IdentifierInfo &IdentifierInfo::clear() {
fRequiredScripts->resetAll();
uhash_removeAll(fScriptSetSet);
fNumerics->clear();
fCommonAmongAlternates->resetAll();
return *this;
}
static void TestOtherAPI(void){
UErrorCode status = U_ZERO_ERROR;
UHashtable *hash;
/* Use the correct type when cast to void * */
const UChar one[4] = {0x006F, 0x006E, 0x0065, 0}; /* L"one" */
const UChar one2[4] = {0x006F, 0x006E, 0x0065, 0}; /* Get around compiler optimizations */
const UChar two[4] = {0x0074, 0x0077, 0x006F, 0}; /* L"two" */
const UChar two2[4] = {0x0074, 0x0077, 0x006F, 0}; /* L"two" */
const UChar three[6] = {0x0074, 0x0068, 0x0072, 0x0065, 0x0065, 0}; /* L"three" */
const UChar four[6] = {0x0066, 0x006F, 0x0075, 0x0072, 0}; /* L"four" */
const UChar five[6] = {0x0066, 0x0069, 0x0076, 0x0065, 0}; /* L"five" */
const UChar five2[6] = {0x0066, 0x0069, 0x0076, 0x0065, 0}; /* L"five" */
hash = uhash_open(uhash_hashUChars, uhash_compareUChars, &status);
if (U_FAILURE(status)) {
log_err("FAIL: uhash_open failed with %s and returned 0x%08x\n",
u_errorName(status), hash);
return;
}
if (hash == NULL) {
log_err("FAIL: uhash_open returned NULL\n");
return;
}
log_verbose("Ok: uhash_open returned 0x%08X\n", hash);
uhash_puti(hash, (void*)one, 1, &status);
if(uhash_count(hash) != 1){
log_err("FAIL: uhas_count() failed. Expected: 1, Got: %d\n", uhash_count(hash));
}
uhash_puti(hash, (void*)two, 2, &status);
uhash_puti(hash, (void*)three, 3, &status);
uhash_puti(hash, (void*)four, 4, &status);
uhash_puti(hash, (void*)five, 5, &status);
if(uhash_count(hash) != 5){
log_err("FAIL: uhas_count() failed. Expected: 5, Got: %d\n", uhash_count(hash));
}
if(uhash_geti(hash, (void*)two2) != 2){
log_err("FAIL: uhash_geti failed\n");
}
if(uhash_removei(hash, (void*)five2) != 5){
log_err("FAIL: uhash_remove() failed\n");
}
if(uhash_count(hash) != 4){
log_err("FAIL: uhas_count() failed. Expected: 4, Got: %d\n", uhash_count(hash));
}
uhash_put(hash, (void*)one, NULL, &status);
if(uhash_count(hash) != 3){
log_err("FAIL: uhash_put() with value=NULL didn't remove the key value pair\n");
}
status=U_ILLEGAL_ARGUMENT_ERROR;
uhash_puti(hash, (void*)one, 1, &status);
if(uhash_count(hash) != 3){
log_err("FAIL: uhash_put() with value!=NULL should fail when status != U_ZERO_ERROR \n");
}
status=U_ZERO_ERROR;
uhash_puti(hash, (void*)one, 1, &status);
if(uhash_count(hash) != 4){
log_err("FAIL: uhash_put() with value!=NULL didn't replace the key value pair\n");
}
if(_compareUChars((void*)one, (void*)two) == TRUE ||
_compareUChars((void*)one, (void*)one) != TRUE ||
_compareUChars((void*)one, (void*)one2) != TRUE ||
_compareUChars((void*)one, NULL) == TRUE ) {
log_err("FAIL: compareUChars failed\n");
}
uhash_removeAll(hash);
if(uhash_count(hash) != 0){
log_err("FAIL: uhas_count() failed. Expected: 0, Got: %d\n", uhash_count(hash));
}
uhash_setKeyComparator(hash, uhash_compareLong);
uhash_setKeyHasher(hash, uhash_hashLong);
uhash_iputi(hash, 1001, 1, &status);
uhash_iputi(hash, 1002, 2, &status);
uhash_iputi(hash, 1003, 3, &status);
if(_compareLong(1001, 1002) == TRUE ||
_compareLong(1001, 1001) != TRUE ||
_compareLong(1001, 0) == TRUE ) {
log_err("FAIL: compareLong failed\n");
}
/*set the resize policy to just GROW and SHRINK*/
/*how to test this??*/
uhash_setResizePolicy(hash, U_GROW_AND_SHRINK);
uhash_iputi(hash, 1004, 4, &status);
uhash_iputi(hash, 1005, 5, &status);
uhash_iputi(hash, 1006, 6, &status);
if(uhash_count(hash) != 6){
log_err("FAIL: uhash_count() failed. Expected: 6, Got: %d\n", uhash_count(hash));
}
if(uhash_iremovei(hash, 1004) != 4){
log_err("FAIL: uhash_remove failed\n");
}
if(uhash_iremovei(hash, 1004) != 0){
log_err("FAIL: uhash_remove failed\n");
}
uhash_close(hash);
}
void AlphabeticIndex::buildIndex(UErrorCode &status) {
if (U_FAILURE(status)) {
return;
}
if (!indexBuildRequired_) {
return;
}
// Discard any already-built data.
// This is important when the user builds and uses an index, then subsequently modifies it,
// necessitating a rebuild.
bucketList_->removeAllElements();
labels_->removeAllElements();
uhash_removeAll(alreadyIn_);
noDistinctSorting_->clear();
notAlphabetic_->clear();
// first sort the incoming Labels, with a "best" ordering among items
// that are the same according to the collator
UVector preferenceSorting(status); // Vector of UnicodeStrings; owned by the vector.
preferenceSorting.setDeleter(uprv_deleteUObject);
appendUnicodeSetToUVector(preferenceSorting, *initialLabels_, status);
preferenceSorting.sortWithUComparator(PreferenceComparator, &status, status);
// We now make a set of Labels.
// Some of the input may, however, be redundant.
// That is, we might have c, ch, d, where "ch" sorts just like "c", "h"
// So we make a pass through, filtering out those cases.
// TODO: filtering these out would seem to be at odds with the eventual goal
// of being able to split buckets that contain too many items.
UnicodeSet labelSet;
for (int32_t psIndex=0; psIndex<preferenceSorting.size(); psIndex++) {
UnicodeString item = *static_cast<const UnicodeString *>(preferenceSorting.elementAt(psIndex));
// TODO: Since preferenceSorting was originally populated from the contents of a UnicodeSet,
// is it even possible for duplicates to show up in this check?
if (labelSet.contains(item)) {
UnicodeSetIterator itemAlreadyInIter(labelSet);
while (itemAlreadyInIter.next()) {
const UnicodeString &itemAlreadyIn = itemAlreadyInIter.getString();
if (collatorPrimaryOnly_->compare(item, itemAlreadyIn) == 0) {
UnicodeSet *targets = static_cast<UnicodeSet *>(uhash_get(alreadyIn_, &itemAlreadyIn));
if (targets == NULL) {
// alreadyIn.put(itemAlreadyIn, targets = new LinkedHashSet<String>());
targets = new UnicodeSet();
uhash_put(alreadyIn_, itemAlreadyIn.clone(), targets, &status);
}
targets->add(item);
break;
}
}
} else if (item.moveIndex32(0, 1) < item.length() && // Label contains more than one code point.
collatorPrimaryOnly_->compare(item, separated(item)) == 0) {
noDistinctSorting_->add(item);
} else if (!ALPHABETIC->containsSome(item)) {
notAlphabetic_->add(item);
} else {
labelSet.add(item);
}
}
// If we have no labels, hard-code a fallback default set of [A-Z]
// This case can occur with locales that don't have exemplar character data, including root.
// A no-labels situation will cause other problems; it needs to be avoided.
if (labelSet.isEmpty()) {
labelSet.add((UChar32)0x41, (UChar32)0x5A);
}
// Move the set of Labels from the set into a vector, and sort
// according to the collator.
appendUnicodeSetToUVector(*labels_, labelSet, status);
labels_->sortWithUComparator(sortCollateComparator, collatorPrimaryOnly_, status);
// if the result is still too large, cut down to maxLabelCount_ elements, by removing every nth element
// Implemented by copying the elements to be retained to a new UVector.
const int32_t size = labelSet.size() - 1;
if (size > maxLabelCount_) {
UVector *newLabels = new UVector(status);
newLabels->setDeleter(uprv_deleteUObject);
int32_t count = 0;
int32_t old = -1;
for (int32_t srcIndex=0; srcIndex<labels_->size(); srcIndex++) {
const UnicodeString *str = static_cast<const UnicodeString *>(labels_->elementAt(srcIndex));
++count;
const int32_t bump = count * maxLabelCount_ / size;
if (bump == old) {
// it.remove();
} else {
newLabels->addElement(str->clone(), status);
old = bump;
}
}
delete labels_;
labels_ = newLabels;
}
// We now know the list of labels.
// Create a corresponding list of buckets, one per label.
buildBucketList(status); // Corresponds to Java BucketList constructor.
// Bin the Records into the Buckets.
bucketRecords(status);
indexBuildRequired_ = FALSE;
resetBucketIterator(status);
}
