const CEI *Target::nextCE(int32_t offset)
{
UErrorCode status = U_ZERO_ERROR;
int32_t low = -1, high = -1;
uint32_t order;
UBool cont = FALSE;
if (offset >= bufferMin && offset < bufferMax) {
return &ceb[offset];
}
if (bufferMax >= bufferSize || offset != bufferMax) {
return NULL;
}
do {
low = ucol_getOffset(elements);
order = ucol_next(elements, &status);
high = ucol_getOffset(elements);
if (order == (uint32_t)UCOL_NULLORDER) {
//high = low = -1;
break;
}
cont = isContinuation(order);
order &= strengthMask;
if (toShift && variableTop > order && (order & UCOL_PRIMARYORDERMASK) != 0) {
if (strength >= UCOL_QUATERNARY) {
order &= UCOL_PRIMARYORDERMASK;
} else {
order = UCOL_IGNORABLE;
}
}
} while (order == UCOL_IGNORABLE);
if (cont) {
order |= UCOL_CONTINUATION_MARKER;
}
ceb[offset].order = order;
ceb[offset].lowOffset = low;
ceb[offset].highOffset = high;
bufferMax += 1;
return &ceb[offset];
}
// Collator.startswith {{{
static PyObject *
icu_Collator_collation_order(icu_Collator *self, PyObject *args, PyObject *kwargs) {
PyObject *a_;
int32_t asz;
int32_t actual_a;
UChar *a;
wchar_t *aw;
UErrorCode status = U_ZERO_ERROR;
UCollationElements *iter = NULL;
int order = 0, len = -1;
if (!PyArg_ParseTuple(args, "U", &a_)) return NULL;
asz = (int32_t)PyUnicode_GetSize(a_);
a = (UChar*)calloc(asz*4 + 2, sizeof(UChar));
aw = (wchar_t*)calloc(asz*4 + 2, sizeof(wchar_t));
if (a == NULL || aw == NULL ) return PyErr_NoMemory();
actual_a = (int32_t)PyUnicode_AsWideChar((PyUnicodeObject*)a_, aw, asz*4+1);
if (actual_a > -1) {
u_strFromWCS(a, asz*4 + 1, &actual_a, aw, -1, &status);
iter = ucol_openElements(self->collator, a, actual_a, &status);
if (iter != NULL && U_SUCCESS(status)) {
order = ucol_next(iter, &status);
len = ucol_getOffset(iter);
ucol_closeElements(iter); iter = NULL;
}
}
free(a); free(aw);
return Py_BuildValue("ii", order, len);
} // }}}
/**
* Return an integer array containing all of the collation orders
* returned by calls to next on the specified iterator
*/
OrderAndOffset* getOrders(UCollationElements *iter, int32_t *orderLength)
{
UErrorCode status;
int32_t order;
int32_t maxSize = 100;
int32_t size = 0;
int32_t offset = ucol_getOffset(iter);
OrderAndOffset *temp;
OrderAndOffset *orders =(OrderAndOffset *)malloc(sizeof(OrderAndOffset) * maxSize);
status= U_ZERO_ERROR;
while ((order=ucol_next(iter, &status)) != UCOL_NULLORDER)
{
if (size == maxSize)
{
maxSize *= 2;
temp = (OrderAndOffset *)malloc(sizeof(OrderAndOffset) * maxSize);
memcpy(temp, orders, size * sizeof(OrderAndOffset));
free(orders);
orders = temp;
}
orders[size].order = order;
orders[size].offset = offset;
offset = ucol_getOffset(iter);
size += 1;
}
if (maxSize > size && size > 0)
{
temp = (OrderAndOffset *)malloc(sizeof(OrderAndOffset) * size);
memcpy(temp, orders, size * sizeof(OrderAndOffset));
free(orders);
orders = temp;
}
*orderLength = size;
return orders;
}
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);
}
// Collator.collation_order {{{
static PyObject *
icu_Collator_collation_order(icu_Collator *self, PyObject *a_) {
int32_t asz = 0;
UChar *a = NULL;
UErrorCode status = U_ZERO_ERROR;
UCollationElements *iter = NULL;
int order = 0, len = -1;
a = python_to_icu(a_, &asz);
if (a == NULL) goto end;
iter = ucol_openElements(self->collator, a, asz, &status);
if (U_FAILURE(status)) { PyErr_SetString(PyExc_ValueError, u_errorName(status)); goto end; }
order = ucol_next(iter, &status);
len = ucol_getOffset(iter);
end:
if (iter != NULL) ucol_closeElements(iter); iter = NULL;
if (a != NULL) free(a);
if (PyErr_Occurred()) return NULL;
return Py_BuildValue("ii", order, len);
} // }}}
//static jint NativeCollation_getOffset(JNIEnv*, jclass, jint address) {
JNIEXPORT jint JNICALL
Java_com_ibm_icu4jni_text_NativeCollation_getOffset(JNIEnv*, jclass,
jint address) {
return ucol_getOffset(toCollationElements(address));
}
int32_t CollationElementIterator::getOffset() const
{
return ucol_getOffset(m_data_);
}
static jint NativeCollation_getOffset(JNIEnv*, jclass, jlong address) {
return ucol_getOffset(toCollationElements(address));
}
void
backAndForth(UCollationElements *iter)
{
/* Run through the iterator forwards and stick it into an array */
int32_t idx, o;
UErrorCode status = U_ZERO_ERROR;
int32_t orderLength = 0;
OrderAndOffset *orders = getOrders(iter, &orderLength);
/* Now go through it backwards and make sure we get the same values */
idx = orderLength;
ucol_reset(iter);
/* synwee : changed */
while ((o = ucol_previous(iter, &status)) != UCOL_NULLORDER) {
#if TEST_OFFSETS
int32_t offset =
#endif
ucol_getOffset(iter);
idx -= 1;
if (o != orders[idx].order) {
if (o == 0)
idx ++;
else {
while (idx > 0 && orders[-- idx].order == 0) {
/* nothing... */
}
if (o != orders[idx].order) {
log_err("Mismatched order at index %d: 0x%8.8X vs. 0x%8.8X\n", idx,
orders[idx].order, o);
goto bail;
}
}
}
#if TEST_OFFSETS
if (offset != orders[idx].offset) {
log_err("Mismatched offset at index %d: %d vs. %d\n", idx,
orders[idx].offset, offset);
goto bail;
}
#endif
}
while (idx != 0 && orders[idx - 1].order == 0) {
idx -= 1;
}
if (idx != 0) {
log_err("Didn't get back to beginning - index is %d\n", idx);
ucol_reset(iter);
log_err("\nnext: ");
if ((o = ucol_next(iter, &status)) != UCOL_NULLORDER) {
log_err("Error at %x\n", o);
}
log_err("\nprev: ");
if ((o = ucol_previous(iter, &status)) != UCOL_NULLORDER) {
log_err("Error at %x\n", o);
}
log_verbose("\n");
}
bail:
free(orders);
}
int32_t Target::getOffset()
{
return ucol_getOffset(elements);
}
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) );
}
}
}
