int32_t StringSearch::handleNext(int32_t position, UErrorCode &status)
{
// values passed here are already in the pre-shift position
if (U_SUCCESS(status)) {
if (m_strsrch_->pattern.CELength == 0) {
m_search_->matchedIndex =
m_search_->matchedIndex == USEARCH_DONE ?
getOffset() : m_search_->matchedIndex + 1;
m_search_->matchedLength = 0;
ucol_setOffset(m_strsrch_->textIter, m_search_->matchedIndex,
&status);
if (m_search_->matchedIndex == m_search_->textLength) {
m_search_->matchedIndex = USEARCH_DONE;
}
}
else {
// looking at usearch.cpp, this part is shifted out to
// StringSearch instead of SearchIterator because m_strsrch_ is
// not accessible in SearchIterator
if (position + m_strsrch_->pattern.defaultShiftSize
> m_search_->textLength) {
setMatchNotFound();
return USEARCH_DONE;
}
ucol_setOffset(m_strsrch_->textIter, position, &status);
while (TRUE) {
if (m_search_->isCanonicalMatch) {
// can't use exact here since extra accents are allowed.
usearch_handleNextCanonical(m_strsrch_, &status);
}
else {
usearch_handleNextExact(m_strsrch_, &status);
}
if (U_FAILURE(status)) {
return USEARCH_DONE;
}
if (m_breakiterator_ == NULL
#if !UCONFIG_NO_BREAK_ITERATION
||
m_search_->matchedIndex == USEARCH_DONE ||
(m_breakiterator_->isBoundary(m_search_->matchedIndex) &&
m_breakiterator_->isBoundary(m_search_->matchedIndex +
m_search_->matchedLength))
#endif
) {
if (m_search_->matchedIndex == USEARCH_DONE) {
ucol_setOffset(m_strsrch_->textIter,
m_search_->textLength, &status);
}
else {
ucol_setOffset(m_strsrch_->textIter,
m_search_->matchedIndex, &status);
}
return m_search_->matchedIndex;
}
}
}
}
return USEARCH_DONE;
}
//static void NativeCollation_setOffset(JNIEnv* env, jclass, jint address, jint offset) {
JNIEXPORT void JNICALL
Java_com_ibm_icu4jni_text_NativeCollation_setOffset(JNIEnv* env, jclass,
jint address, jint offset) {
UErrorCode status = U_ZERO_ERROR;
ucol_setOffset(toCollationElements(address), offset, &status);
icu4jni_error(env, status);
}
void Target::setOffset(int32_t offset)
{
UErrorCode status = U_ZERO_ERROR;
bufferMin = 0;
bufferMax = 0;
ucol_setOffset(elements, offset, &status);
}
void Target::setLast(int32_t last)
{
UErrorCode status = U_ZERO_ERROR;
bufferMin = 0;
bufferMax = 1;
ceb[0].order = UCOL_NULLORDER;
ceb[0].lowOffset = last;
ceb[0].highOffset = last;
ucol_setOffset(elements, last, &status);
}
OrderList::OrderList(UCollator *coll, const UnicodeString &string, int32_t stringOffset)
: list(NULL), listMax(16), listSize(0)
{
UErrorCode status = U_ZERO_ERROR;
UCollationElements *elems = ucol_openElements(coll, string.getBuffer(), string.length(), &status);
uint32_t strengthMask = 0;
int32_t order, low, high;
switch (ucol_getStrength(coll))
{
default:
strengthMask |= UCOL_TERTIARYORDERMASK;
/* fall through */
case UCOL_SECONDARY:
strengthMask |= UCOL_SECONDARYORDERMASK;
/* fall through */
case UCOL_PRIMARY:
strengthMask |= UCOL_PRIMARYORDERMASK;
}
list = new Order[listMax];
ucol_setOffset(elems, stringOffset, &status);
do {
low = ucol_getOffset(elems);
order = ucol_next(elems, &status);
high = ucol_getOffset(elems);
if (order != UCOL_NULLORDER) {
order &= strengthMask;
}
if (order != UCOL_IGNORABLE) {
add(order, low, high);
}
} while (order != UCOL_NULLORDER);
ucol_closeElements(elems);
}
void CollationElementIterator::setOffset(int32_t newOffset,
UErrorCode& status)
{
ucol_setOffset(m_data_, newOffset, &status);
}
static void NativeCollation_setOffset(JNIEnv* env, jclass, jlong address, jint offset) {
UErrorCode status = U_ZERO_ERROR;
ucol_setOffset(toCollationElements(address), offset, &status);
maybeThrowIcuException(env, "ucol_setOffset", status);
}
int32_t StringSearch::handlePrev(int32_t position, UErrorCode &status)
{
// values passed here are already in the pre-shift position
if (U_SUCCESS(status)) {
if (m_strsrch_->pattern.cesLength == 0) {
m_search_->matchedIndex =
(m_search_->matchedIndex == USEARCH_DONE ? getOffset() :
m_search_->matchedIndex);
if (m_search_->matchedIndex == 0) {
setMatchNotFound();
}
else {
m_search_->matchedIndex --;
ucol_setOffset(m_strsrch_->textIter, m_search_->matchedIndex,
&status);
m_search_->matchedLength = 0;
}
}
else {
// looking at usearch.cpp, this part is shifted out to
// StringSearch instead of SearchIterator because m_strsrch_ is
// not accessible in SearchIterator
#if 0
if (!m_search_->isOverlap &&
position - m_strsrch_->pattern.defaultShiftSize < 0) {
setMatchNotFound();
return USEARCH_DONE;
}
for (;;) {
if (m_search_->isCanonicalMatch) {
// can't use exact here since extra accents are allowed.
usearch_handlePreviousCanonical(m_strsrch_, &status);
}
else {
usearch_handlePreviousExact(m_strsrch_, &status);
}
if (U_FAILURE(status)) {
return USEARCH_DONE;
}
if (m_breakiterator_ == NULL
#if !UCONFIG_NO_BREAK_ITERATION
||
m_search_->matchedIndex == USEARCH_DONE ||
(m_breakiterator_->isBoundary(m_search_->matchedIndex) &&
m_breakiterator_->isBoundary(m_search_->matchedIndex +
m_search_->matchedLength))
#endif
) {
return m_search_->matchedIndex;
}
}
#else
ucol_setOffset(m_strsrch_->textIter, position, &status);
if (m_search_->isCanonicalMatch) {
// *could* use exact match here since extra accents *not* allowed!
usearch_handlePreviousCanonical(m_strsrch_, &status);
} else {
usearch_handlePreviousExact(m_strsrch_, &status);
}
if (U_FAILURE(status)) {
return USEARCH_DONE;
}
return m_search_->matchedIndex;
#endif
}
return m_search_->matchedIndex;
}
return USEARCH_DONE;
}
int32_t StringSearch::handleNext(int32_t position, UErrorCode &status)
{
// values passed here are already in the pre-shift position
if (U_SUCCESS(status)) {
if (m_strsrch_->pattern.cesLength == 0) {
m_search_->matchedIndex =
m_search_->matchedIndex == USEARCH_DONE ?
getOffset() : m_search_->matchedIndex + 1;
m_search_->matchedLength = 0;
ucol_setOffset(m_strsrch_->textIter, m_search_->matchedIndex,
&status);
if (m_search_->matchedIndex == m_search_->textLength) {
m_search_->matchedIndex = USEARCH_DONE;
}
}
else {
// looking at usearch.cpp, this part is shifted out to
// StringSearch instead of SearchIterator because m_strsrch_ is
// not accessible in SearchIterator
#if 0
if (position + m_strsrch_->pattern.defaultShiftSize
> m_search_->textLength) {
setMatchNotFound();
return USEARCH_DONE;
}
#endif
if (m_search_->matchedLength <= 0) {
// the flipping direction issue has already been handled
// in next()
// for boundary check purposes. this will ensure that the
// next match will not preceed the current offset
// note search->matchedIndex will always be set to something
// in the code
m_search_->matchedIndex = position - 1;
}
ucol_setOffset(m_strsrch_->textIter, position, &status);
#if 0
for (;;) {
if (m_search_->isCanonicalMatch) {
// can't use exact here since extra accents are allowed.
usearch_handleNextCanonical(m_strsrch_, &status);
}
else {
usearch_handleNextExact(m_strsrch_, &status);
}
if (U_FAILURE(status)) {
return USEARCH_DONE;
}
if (m_breakiterator_ == NULL
#if !UCONFIG_NO_BREAK_ITERATION
||
m_search_->matchedIndex == USEARCH_DONE ||
(m_breakiterator_->isBoundary(m_search_->matchedIndex) &&
m_breakiterator_->isBoundary(m_search_->matchedIndex +
m_search_->matchedLength))
#endif
) {
if (m_search_->matchedIndex == USEARCH_DONE) {
ucol_setOffset(m_strsrch_->textIter,
m_search_->textLength, &status);
}
else {
ucol_setOffset(m_strsrch_->textIter,
m_search_->matchedIndex, &status);
}
return m_search_->matchedIndex;
}
}
#else
// if m_strsrch_->breakIter is always the same as m_breakiterator_
// then we don't need to check the match boundaries here because
// usearch_handleNextXXX will already have done it.
if (m_search_->isCanonicalMatch) {
// *could* actually use exact here 'cause no extra accents allowed...
usearch_handleNextCanonical(m_strsrch_, &status);
} else {
usearch_handleNextExact(m_strsrch_, &status);
}
if (U_FAILURE(status)) {
return USEARCH_DONE;
}
if (m_search_->matchedIndex == USEARCH_DONE) {
ucol_setOffset(m_strsrch_->textIter, m_search_->textLength, &status);
} else {
ucol_setOffset(m_strsrch_->textIter, m_search_->matchedIndex, &status);
}
return m_search_->matchedIndex;
#endif
}
}
return USEARCH_DONE;
}
static void NativeCollation_setOffset(JNIEnv* env, jclass, jint address, jint offset) {
UErrorCode status = U_ZERO_ERROR;
ucol_setOffset(toCollationElements(address), offset, &status);
icu4jni_error(env, status);
}
static void TestBug672() {
UErrorCode status = U_ZERO_ERROR;
UChar pattern[20];
UChar text[50];
int i;
int result[3][3];
u_uastrcpy(pattern, "resume");
u_uastrcpy(text, "Time to resume updating my resume.");
for (i = 0; i < 3; ++ i) {
UCollator *coll = ucol_open(LOCALES[i], &status);
UCollationElements *pitr = ucol_openElements(coll, pattern, -1,
&status);
UCollationElements *titer = ucol_openElements(coll, text, -1,
&status);
if (U_FAILURE(status)) {
log_err_status(status, "ERROR: in creation of either the collator or the collation iterator :%s\n",
myErrorName(status));
return;
}
log_verbose("locale tested %s\n", LOCALES[i]);
while (ucol_next(pitr, &status) != UCOL_NULLORDER &&
U_SUCCESS(status)) {
}
if (U_FAILURE(status)) {
log_err("ERROR: reversing collation iterator :%s\n",
myErrorName(status));
return;
}
ucol_reset(pitr);
ucol_setOffset(titer, u_strlen(pattern), &status);
if (U_FAILURE(status)) {
log_err("ERROR: setting offset in collator :%s\n",
myErrorName(status));
return;
}
result[i][0] = ucol_getOffset(titer);
log_verbose("Text iterator set to offset %d\n", result[i][0]);
/* Use previous() */
ucol_previous(titer, &status);
result[i][1] = ucol_getOffset(titer);
log_verbose("Current offset %d after previous\n", result[i][1]);
/* Add one to index */
log_verbose("Adding one to current offset...\n");
ucol_setOffset(titer, ucol_getOffset(titer) + 1, &status);
if (U_FAILURE(status)) {
log_err("ERROR: setting offset in collator :%s\n",
myErrorName(status));
return;
}
result[i][2] = ucol_getOffset(titer);
log_verbose("Current offset in text = %d\n", result[i][2]);
ucol_closeElements(pitr);
ucol_closeElements(titer);
ucol_close(coll);
}
if (uprv_memcmp(result[0], result[1], 3) != 0 ||
uprv_memcmp(result[1], result[2], 3) != 0) {
log_err("ERROR: Different locales have different offsets at the same character\n");
}
}
/**
* Test for getOffset() and setOffset()
*/
static void TestOffset()
{
UErrorCode status= U_ZERO_ERROR;
UCollator *en_us=NULL;
UCollationElements *iter, *pristine;
int32_t offset;
OrderAndOffset *orders;
int32_t orderLength=0;
int count = 0;
UChar test1[50];
UChar test2[50];
u_uastrcpy(test1, "What subset of all possible test cases?");
u_uastrcpy(test2, "has the highest probability of detecting");
en_us = ucol_open("en_US", &status);
log_verbose("Testing getOffset and setOffset for collations\n");
iter = ucol_openElements(en_us, test1, u_strlen(test1), &status);
if(U_FAILURE(status)){
log_err_status(status, "ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
myErrorName(status));
ucol_close(en_us);
return;
}
/* testing boundaries */
ucol_setOffset(iter, 0, &status);
if (U_FAILURE(status) || ucol_previous(iter, &status) != UCOL_NULLORDER) {
log_err("Error: After setting offset to 0, we should be at the end "
"of the backwards iteration");
}
ucol_setOffset(iter, u_strlen(test1), &status);
if (U_FAILURE(status) || ucol_next(iter, &status) != UCOL_NULLORDER) {
log_err("Error: After setting offset to end of the string, we should "
"be at the end of the backwards iteration");
}
/* Run all the way through the iterator, then get the offset */
orders = getOrders(iter, &orderLength);
offset = ucol_getOffset(iter);
if (offset != u_strlen(test1))
{
log_err("offset at end != length %d vs %d\n", offset,
u_strlen(test1) );
}
/* Now set the offset back to the beginning and see if it works */
pristine=ucol_openElements(en_us, test1, u_strlen(test1), &status);
if(U_FAILURE(status)){
log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
myErrorName(status));
ucol_close(en_us);
return;
}
status = U_ZERO_ERROR;
ucol_setOffset(iter, 0, &status);
if (U_FAILURE(status))
{
log_err("setOffset failed. %s\n", myErrorName(status));
}
else
{
assertEqual(iter, pristine);
}
ucol_closeElements(pristine);
ucol_closeElements(iter);
free(orders);
/* testing offsets in normalization buffer */
test1[0] = 0x61;
test1[1] = 0x300;
test1[2] = 0x316;
test1[3] = 0x62;
test1[4] = 0;
ucol_setAttribute(en_us, UCOL_NORMALIZATION_MODE, UCOL_ON, &status);
iter = ucol_openElements(en_us, test1, 4, &status);
if(U_FAILURE(status)){
log_err("ERROR: in creation of collation element iterator using ucol_openElements()\n %s\n",
myErrorName(status));
ucol_close(en_us);
return;
}
count = 0;
while (ucol_next(iter, &status) != UCOL_NULLORDER &&
U_SUCCESS(status)) {
switch (count) {
case 0:
if (ucol_getOffset(iter) != 1) {
log_err("ERROR: Offset of iteration should be 1\n");
}
break;
case 3:
if (ucol_getOffset(iter) != 4) {
log_err("ERROR: Offset of iteration should be 4\n");
}
break;
default:
if (ucol_getOffset(iter) != 3) {
log_err("ERROR: Offset of iteration should be 3\n");
}
}
count ++;
}
ucol_reset(iter);
count = 0;
while (ucol_previous(iter, &status) != UCOL_NULLORDER &&
U_SUCCESS(status)) {
switch (count) {
case 0:
case 1:
if (ucol_getOffset(iter) != 3) {
log_err("ERROR: Offset of iteration should be 3\n");
}
break;
case 2:
if (ucol_getOffset(iter) != 1) {
log_err("ERROR: Offset of iteration should be 1\n");
}
break;
default:
if (ucol_getOffset(iter) != 0) {
log_err("ERROR: Offset of iteration should be 0\n");
}
}
count ++;
}
if(U_FAILURE(status)){
log_err("ERROR: in iterating collation elements %s\n",
myErrorName(status));
}
ucol_closeElements(iter);
ucol_close(en_us);
}
static void TestSearchCollatorElements(void)
{
const TSCEItem * tsceItemPtr;
for (tsceItemPtr = tsceItems; tsceItemPtr->locale != NULL; tsceItemPtr++) {
UErrorCode status = U_ZERO_ERROR;
UCollator* ucol = ucol_open(tsceItemPtr->locale, &status);
if ( U_SUCCESS(status) ) {
UCollationElements * uce = ucol_openElements(ucol, tsceText, kLen_tsceText, &status);
if ( U_SUCCESS(status) ) {
int32_t offset, element;
const int32_t * nextOffsetPtr;
const int32_t * limitOffsetPtr;
nextOffsetPtr = tsceItemPtr->offsets;
limitOffsetPtr = tsceItemPtr->offsets + tsceItemPtr->offsetsLen;
do {
offset = ucol_getOffset(uce);
element = ucol_next(uce, &status);
log_verbose("(%s) offset=%2d ce=%08x\n", tsceItemPtr->locale, offset, element);
if ( element == 0 ) {
log_err("error, locale %s, ucol_next returned element 0\n", tsceItemPtr->locale );
}
if ( nextOffsetPtr < limitOffsetPtr ) {
if (offset != *nextOffsetPtr) {
log_err("error, locale %s, expected ucol_next -> ucol_getOffset %d, got %d\n",
tsceItemPtr->locale, *nextOffsetPtr, offset );
nextOffsetPtr = limitOffsetPtr;
break;
}
nextOffsetPtr++;
} else {
log_err("error, locale %s, ucol_next returned more elements than expected\n", tsceItemPtr->locale );
}
} while ( U_SUCCESS(status) && element != UCOL_NULLORDER );
if ( nextOffsetPtr < limitOffsetPtr ) {
log_err("error, locale %s, ucol_next returned fewer elements than expected\n", tsceItemPtr->locale );
}
ucol_setOffset(uce, kLen_tsceText, &status);
status = U_ZERO_ERROR;
nextOffsetPtr = tsceItemPtr->offsets + tsceItemPtr->offsetsLen;
limitOffsetPtr = tsceItemPtr->offsets;
do {
offset = ucol_getOffset(uce);
element = ucol_previous(uce, &status);
if ( element == 0 ) {
log_err("error, locale %s, ucol_previous returned element 0\n", tsceItemPtr->locale );
}
if ( nextOffsetPtr > limitOffsetPtr ) {
nextOffsetPtr--;
if (offset != *nextOffsetPtr) {
log_err("error, locale %s, expected ucol_previous -> ucol_getOffset %d, got %d\n",
tsceItemPtr->locale, *nextOffsetPtr, offset );
nextOffsetPtr = limitOffsetPtr;
break;
}
} else {
log_err("error, locale %s, ucol_previous returned more elements than expected\n", tsceItemPtr->locale );
}
} while ( U_SUCCESS(status) && element != UCOL_NULLORDER );
if ( nextOffsetPtr > limitOffsetPtr ) {
log_err("error, locale %s, ucol_previous returned fewer elements than expected\n", tsceItemPtr->locale );
}
ucol_closeElements(uce);
} else {
log_err("error, locale %s, ucol_openElements failed: %s\n", tsceItemPtr->locale, u_errorName(status) );
}
ucol_close(ucol);
} else {
log_data_err("error, locale %s, ucol_open failed: %s\n", tsceItemPtr->locale, u_errorName(status) );
}
}
}
