virtual void put(const char *key, ResourceValue &value, UBool, UErrorCode &errorCode) {
ResourceTable dayPeriodData = value.getTable(errorCode);
if (U_FAILURE(errorCode)) { return; }
for (int32_t i = 0; dayPeriodData.getKeyAndValue(i, key, value); ++i) {
if (uprv_strcmp(key, "locales") == 0) {
ResourceTable locales = value.getTable(errorCode);
if (U_FAILURE(errorCode)) { return; }
for (int32_t j = 0; locales.getKeyAndValue(j, key, value); ++j) {
UnicodeString setNum_str = value.getUnicodeString(errorCode);
int32_t setNum = parseSetNum(setNum_str, errorCode);
uhash_puti(data->localeToRuleSetNumMap, const_cast<char *>(key), setNum, &errorCode);
}
} else if (uprv_strcmp(key, "rules") == 0) {
// Allocate one more than needed to skip [0]. See comment in parseSetNum().
data->rules = new DayPeriodRules[data->maxRuleSetNum + 1];
if (data->rules == NULL) {
errorCode = U_MEMORY_ALLOCATION_ERROR;
return;
}
ResourceTable rules = value.getTable(errorCode);
processRules(rules, key, value, errorCode);
if (U_FAILURE(errorCode)) { return; }
}
}
}
StringTrieBuilder::Node *
StringTrieBuilder::registerFinalValue(int32_t value, UErrorCode &errorCode) {
if(U_FAILURE(errorCode)) {
return NULL;
}
FinalValueNode key(value);
const UHashElement *old=uhash_find(nodes, &key);
if(old!=NULL) {
return (Node *)old->key.pointer;
}
Node *newNode=new FinalValueNode(value);
if(newNode==NULL) {
errorCode=U_MEMORY_ALLOCATION_ERROR;
return NULL;
}
// If uhash_puti() returns a non-zero value from an equivalent, previously
// registered node, then uhash_find() failed to find that and we will leak newNode.
#if !U_RELEASE
int32_t oldValue= // Only in debug mode to avoid a compiler warning about unused oldValue.
#endif
uhash_puti(nodes, newNode, 1, &errorCode);
U_ASSERT(oldValue==0);
if(U_FAILURE(errorCode)) {
delete newNode;
return NULL;
}
return newNode;
}
virtual void put(const char *key, const ResourceValue &value, UErrorCode &errorCode) {
if (U_FAILURE(errorCode)) { return; }
UnicodeString setNum_str = value.getUnicodeString(errorCode);
int32_t setNum = parseSetNum(setNum_str, errorCode);
uhash_puti(data->localeToRuleSetNumMap, const_cast<char *>(key), setNum, &errorCode);
}
static void _put(UHashtable* hash,
const char* key,
int32_t value,
int32_t expectedOldValue) {
UErrorCode status = U_ZERO_ERROR;
int32_t oldValue =
uhash_puti(hash, (void*) key, value, &status);
if (U_FAILURE(status)) {
log_err("FAIL: uhash_put(%s) failed with %s and returned %ld\n",
key, u_errorName(status), oldValue);
} else if (oldValue != expectedOldValue) {
log_err("FAIL: uhash_put(%s) returned old value %ld; expected %ld\n",
key, oldValue, expectedOldValue);
} else {
log_verbose("Ok: uhash_put(%s, %d) returned old value %ld\n",
key, value, oldValue);
}
}
int32_t writeValueAliases(const Value &value, UErrorCode &errorCode) {
int32_t nameOffset=uhash_geti(nameGroupToOffset, (void *)value.joinedAliases);
if(nameOffset!=0) {
// The same list of aliases has been written already.
return nameOffset-1; // Was incremented to reserve 0 for "not found".
}
// Write this not-yet-seen list of aliases.
nameOffset=nameGroups.length();
uhash_puti(nameGroupToOffset, (void *)value.joinedAliases,
nameOffset+1, &errorCode);
// The first byte tells us how many aliases there are.
// We use only values 0..0x1f in the first byte because when we write
// the name groups as an invariant-character string into a source file,
// those values (C0 control codes) are written as numbers rather than as characters.
int32_t count=value.count;
if(count>=0x20) {
fprintf(stderr, "Error: Too many aliases in \"%s\"\n", value.joinedAliases);
exit(U_INDEX_OUTOFBOUNDS_ERROR);
}
nameGroups.append((char)count, errorCode);
// There is at least a short name (sometimes empty) and a long name. (count>=2)
// Note: Sometimes the short and long names are the same.
// In such a case, we could set a flag and omit the duplicate,
// but that would save only about 1.35% of total data size (Unicode 6.0/ICU 4.6)
// which is not worth the trouble.
// Note: In Unicode 6.1, there are more duplicates due to newly added
// short names for blocks and other properties.
// It might now be worth changing the data structure.
for(int32_t i=0; i<count; ++i) {
const char *s=value.aliases[i];
int32_t sLength=uprv_strlen(s)+1;
if(sLength>maxNameLength) {
maxNameLength=sLength;
}
nameGroups.append(s, sLength, errorCode); // including NUL
}
return nameOffset;
}
IdentifierInfo &IdentifierInfo::setIdentifier(const UnicodeString &identifier, UErrorCode &status) {
if (U_FAILURE(status)) {
return *this;
}
*fIdentifier = identifier;
clear();
ScriptSet scriptsForCP;
UChar32 cp;
for (int32_t i = 0; i < identifier.length(); i += U16_LENGTH(cp)) {
cp = identifier.char32At(i);
// Store a representative character for each kind of decimal digit
if (u_charType(cp) == U_DECIMAL_DIGIT_NUMBER) {
// Just store the zero character as a representative for comparison. Unicode guarantees it is cp - value
fNumerics->add(cp - (UChar32)u_getNumericValue(cp));
}
UScriptCode extensions[500];
int32_t extensionsCount = uscript_getScriptExtensions(cp, extensions, UPRV_LENGTHOF(extensions), &status);
if (U_FAILURE(status)) {
return *this;
}
scriptsForCP.resetAll();
for (int32_t j=0; j<extensionsCount; j++) {
scriptsForCP.set(extensions[j], status);
}
scriptsForCP.reset(USCRIPT_COMMON, status);
scriptsForCP.reset(USCRIPT_INHERITED, status);
switch (scriptsForCP.countMembers()) {
case 0: break;
case 1:
// Single script, record it.
fRequiredScripts->Union(scriptsForCP);
break;
default:
if (!fRequiredScripts->intersects(scriptsForCP)
&& !uhash_geti(fScriptSetSet, &scriptsForCP)) {
// If the set hasn't been added already, add it
// (Add a copy, fScriptSetSet takes ownership of the copy.)
uhash_puti(fScriptSetSet, new ScriptSet(scriptsForCP), 1, &status);
}
break;
}
}
// Now make a final pass through ScriptSetSet to remove alternates that came before singles.
// [Kana], [Kana Hira] => [Kana]
// This is relatively infrequent, so doesn't have to be optimized.
// We also compute any commonalities among the alternates.
if (uhash_count(fScriptSetSet) > 0) {
fCommonAmongAlternates->setAll();
for (int32_t it = UHASH_FIRST;;) {
const UHashElement *nextHashEl = uhash_nextElement(fScriptSetSet, &it);
if (nextHashEl == NULL) {
break;
}
ScriptSet *next = static_cast<ScriptSet *>(nextHashEl->key.pointer);
// [Kana], [Kana Hira] => [Kana]
if (fRequiredScripts->intersects(*next)) {
uhash_removeElement(fScriptSetSet, nextHashEl);
} else {
fCommonAmongAlternates->intersect(*next);
// [[Arab Syrc Thaa]; [Arab Syrc]] => [[Arab Syrc]]
for (int32_t otherIt = UHASH_FIRST;;) {
const UHashElement *otherHashEl = uhash_nextElement(fScriptSetSet, &otherIt);
if (otherHashEl == NULL) {
break;
}
ScriptSet *other = static_cast<ScriptSet *>(otherHashEl->key.pointer);
if (next != other && next->contains(*other)) {
uhash_removeElement(fScriptSetSet, nextHashEl);
break;
}
}
}
}
}
if (uhash_count(fScriptSetSet) == 0) {
fCommonAmongAlternates->resetAll();
}
return *this;
}
//--------------------------------------------------------------------------
//
// Assignment Operator
//
//--------------------------------------------------------------------------
RegexPattern &RegexPattern::operator = (const RegexPattern &other) {
if (this == &other) {
// Source and destination are the same. Don't do anything.
return *this;
}
// Clean out any previous contents of object being assigned to.
zap();
// Give target object a default initialization
init();
// Copy simple fields
fDeferredStatus = other.fDeferredStatus;
if (U_FAILURE(fDeferredStatus)) {
return *this;
}
if (other.fPatternString == NULL) {
fPatternString = NULL;
fPattern = utext_clone(fPattern, other.fPattern, FALSE, TRUE, &fDeferredStatus);
} else {
fPatternString = new UnicodeString(*(other.fPatternString));
if (fPatternString == NULL) {
fDeferredStatus = U_MEMORY_ALLOCATION_ERROR;
} else {
fPattern = utext_openConstUnicodeString(NULL, fPatternString, &fDeferredStatus);
}
}
if (U_FAILURE(fDeferredStatus)) {
return *this;
}
fFlags = other.fFlags;
fLiteralText = other.fLiteralText;
fMinMatchLen = other.fMinMatchLen;
fFrameSize = other.fFrameSize;
fDataSize = other.fDataSize;
fStaticSets = other.fStaticSets;
fStaticSets8 = other.fStaticSets8;
fStartType = other.fStartType;
fInitialStringIdx = other.fInitialStringIdx;
fInitialStringLen = other.fInitialStringLen;
*fInitialChars = *other.fInitialChars;
fInitialChar = other.fInitialChar;
*fInitialChars8 = *other.fInitialChars8;
fNeedsAltInput = other.fNeedsAltInput;
// Copy the pattern. It's just values, nothing deep to copy.
fCompiledPat->assign(*other.fCompiledPat, fDeferredStatus);
fGroupMap->assign(*other.fGroupMap, fDeferredStatus);
// Copy the Unicode Sets.
// Could be made more efficient if the sets were reference counted and shared,
// but I doubt that pattern copying will be particularly common.
// Note: init() already added an empty element zero to fSets
int32_t i;
int32_t numSets = other.fSets->size();
fSets8 = new Regex8BitSet[numSets];
if (fSets8 == NULL) {
fDeferredStatus = U_MEMORY_ALLOCATION_ERROR;
return *this;
}
for (i=1; i<numSets; i++) {
if (U_FAILURE(fDeferredStatus)) {
return *this;
}
UnicodeSet *sourceSet = (UnicodeSet *)other.fSets->elementAt(i);
UnicodeSet *newSet = new UnicodeSet(*sourceSet);
if (newSet == NULL) {
fDeferredStatus = U_MEMORY_ALLOCATION_ERROR;
break;
}
fSets->addElement(newSet, fDeferredStatus);
fSets8[i] = other.fSets8[i];
}
// Copy the named capture group hash map.
int32_t hashPos = UHASH_FIRST;
while (const UHashElement *hashEl = uhash_nextElement(other.fNamedCaptureMap, &hashPos)) {
if (U_FAILURE(fDeferredStatus)) {
break;
}
const UnicodeString *name = (const UnicodeString *)hashEl->key.pointer;
UnicodeString *key = new UnicodeString(*name);
int32_t val = hashEl->value.integer;
if (key == NULL) {
fDeferredStatus = U_MEMORY_ALLOCATION_ERROR;
} else {
uhash_puti(fNamedCaptureMap, key, val, &fDeferredStatus);
}
}
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 PluralRulesTest::testAvailbleLocales() {
// Hash set of (char *) strings.
UErrorCode status = U_ZERO_ERROR;
UHashtable *localeSet = uhash_open(uhash_hashUnicodeString, uhash_compareUnicodeString, uhash_compareLong, &status);
uhash_setKeyDeleter(localeSet, uprv_deleteUObject);
if (U_FAILURE(status)) {
errln("file %s, line %d: Error status = %s", __FILE__, __LINE__, u_errorName(status));
return;
}
// Check that each locale returned by the iterator is unique.
StringEnumeration *localesEnum = PluralRules::getAvailableLocales(status);
int localeCount = 0;
for (;;) {
const char *locale = localesEnum->next(NULL, status);
if (U_FAILURE(status)) {
dataerrln("file %s, line %d: Error status = %s", __FILE__, __LINE__, u_errorName(status));
return;
}
if (locale == NULL) {
break;
}
localeCount++;
int32_t oldVal = uhash_puti(localeSet, new UnicodeString(locale), 1, &status);
if (oldVal != 0) {
errln("file %s, line %d: locale %s was seen before.", __FILE__, __LINE__, locale);
}
}
// Reset the iterator, verify that we get the same count.
localesEnum->reset(status);
int32_t localeCount2 = 0;
while (localesEnum->next(NULL, status) != NULL) {
if (U_FAILURE(status)) {
errln("file %s, line %d: Error status = %s", __FILE__, __LINE__, u_errorName(status));
break;
}
localeCount2++;
}
if (localeCount != localeCount2) {
errln("file %s, line %d: locale counts differ. They are (%d, %d)",
__FILE__, __LINE__, localeCount, localeCount2);
}
// Instantiate plural rules for each available locale.
localesEnum->reset(status);
for (;;) {
status = U_ZERO_ERROR;
const char *localeName = localesEnum->next(NULL, status);
if (U_FAILURE(status)) {
errln("file %s, line %d: Error status = %s, locale = %s",
__FILE__, __LINE__, u_errorName(status), localeName);
return;
}
if (localeName == NULL) {
break;
}
Locale locale = Locale::createFromName(localeName);
PluralRules *pr = PluralRules::forLocale(locale, status);
if (U_FAILURE(status)) {
errln("file %s, line %d: Error %s creating plural rules for locale %s",
__FILE__, __LINE__, u_errorName(status), localeName);
continue;
}
if (pr == NULL) {
errln("file %s, line %d: Null plural rules for locale %s", __FILE__, __LINE__, localeName);
continue;
}
// Pump some numbers through the plural rules. Can't check for correct results,
// mostly this to tickle any asserts or crashes that may be lurking.
for (double n=0; n<120.0; n+=0.5) {
UnicodeString keyword = pr->select(n);
if (keyword.length() == 0) {
errln("file %s, line %d, empty keyword for n = %g, locale %s",
__FILE__, __LINE__, n, localeName);
}
}
delete pr;
}
uhash_close(localeSet);
delete localesEnum;
}
