv_malloc(U_SIZEOF_UCHAR);
1.224 + /* test for NULL */
1.225 + if (buffer == NULL) {
1.226 + status = U_MEMORY_ALLOCATION_ERROR;
1.227 + return;
1.228 + }
1.229 + *buffer = 0;
1.230 + }
1.231 + else {
1.232 + buffer = (UChar *)uprv_malloc(U_SIZEOF_UCHAR * length);
1.233 + /* test for NULL */
1.234 + if (buffer == NULL) {
1.235 + status = U_MEMORY_ALLOCATION_ERROR;
1.236 + return;
1.237 + }
1.238 + /*
1.239 + Using this constructor will prevent buffer from being removed when
1.240 + string gets removed
1.241 + */
1.242 + UnicodeString string;
1.243 + source.getText(string);
1.244 + u_memcpy(buffer, string.getBuffer(), length);
1.245 + }
1.246 +
1.247 + if (m_data_->isWritable && m_data_->iteratordata_.string != NULL) {
1.248 + uprv_free((UChar *)m_data_->iteratordata_.string);
1.249 + }
1.250 + m_data_->isWritable = TRUE;
1.251 + /* Free offsetBuffer before initializing it. */
1.252 + ucol_freeOffsetBuffer(&(m_data_->iteratordata_));
1.253 + uprv_init_collIterate(m_data_->iteratordata_.coll, buffer, length,
1.254 + &m_data_->iteratordata_, &status);
1.255 + m_data_->reset_ = TRUE;
1.256 +}
1.257 +
1.258 +int32_t CollationElementIterator::strengthOrder(int32_t order) const
1.259 +{
1.260 + UCollationStrength s = ucol_getStrength(m_data_->iteratordata_.coll);
1.261 + // Mask off the unwanted differences.
1.262 + if (s == UCOL_PRIMARY) {
1.263 + order &= RuleBasedCollator::PRIMARYDIFFERENCEONLY;
1.264 + }
1.265 + else if (s == UCOL_SECONDARY) {
1.266 + order &= RuleBasedCollator::SECONDARYDIFFERENCEONLY;
1.267 + }
1.268 +
1.269 + return order;
1.270 +}
1.271 +
1.272 +/* CollationElementIterator private constructors/destructors --------------- */
1.273 +
1.274 +/**
1.275 +* This is the "real" constructor for this class; it constructs an iterator
1.276 +* over the source text using the specified collator
1.277 +*/
1.278 +CollationElementIterator::CollationElementIterator(
1.279 + const UnicodeString& sourceText,
1.280 + const RuleBasedCollator* order,
1.281 + UErrorCode& status)
1.282 + : isDataOwned_(TRUE)
1.283 +{
1.284 + if (U_FAILURE(status)) {
1.285 + return;
1.286 + }
1.287 +
1.288 + int32_t length = sourceText.length();
1.289 + UChar *string = NULL;
1.290 +
1.291 + if (length > 0) {
1.292 + string = (UChar *)uprv_malloc(U_SIZEOF_UCHAR * length);
1.293 + /* test for NULL */
1.294 + if (string == NULL) {
1.295 + status = U_MEMORY_ALLOCATION_ERROR;
1.296 + return;
1.297 + }
1.298 + /*
1.299 + Using this constructor will prevent buffer from being removed when
1.300 + string gets removed
1.301 + */
1.302 + u_memcpy(string, sourceText.getBuffer(), length);
1.303 + }
1.304 + else {
1.305 + string = (UChar *)uprv_malloc(U_SIZEOF_UCHAR);
1.306 + /* test for NULL */
1.307 + if (string == NULL) {
1.308 + status = U_MEMORY_ALLOCATION_ERROR;
1.309 + return;
1.310 + }
1.311 + *string = 0;
1.312 + }
1.313 + m_data_ = ucol_openElements(order->ucollator, string, length, &status);
1.314 +
1.315 + /* Test for buffer overflows */
1.316 + if (U_FAILURE(status)) {
1.317 + return;
1.318 + }
1.319 + m_data_->isWritable = TRUE;
1.320 +}
1.321 +
1.322 +/**
1.323 +* This is the "real" constructor for this class; it constructs an iterator over
1.324 +* the source text using the specified collator
1.325 +*/
1.326 +CollationElementIterator::CollationElementIterator(
1.327 + const CharacterIterator& sourceText,
1.328 + const RuleBasedCollator* order,
1.329 + UErrorCode& status)
1.330 + : isDataOwned_(TRUE)
1.331 +{
1.332 + if (U_FAILURE(status))
1.333 + return;
1.334 +
1.335 + // **** should I just drop this test? ****
1.336 + /*
1.337 + if ( sourceText.endIndex() != 0 )
1.338 + {
1.339 + // A CollationElementIterator is really a two-layered beast.
1.340 + // Internally it uses a Normalizer to munge the source text into a form
1.341 + // where all "composed" Unicode characters (such as \u00FC) are split into a
1.342 + // normal character and a combining accent character.
1.343 + // Afterward, CollationElementIterator does its own processing to handle
1.344 + // expanding and contracting collation sequences, ignorables, and so on.
1.345 +
1.346 + Normalizer::EMode decomp = order->getStrength() == Collator::IDENTICAL
1.347 + ? Normalizer::NO_OP : order->getDecomposition();
1.348 +
1.349 + text = new Normalizer(sourceText, decomp);
1.350 + if (text == NULL)
1.351 + status = U_MEMORY_ALLOCATION_ERROR;
1.352 + }
1.353 + */
1.354 + int32_t length = sourceText.getLength();
1.355 + UChar *buffer;
1.356 + if (length > 0) {
1.357 + buffer = (UChar *)uprv_malloc(U_SIZEOF_UCHAR * length);
1.358 + /* test for NULL */
1.359 + if (buffer == NULL) {
1.360 + status = U_MEMORY_ALLOCATION_ERROR;
1.361 + return;
1.362 + }
1.363 + /*
1.364 + Using this constructor will prevent buffer from being removed when
1.365 + string gets removed
1.366 + */
1.367 + UnicodeString string(buffer, length, length);
1.368 + ((CharacterIterator &)sourceText).getText(string);
1.369 + const UChar *temp = string.getBuffer();
1.370 + u_memcpy(buffer, temp, length);
1.371 + }
1.372 + else {
1.373 + buffer = (UChar *)uprv_malloc(U_SIZEOF_UCHAR);
1.374 + /* test for NULL */
1.375 + if (buffer == NULL) {
1.376 + status = U_MEMORY_ALLOCATION_ERROR;
1.377 + return;
1.378 + }
1.379 + *buffer = 0;
1.380 + }
1.381 + m_data_ = ucol_openElements(order->ucollator, buffer, length, &status);
1.382 +
1.383 + /* Test for buffer overflows */
1.384 + if (U_FAILURE(status)) {
1.385 + return;
1.386 + }
1.387 + m_data_->isWritable = TRUE;
1.388 +}
1.389 +
1.390 +/* CollationElementIterator protected methods ----------------------------- */
1.391 +
1.392 +const CollationElementIterator& CollationElementIterator::operator=(
1.393 + const CollationElementIterator& other)
1.394 +{
1.395 + if (this != &other)
1.396 + {
1.397 + UCollationElements *ucolelem = this->m_data_;
1.398 + UCollationElements *otherucolelem = other.m_data_;
1.399 + collIterate *coliter = &(ucolelem->iteratordata_);
1.400 + collIterate *othercoliter = &(otherucolelem->iteratordata_);
1.401 + int length = 0;
1.402 +
1.403 + // checking only UCOL_ITER_HASLEN is not enough here as we may be in
1.404 + // the normalization buffer
1.405 + length = (int)(othercoliter->endp - othercoliter->string);
1.406 +
1.407 + ucolelem->reset_ = otherucolelem->reset_;
1.408 + ucolelem->isWritable = TRUE;
1.409 +
1.410 + /* create a duplicate of string */
1.411 + if (length > 0) {
1.412 + coliter->string = (UChar *)uprv_malloc(length * U_SIZEOF_UCHAR);
1.413 + if(coliter->string != NULL) {
1.414 + uprv_memcpy((UChar *)coliter->string, othercoliter->string,
1.415 + length * U_SIZEOF_UCHAR);
1.416 + } else { // Error: couldn't allocate memory. No copying should be done
1.417 + length = 0;
1.418 + }
1.419 + }
1.420 + else {
1.421 + coliter->string = NULL;
1.422 + }
1.423 +
1.424 + /* start and end of string */
1.425 + coliter->endp = coliter->string == NULL ? NULL : coliter->string + length;
1.426 +
1.427 + /* handle writable buffer here */
1.428 +
1.429 + if (othercoliter->flags & UCOL_ITER_INNORMBUF) {
1.430 + coliter->writableBuffer = othercoliter->writableBuffer;
1.431 + coliter->writableBuffer.getTerminatedBuffer();
1.432 + }
1.433 +
1.434 + /* current position */
1.435 + if (othercoliter->pos >= othercoliter->string &&
1.436 + othercoliter->pos <= othercoliter->endp)
1.437 + {
1.438 + U_ASSERT(coliter->string != NULL);
1.439 + coliter->pos = coliter->string +
1.440 + (othercoliter->pos - othercoliter->string);
1.441 + }
1.442 + else {
1.443 + coliter->pos = coliter->writableBuffer.getTerminatedBuffer() +
1.444 + (othercoliter->pos - othercoliter->writableBuffer.getBuffer());
1.445 + }
1.446 +
1.447 + /* CE buffer */
1.448 + int32_t CEsize;
1.449 + if (coliter->extendCEs) {
1.450 + uprv_memcpy(coliter->CEs, othercoliter->CEs, sizeof(uint32_t) * UCOL_EXPAND_CE_BUFFER_SIZE);
1.451 + CEsize = sizeof(othercoliter->extendCEs);
1.452 + if (CEsize > 0) {
1.453 + othercoliter->extendCEs = (uint32_t *)uprv_malloc(CEsize);
1.454 + uprv_memcpy(coliter->extendCEs, othercoliter->extendCEs, CEsize);
1.455 + }
1.456 + coliter->toReturn = coliter->extendCEs +
1.457 + (othercoliter->toReturn - othercoliter->extendCEs);
1.458 + coliter->CEpos = coliter->extendCEs + CEsize;
1.459 + } else {
1.460 + CEsize = (int32_t)(othercoliter->CEpos - othercoliter->CEs);
1.461 + if (CEsize > 0) {
1.462 + uprv_memcpy(coliter->CEs, othercoliter->CEs, CEsize);
1.463 + }
1.464 + coliter->toReturn = coliter->CEs +
1.465 + (othercoliter->toReturn - othercoliter->CEs);
1.466 + coliter->CEpos = coliter->CEs + CEsize;
1.467 + }
1.468 +
1.469 + if (othercoliter->fcdPosition != NULL) {
1.470 + U_ASSERT(coliter->string != NULL);
1.471 + coliter->fcdPosition = coliter->string +
1.472 + (othercoliter->fcdPosition
1.473 + - othercoliter->string);
1.474 + }
1.475 + else {
1.476 + coliter->fcdPosition = NULL;
1.477 + }
1.478 + coliter->flags = othercoliter->flags/*| UCOL_ITER_HASLEN*/;
1.479 + coliter->origFlags = othercoliter->origFlags;
1.480 + coliter->coll = othercoliter->coll;
1.481 + this->isDataOwned_ = TRUE;
1.482 + }
1.483 +
1.484 + return *this;
1.485 +}
1.486 +
1.487 +U_NAMESPACE_END
1.488 +
1.489 +#endif /* #if !UCONFIG_NO_COLLATION */
1.490 +
1.491 +/* eof */
