SimpleFilteredBreakIteratorBuilder::SimpleFilteredBreakIteratorBuilder(const Locale &fromLocale, UErrorCode &status)
: fSet()
{
if(U_SUCCESS(status)) {
LocalUResourceBundlePointer b(ures_open(U_ICUDATA_BRKITR, fromLocale.getBaseName(), &status));
LocalUResourceBundlePointer exceptions(ures_getByKeyWithFallback(b.getAlias(), "exceptions", NULL, &status));
LocalUResourceBundlePointer breaks(ures_getByKeyWithFallback(exceptions.getAlias(), "SentenceBreak", NULL, &status));
if(U_FAILURE(status)) return; // leaves the builder empty, if you try to use it.
LocalUResourceBundlePointer strs;
UErrorCode subStatus = status;
do {
strs.adoptInstead(ures_getNextResource(breaks.getAlias(), strs.orphan(), &subStatus));
if(strs.isValid() && U_SUCCESS(subStatus)) {
UnicodeString str(ures_getUnicodeString(strs.getAlias(), &status));
suppressBreakAfter(str, status); // load the string
}
} while (strs.isValid() && U_SUCCESS(subStatus));
if(U_FAILURE(subStatus)&&subStatus!=U_INDEX_OUTOFBOUNDS_ERROR&&U_SUCCESS(status)) {
status = subStatus;
}
}
}
void
DecimalFormatSymbols::initialize(const Locale& loc, UErrorCode& status, UBool useLastResortData)
{
static const char *gNumberElementKeys[kFormatSymbolCount] = {
"decimal",
"group",
"list",
"percentSign",
NULL, /* Native zero digit is deprecated from CLDR - get it from the numbering system */
NULL, /* Pattern digit character is deprecated from CLDR - use # by default always */
"minusSign",
"plusSign",
NULL, /* currency symbol - We don't really try to load this directly from CLDR until we know the currency */
NULL, /* intl currency symbol - We don't really try to load this directly from CLDR until we know the currency */
"currencyDecimal",
"exponential",
"perMille",
NULL, /* Escape padding character - not in CLDR */
"infinity",
"nan",
NULL, /* Significant digit symbol - not in CLDR */
"currencyGroup",
NULL, /* one digit - get it from the numbering system */
NULL, /* two digit - get it from the numbering system */
NULL, /* three digit - get it from the numbering system */
NULL, /* four digit - get it from the numbering system */
NULL, /* five digit - get it from the numbering system */
NULL, /* six digit - get it from the numbering system */
NULL, /* seven digit - get it from the numbering system */
NULL, /* eight digit - get it from the numbering system */
NULL, /* nine digit - get it from the numbering system */
"superscriptingExponent", /* Multiplication (x) symbol for exponents */
};
static const char *gLatn = "latn";
static const char *gSymbols = "symbols";
const char *nsName;
const UChar *sym = NULL;
int32_t len = 0;
*validLocale = *actualLocale = 0;
currPattern = NULL;
if (U_FAILURE(status))
return;
const char* locStr = loc.getName();
LocalUResourceBundlePointer resource(ures_open(NULL, locStr, &status));
LocalUResourceBundlePointer numberElementsRes(
ures_getByKeyWithFallback(resource.getAlias(), gNumberElements, NULL, &status));
if (U_FAILURE(status)) {
if ( useLastResortData ) {
status = U_USING_DEFAULT_WARNING;
initialize();
}
return;
}
// First initialize all the symbols to the fallbacks for anything we can't find
initialize();
//
// Next get the numbering system for this locale and set zero digit
// and the digit string based on the numbering system for the locale
//
LocalPointer<NumberingSystem> ns(NumberingSystem::createInstance(loc, status));
if (U_SUCCESS(status) && ns->getRadix() == 10 && !ns->isAlgorithmic()) {
nsName = ns->getName();
UnicodeString digitString(ns->getDescription());
int32_t digitIndex = 0;
UChar32 digit = digitString.char32At(0);
fSymbols[kZeroDigitSymbol].setTo(digit);
for (int32_t i = kOneDigitSymbol; i <= kNineDigitSymbol; ++i) {
digitIndex += U16_LENGTH(digit);
digit = digitString.char32At(digitIndex);
fSymbols[i].setTo(digit);
}
} else {
nsName = gLatn;
}
UBool isLatn = !uprv_strcmp(nsName,gLatn);
UErrorCode nlStatus = U_ZERO_ERROR;
LocalUResourceBundlePointer nonLatnSymbols;
if ( !isLatn ) {
nonLatnSymbols.adoptInstead(
ures_getByKeyWithFallback(numberElementsRes.getAlias(), nsName, NULL, &nlStatus));
ures_getByKeyWithFallback(nonLatnSymbols.getAlias(), gSymbols, nonLatnSymbols.getAlias(), &nlStatus);
}
LocalUResourceBundlePointer latnSymbols(
ures_getByKeyWithFallback(numberElementsRes.getAlias(), gLatn, NULL, &status));
ures_getByKeyWithFallback(latnSymbols.getAlias(), gSymbols, latnSymbols.getAlias(), &status);
UBool kMonetaryDecimalSet = FALSE;
UBool kMonetaryGroupingSet = FALSE;
for(int32_t i = 0; i<kFormatSymbolCount; i++) {
if ( gNumberElementKeys[i] != NULL ) {
UErrorCode localStatus = U_ZERO_ERROR;
if ( !isLatn ) {
sym = ures_getStringByKeyWithFallback(nonLatnSymbols.getAlias(),
gNumberElementKeys[i], &len, &localStatus);
// If we can't find the symbol in the numbering system specific resources,
// use the "latn" numbering system as the fallback.
if ( U_FAILURE(localStatus) ) {
localStatus = U_ZERO_ERROR;
sym = ures_getStringByKeyWithFallback(latnSymbols.getAlias(),
gNumberElementKeys[i], &len, &localStatus);
}
} else {
sym = ures_getStringByKeyWithFallback(latnSymbols.getAlias(),
gNumberElementKeys[i], &len, &localStatus);
}
if ( U_SUCCESS(localStatus) ) {
setSymbol((ENumberFormatSymbol)i, UnicodeString(TRUE, sym, len));
if ( i == kMonetarySeparatorSymbol ) {
kMonetaryDecimalSet = TRUE;
} else if ( i == kMonetaryGroupingSeparatorSymbol ) {
kMonetaryGroupingSet = TRUE;
}
}
}
}
// If monetary decimal or grouping were not explicitly set, then set them to be the
// same as their non-monetary counterparts.
if ( !kMonetaryDecimalSet ) {
setSymbol(kMonetarySeparatorSymbol,fSymbols[kDecimalSeparatorSymbol]);
}
if ( !kMonetaryGroupingSet ) {
setSymbol(kMonetaryGroupingSeparatorSymbol,fSymbols[kGroupingSeparatorSymbol]);
}
// Obtain currency data from the currency API. This is strictly
// for backward compatibility; we don't use DecimalFormatSymbols
// for currency data anymore.
UErrorCode internalStatus = U_ZERO_ERROR; // don't propagate failures out
UChar curriso[4];
UnicodeString tempStr;
ucurr_forLocale(locStr, curriso, 4, &internalStatus);
uprv_getStaticCurrencyName(curriso, locStr, tempStr, internalStatus);
if (U_SUCCESS(internalStatus)) {
fSymbols[kIntlCurrencySymbol].setTo(curriso, -1);
fSymbols[kCurrencySymbol] = tempStr;
}
/* else use the default values. */
U_LOCALE_BASED(locBased, *this);
locBased.setLocaleIDs(ures_getLocaleByType(numberElementsRes.getAlias(),
ULOC_VALID_LOCALE, &status),
ures_getLocaleByType(numberElementsRes.getAlias(),
ULOC_ACTUAL_LOCALE, &status));
//load the currency data
UChar ucc[4]={0}; //Currency Codes are always 3 chars long
int32_t uccLen = 4;
const char* locName = loc.getName();
UErrorCode localStatus = U_ZERO_ERROR;
uccLen = ucurr_forLocale(locName, ucc, uccLen, &localStatus);
if(U_SUCCESS(localStatus) && uccLen > 0) {
char cc[4]={0};
u_UCharsToChars(ucc, cc, uccLen);
/* An explicit currency was requested */
LocalUResourceBundlePointer currencyResource(ures_open(U_ICUDATA_CURR, locStr, &localStatus));
LocalUResourceBundlePointer currency(
ures_getByKeyWithFallback(currencyResource.getAlias(), "Currencies", NULL, &localStatus));
ures_getByKeyWithFallback(currency.getAlias(), cc, currency.getAlias(), &localStatus);
if(U_SUCCESS(localStatus) && ures_getSize(currency.getAlias())>2) { // the length is 3 if more data is present
ures_getByIndex(currency.getAlias(), 2, currency.getAlias(), &localStatus);
int32_t currPatternLen = 0;
currPattern =
ures_getStringByIndex(currency.getAlias(), (int32_t)0, &currPatternLen, &localStatus);
UnicodeString decimalSep =
ures_getUnicodeStringByIndex(currency.getAlias(), (int32_t)1, &localStatus);
UnicodeString groupingSep =
ures_getUnicodeStringByIndex(currency.getAlias(), (int32_t)2, &localStatus);
if(U_SUCCESS(localStatus)){
fSymbols[kMonetaryGroupingSeparatorSymbol] = groupingSep;
fSymbols[kMonetarySeparatorSymbol] = decimalSep;
//pattern.setTo(TRUE, currPattern, currPatternLen);
status = localStatus;
}
}
/* else An explicit currency was requested and is unknown or locale data is malformed. */
/* ucurr_* API will get the correct value later on. */
}
// else ignore the error if no currency
// Currency Spacing.
localStatus = U_ZERO_ERROR;
LocalUResourceBundlePointer currencyResource(ures_open(U_ICUDATA_CURR, locStr, &localStatus));
LocalUResourceBundlePointer currencySpcRes(
ures_getByKeyWithFallback(currencyResource.getAlias(),
gCurrencySpacingTag, NULL, &localStatus));
if (localStatus == U_USING_FALLBACK_WARNING || U_SUCCESS(localStatus)) {
const char* keywords[UNUM_CURRENCY_SPACING_COUNT] = {
gCurrencyMatchTag, gCurrencySudMatchTag, gCurrencyInsertBtnTag
};
localStatus = U_ZERO_ERROR;
LocalUResourceBundlePointer dataRes(
ures_getByKeyWithFallback(currencySpcRes.getAlias(),
gBeforeCurrencyTag, NULL, &localStatus));
if (localStatus == U_USING_FALLBACK_WARNING || U_SUCCESS(localStatus)) {
localStatus = U_ZERO_ERROR;
for (int32_t i = 0; i < UNUM_CURRENCY_SPACING_COUNT; i++) {
currencySpcBeforeSym[i] =
ures_getUnicodeStringByKey(dataRes.getAlias(), keywords[i], &localStatus);
}
}
dataRes.adoptInstead(
ures_getByKeyWithFallback(currencySpcRes.getAlias(),
gAfterCurrencyTag, NULL, &localStatus));
if (localStatus == U_USING_FALLBACK_WARNING || U_SUCCESS(localStatus)) {
localStatus = U_ZERO_ERROR;
for (int32_t i = 0; i < UNUM_CURRENCY_SPACING_COUNT; i++) {
currencySpcAfterSym[i] =
ures_getUnicodeStringByKey(dataRes.getAlias(), keywords[i], &localStatus);
}
}
}
}
U_CDECL_END
static void U_CALLCONV
initFromResourceBundle(UErrorCode& sts) {
U_NAMESPACE_USE
ucln_common_registerCleanup(UCLN_COMMON_LOCALE_KEY_TYPE, uloc_key_type_cleanup);
gLocExtKeyMap = uhash_open(uhash_hashIChars, uhash_compareIChars, NULL, &sts);
LocalUResourceBundlePointer keyTypeDataRes(ures_openDirect(NULL, "keyTypeData", &sts));
LocalUResourceBundlePointer keyMapRes(ures_getByKey(keyTypeDataRes.getAlias(), "keyMap", NULL, &sts));
LocalUResourceBundlePointer typeMapRes(ures_getByKey(keyTypeDataRes.getAlias(), "typeMap", NULL, &sts));
if (U_FAILURE(sts)) {
return;
}
UErrorCode tmpSts = U_ZERO_ERROR;
LocalUResourceBundlePointer typeAliasRes(ures_getByKey(keyTypeDataRes.getAlias(), "typeAlias", NULL, &tmpSts));
tmpSts = U_ZERO_ERROR;
LocalUResourceBundlePointer bcpTypeAliasRes(ures_getByKey(keyTypeDataRes.getAlias(), "bcpTypeAlias", NULL, &tmpSts));
// initialize vectors storing dynamically allocated objects
gKeyTypeStringPool = new UVector(uloc_deleteKeyTypeStringPoolEntry, NULL, sts);
if (gKeyTypeStringPool == NULL) {
if (U_SUCCESS(sts)) {
sts = U_MEMORY_ALLOCATION_ERROR;
}
}
if (U_FAILURE(sts)) {
return;
}
gLocExtKeyDataEntries = new UVector(uloc_deleteKeyDataEntry, NULL, sts);
if (gLocExtKeyDataEntries == NULL) {
if (U_SUCCESS(sts)) {
sts = U_MEMORY_ALLOCATION_ERROR;
}
}
if (U_FAILURE(sts)) {
return;
}
gLocExtTypeEntries = new UVector(uloc_deleteTypeEntry, NULL, sts);
if (gLocExtTypeEntries == NULL) {
if (U_SUCCESS(sts)) {
sts = U_MEMORY_ALLOCATION_ERROR;
}
}
if (U_FAILURE(sts)) {
return;
}
// iterate through keyMap resource
LocalUResourceBundlePointer keyMapEntry;
while (ures_hasNext(keyMapRes.getAlias())) {
keyMapEntry.adoptInstead(ures_getNextResource(keyMapRes.getAlias(), keyMapEntry.orphan(), &sts));
if (U_FAILURE(sts)) {
break;
}
const char* legacyKeyId = ures_getKey(keyMapEntry.getAlias());
int32_t bcpKeyIdLen = 0;
const UChar* uBcpKeyId = ures_getString(keyMapEntry.getAlias(), &bcpKeyIdLen, &sts);
if (U_FAILURE(sts)) {
break;
}
// empty value indicates that BCP key is same with the legacy key.
const char* bcpKeyId = legacyKeyId;
if (bcpKeyIdLen > 0) {
char* bcpKeyIdBuf = (char*)uprv_malloc(bcpKeyIdLen + 1);
if (bcpKeyIdBuf == NULL) {
sts = U_MEMORY_ALLOCATION_ERROR;
break;
}
u_UCharsToChars(uBcpKeyId, bcpKeyIdBuf, bcpKeyIdLen);
bcpKeyIdBuf[bcpKeyIdLen] = 0;
gKeyTypeStringPool->addElement(bcpKeyIdBuf, sts);
if (U_FAILURE(sts)) {
break;
}
bcpKeyId = bcpKeyIdBuf;
}
UBool isTZ = uprv_strcmp(legacyKeyId, "timezone") == 0;
UHashtable* typeDataMap = uhash_open(uhash_hashIChars, uhash_compareIChars, NULL, &sts);
if (U_FAILURE(sts)) {
break;
}
uint32_t specialTypes = SPECIALTYPE_NONE;
LocalUResourceBundlePointer typeAliasResByKey;
LocalUResourceBundlePointer bcpTypeAliasResByKey;
if (typeAliasRes.isValid()) {
tmpSts = U_ZERO_ERROR;
typeAliasResByKey.adoptInstead(ures_getByKey(typeAliasRes.getAlias(), legacyKeyId, NULL, &tmpSts));
if (U_FAILURE(tmpSts)) {
typeAliasResByKey.orphan();
}
}
if (bcpTypeAliasRes.isValid()) {
tmpSts = U_ZERO_ERROR;
bcpTypeAliasResByKey.adoptInstead(ures_getByKey(bcpTypeAliasRes.getAlias(), bcpKeyId, NULL, &tmpSts));
if (U_FAILURE(tmpSts)) {
bcpTypeAliasResByKey.orphan();
}
}
// look up type map for the key, and walk through the mapping data
tmpSts = U_ZERO_ERROR;
LocalUResourceBundlePointer typeMapResByKey(ures_getByKey(typeMapRes.getAlias(), legacyKeyId, NULL, &tmpSts));
if (U_FAILURE(tmpSts)) {
// type map for each key must exist
U_ASSERT(FALSE);
} else {
LocalUResourceBundlePointer typeMapEntry;
while (ures_hasNext(typeMapResByKey.getAlias())) {
typeMapEntry.adoptInstead(ures_getNextResource(typeMapResByKey.getAlias(), typeMapEntry.orphan(), &sts));
if (U_FAILURE(sts)) {
break;
}
const char* legacyTypeId = ures_getKey(typeMapEntry.getAlias());
// special types
if (uprv_strcmp(legacyTypeId, "CODEPOINTS") == 0) {
specialTypes |= SPECIALTYPE_CODEPOINTS;
continue;
}
if (uprv_strcmp(legacyTypeId, "REORDER_CODE") == 0) {
specialTypes |= SPECIALTYPE_REORDER_CODE;
continue;
}
if (uprv_strcmp(legacyTypeId, "RG_KEY_VALUE") == 0) {
specialTypes |= SPECIALTYPE_RG_KEY_VALUE;
continue;
}
if (isTZ) {
// a timezone key uses a colon instead of a slash in the resource.
// e.g. America:Los_Angeles
if (uprv_strchr(legacyTypeId, ':') != NULL) {
int32_t legacyTypeIdLen = uprv_strlen(legacyTypeId);
char* legacyTypeIdBuf = (char*)uprv_malloc(legacyTypeIdLen + 1);
if (legacyTypeIdBuf == NULL) {
sts = U_MEMORY_ALLOCATION_ERROR;
break;
}
const char* p = legacyTypeId;
char* q = legacyTypeIdBuf;
while (*p) {
if (*p == ':') {
*q++ = '/';
} else {
*q++ = *p;
}
p++;
}
*q = 0;
gKeyTypeStringPool->addElement(legacyTypeIdBuf, sts);
if (U_FAILURE(sts)) {
break;
}
legacyTypeId = legacyTypeIdBuf;
}
}
int32_t bcpTypeIdLen = 0;
const UChar* uBcpTypeId = ures_getString(typeMapEntry.getAlias(), &bcpTypeIdLen, &sts);
if (U_FAILURE(sts)) {
break;
}
// empty value indicates that BCP type is same with the legacy type.
const char* bcpTypeId = legacyTypeId;
if (bcpTypeIdLen > 0) {
char* bcpTypeIdBuf = (char*)uprv_malloc(bcpTypeIdLen + 1);
if (bcpTypeIdBuf == NULL) {
sts = U_MEMORY_ALLOCATION_ERROR;
break;
}
u_UCharsToChars(uBcpTypeId, bcpTypeIdBuf, bcpTypeIdLen);
bcpTypeIdBuf[bcpTypeIdLen] = 0;
gKeyTypeStringPool->addElement(bcpTypeIdBuf, sts);
if (U_FAILURE(sts)) {
break;
}
bcpTypeId = bcpTypeIdBuf;
}
// Note: legacy type value should never be
// equivalent to bcp type value of a different
// type under the same key. So we use a single
// map for lookup.
LocExtType* t = (LocExtType*)uprv_malloc(sizeof(LocExtType));
if (t == NULL) {
sts = U_MEMORY_ALLOCATION_ERROR;
break;
}
t->bcpId = bcpTypeId;
t->legacyId = legacyTypeId;
gLocExtTypeEntries->addElement((void*)t, sts);
if (U_FAILURE(sts)) {
break;
}
uhash_put(typeDataMap, (void*)legacyTypeId, t, &sts);
if (bcpTypeId != legacyTypeId) {
// different type value
uhash_put(typeDataMap, (void*)bcpTypeId, t, &sts);
}
if (U_FAILURE(sts)) {
break;
}
// also put aliases in the map
if (typeAliasResByKey.isValid()) {
LocalUResourceBundlePointer typeAliasDataEntry;
ures_resetIterator(typeAliasResByKey.getAlias());
while (ures_hasNext(typeAliasResByKey.getAlias()) && U_SUCCESS(sts)) {
int32_t toLen;
typeAliasDataEntry.adoptInstead(ures_getNextResource(typeAliasResByKey.getAlias(), typeAliasDataEntry.orphan(), &sts));
const UChar* to = ures_getString(typeAliasDataEntry.getAlias(), &toLen, &sts);
if (U_FAILURE(sts)) {
break;
}
// check if this is an alias of canoncal legacy type
if (uprv_compareInvWithUChar(NULL, legacyTypeId, -1, to, toLen) == 0) {
const char* from = ures_getKey(typeAliasDataEntry.getAlias());
if (isTZ) {
// replace colon with slash if necessary
if (uprv_strchr(from, ':') != NULL) {
int32_t fromLen = uprv_strlen(from);
char* fromBuf = (char*)uprv_malloc(fromLen + 1);
if (fromBuf == NULL) {
sts = U_MEMORY_ALLOCATION_ERROR;
break;
}
const char* p = from;
char* q = fromBuf;
while (*p) {
if (*p == ':') {
*q++ = '/';
} else {
*q++ = *p;
}
p++;
}
*q = 0;
gKeyTypeStringPool->addElement(fromBuf, sts);
if (U_FAILURE(sts)) {
break;
}
from = fromBuf;
}
}
uhash_put(typeDataMap, (void*)from, t, &sts);
}
}
if (U_FAILURE(sts)) {
break;
}
}
if (bcpTypeAliasResByKey.isValid()) {
LocalUResourceBundlePointer bcpTypeAliasDataEntry;
ures_resetIterator(bcpTypeAliasResByKey.getAlias());
while (ures_hasNext(bcpTypeAliasResByKey.getAlias()) && U_SUCCESS(sts)) {
int32_t toLen;
bcpTypeAliasDataEntry.adoptInstead(ures_getNextResource(bcpTypeAliasResByKey.getAlias(), bcpTypeAliasDataEntry.orphan(), &sts));
const UChar* to = ures_getString(bcpTypeAliasDataEntry.getAlias(), &toLen, &sts);
if (U_FAILURE(sts)) {
break;
}
// check if this is an alias of bcp type
if (uprv_compareInvWithUChar(NULL, bcpTypeId, -1, to, toLen) == 0) {
const char* from = ures_getKey(bcpTypeAliasDataEntry.getAlias());
uhash_put(typeDataMap, (void*)from, t, &sts);
}
}
if (U_FAILURE(sts)) {
break;
}
}
}
}
if (U_FAILURE(sts)) {
break;
}
LocExtKeyData* keyData = (LocExtKeyData*)uprv_malloc(sizeof(LocExtKeyData));
if (keyData == NULL) {
sts = U_MEMORY_ALLOCATION_ERROR;
break;
}
keyData->bcpId = bcpKeyId;
keyData->legacyId = legacyKeyId;
keyData->specialTypes = specialTypes;
keyData->typeMap = typeDataMap;
gLocExtKeyDataEntries->addElement((void*)keyData, sts);
if (U_FAILURE(sts)) {
break;
}
uhash_put(gLocExtKeyMap, (void*)legacyKeyId, keyData, &sts);
if (legacyKeyId != bcpKeyId) {
// different key value
uhash_put(gLocExtKeyMap, (void*)bcpKeyId, keyData, &sts);
}
if (U_FAILURE(sts)) {
break;
}
}
}