/*
*******************************************************************************
-* Copyright (C) 1997-2004, International Business Machines Corporation and *
-* others. All Rights Reserved. *
+* Copyright (C) 1997-2016, International Business Machines Corporation and
+* others. All Rights Reserved.
*******************************************************************************
*
-* File TXTBDRY.CPP
+* File brkiter.cpp
*
* Modification History:
*
#if !UCONFIG_NO_BREAK_ITERATION
-#include "unicode/dbbi.h"
+#include "unicode/rbbi.h"
#include "unicode/brkiter.h"
#include "unicode/udata.h"
#include "unicode/ures.h"
+#include "unicode/ustring.h"
+#include "unicode/filteredbrk.h"
#include "ucln_cmn.h"
#include "cstring.h"
-#include "mutex.h"
-#include "iculserv.h"
+#include "umutex.h"
+#include "servloc.h"
#include "locbased.h"
#include "uresimp.h"
+#include "uassert.h"
+#include "ubrkimpl.h"
+#include "charstr.h"
+#include "unicode/filteredbrk.h"
// *****************************************************************************
// class BreakIterator
U_NAMESPACE_BEGIN
-const int32_t BreakIterator::DONE = (int32_t)-1;
-
// -------------------------------------
BreakIterator*
-BreakIterator::buildInstance(const Locale& loc, const char *type, UBool dict, UErrorCode &status)
+BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind, UErrorCode &status)
{
char fnbuff[256];
- char actualLocale[ULOC_FULLNAME_CAPACITY];
+ char ext[4]={'\0'};
+ CharString actualLocale;
int32_t size;
const UChar* brkfname = NULL;
- UResourceBundle brkRulesStack, brkNameStack;
- UResourceBundle *brkRules = &brkRulesStack, *brkName = &brkNameStack;
- BreakIterator *result = NULL;
-
+ UResourceBundle brkRulesStack;
+ UResourceBundle brkNameStack;
+ UResourceBundle *brkRules = &brkRulesStack;
+ UResourceBundle *brkName = &brkNameStack;
+ RuleBasedBreakIterator *result = NULL;
+
if (U_FAILURE(status))
return NULL;
ures_initStackObject(brkName);
// Get the locale
- UResourceBundle *b = ures_open(NULL, loc.getName(), &status);
+ UResourceBundle *b = ures_openNoDefault(U_ICUDATA_BRKITR, loc.getName(), &status);
// Get the "boundaries" array.
if (U_SUCCESS(status)) {
brkName = ures_getByKeyWithFallback(brkRules, type, brkName, &status);
// Get the actual string
brkfname = ures_getString(brkName, &size, &status);
+ U_ASSERT((size_t)size<sizeof(fnbuff));
+ if ((size_t)size>=sizeof(fnbuff)) {
+ size=0;
+ if (U_SUCCESS(status)) {
+ status = U_BUFFER_OVERFLOW_ERROR;
+ }
+ }
// Use the string if we found it
if (U_SUCCESS(status) && brkfname) {
- uprv_strncpy(actualLocale,
- ures_getLocale(brkName, &status),
- sizeof(actualLocale)/sizeof(actualLocale[0]));
- u_UCharsToChars(brkfname, fnbuff, size+1);
+ actualLocale.append(ures_getLocaleInternal(brkName, &status), -1, status);
+
+ UChar* extStart=u_strchr(brkfname, 0x002e);
+ int len = 0;
+ if(extStart!=NULL){
+ len = (int)(extStart-brkfname);
+ u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
+ u_UCharsToChars(brkfname, fnbuff, len);
+ }
+ fnbuff[len]=0; // nul terminate
}
}
ures_close(brkRules);
ures_close(brkName);
-
- UDataMemory* file = udata_open(NULL, "brk", fnbuff, &status);
+
+ UDataMemory* file = udata_open(U_ICUDATA_BRKITR, ext, fnbuff, &status);
if (U_FAILURE(status)) {
ures_close(b);
return NULL;
}
- // We found the break rules; now see if a dictionary is needed
- if (dict)
- {
- UErrorCode localStatus = U_ZERO_ERROR;
- brkName = &brkNameStack;
- ures_initStackObject(brkName);
- brkName = ures_getByKeyWithFallback(b, "BreakDictionaryData", brkName, &localStatus);
-#if 0
- if (U_SUCCESS(localStatus)) {
- brkfname = ures_getString(&brkname, &size, &localStatus);
- }
-#endif
- if (U_SUCCESS(localStatus)) {
-#if 0
- u_UCharsToChars(brkfname, fnbuff, size);
- fnbuff[size] = '\0';
-#endif
- result = new DictionaryBasedBreakIterator(file, "thaidict.brk", status);
- }
- ures_close(brkName);
- }
-
- // If there is still no result but we haven't had an error, no dictionary,
- // so make a non-dictionary break iterator
- if (U_SUCCESS(status) && result == NULL) {
- result = new RuleBasedBreakIterator(file, status);
- }
+ // Create a RuleBasedBreakIterator
+ result = new RuleBasedBreakIterator(file, status);
- // If there is a result, set the valid locale and actual locale
+ // If there is a result, set the valid locale and actual locale, and the kind
if (U_SUCCESS(status) && result != NULL) {
- U_LOCALE_BASED(locBased, *result);
- locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), actualLocale);
+ U_LOCALE_BASED(locBased, *(BreakIterator*)result);
+ locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status),
+ actualLocale.data());
+ result->setBreakType(kind);
}
ures_close(b);
-
+
if (U_FAILURE(status) && result != NULL) { // Sometimes redundant check, but simple
delete result;
return NULL;
return Locale::getAvailableLocales(count);
}
-// -------------------------------------
-// Gets the objectLocale display name in the default locale language.
-UnicodeString& U_EXPORT2
-BreakIterator::getDisplayName(const Locale& objectLocale,
- UnicodeString& name)
-{
- return objectLocale.getDisplayName(name);
-}
-
-// -------------------------------------
-// Gets the objectLocale display name in the displayLocale language.
-UnicodeString& U_EXPORT2
-BreakIterator::getDisplayName(const Locale& objectLocale,
- const Locale& displayLocale,
- UnicodeString& name)
-{
- return objectLocale.getDisplayName(displayLocale, name);
-}
-
// ------------------------------------------
//
// Default constructor and destructor
//-------------------------------------------
BreakIterator::BreakIterator()
+: fKeepAll(FALSE)
{
- fBufferClone = FALSE;
*validLocale = *actualLocale = 0;
}
//-------------------------------------------
#if !UCONFIG_NO_SERVICE
-static ICULocaleService* gService = NULL;
-
// -------------------------------------
class ICUBreakIteratorFactory : public ICUResourceBundleFactory {
+public:
+ virtual ~ICUBreakIteratorFactory();
protected:
virtual UObject* handleCreate(const Locale& loc, int32_t kind, const ICUService* /*service*/, UErrorCode& status) const {
return BreakIterator::makeInstance(loc, kind, status);
}
};
+ICUBreakIteratorFactory::~ICUBreakIteratorFactory() {}
+
// -------------------------------------
class ICUBreakIteratorService : public ICULocaleService {
UErrorCode status = U_ZERO_ERROR;
registerFactory(new ICUBreakIteratorFactory(), status);
}
-
+
+ virtual ~ICUBreakIteratorService();
+
virtual UObject* cloneInstance(UObject* instance) const {
return ((BreakIterator*)instance)->clone();
}
-
+
virtual UObject* handleDefault(const ICUServiceKey& key, UnicodeString* /*actualID*/, UErrorCode& status) const {
LocaleKey& lkey = (LocaleKey&)key;
int32_t kind = lkey.kind();
lkey.currentLocale(loc);
return BreakIterator::makeInstance(loc, kind, status);
}
-
+
virtual UBool isDefault() const {
return countFactories() == 1;
}
};
+ICUBreakIteratorService::~ICUBreakIteratorService() {}
+
// -------------------------------------
+// defined in ucln_cmn.h
U_NAMESPACE_END
-// defined in ucln_cmn.h
+static icu::UInitOnce gInitOnce;
+static icu::ICULocaleService* gService = NULL;
+
+
/**
- * Release all static memory held by breakiterator.
+ * Release all static memory held by breakiterator.
*/
U_CDECL_BEGIN
static UBool U_CALLCONV breakiterator_cleanup(void) {
delete gService;
gService = NULL;
}
+ gInitOnce.reset();
#endif
return TRUE;
}
U_CDECL_END
U_NAMESPACE_BEGIN
-static ICULocaleService*
+static void U_CALLCONV
+initService(void) {
+ gService = new ICUBreakIteratorService();
+ ucln_common_registerCleanup(UCLN_COMMON_BREAKITERATOR, breakiterator_cleanup);
+}
+
+static ICULocaleService*
getService(void)
{
- UBool needsInit;
- umtx_lock(NULL);
- needsInit = (UBool)(gService == NULL);
- umtx_unlock(NULL);
-
- if (needsInit) {
- ICULocaleService *tService = new ICUBreakIteratorService();
- umtx_lock(NULL);
- if (gService == NULL) {
- gService = tService;
- tService = NULL;
- ucln_common_registerCleanup(UCLN_COMMON_BREAKITERATOR, breakiterator_cleanup);
- }
- umtx_unlock(NULL);
- delete tService;
- }
+ umtx_initOnce(gInitOnce, &initService);
return gService;
}
+
// -------------------------------------
-static UBool
-hasService(void)
+static inline UBool
+hasService(void)
{
- Mutex mutex;
- return gService != NULL;
+ return !gInitOnce.isReset() && getService() != NULL;
}
// -------------------------------------
URegistryKey U_EXPORT2
-BreakIterator::registerInstance(BreakIterator* toAdopt, const Locale& locale, UBreakIteratorType kind, UErrorCode& status)
+BreakIterator::registerInstance(BreakIterator* toAdopt, const Locale& locale, UBreakIteratorType kind, UErrorCode& status)
{
- return getService()->registerInstance(toAdopt, locale, kind, status);
+ ICULocaleService *service = getService();
+ if (service == NULL) {
+ status = U_MEMORY_ALLOCATION_ERROR;
+ return NULL;
+ }
+ return service->registerInstance(toAdopt, locale, kind, status);
}
// -------------------------------------
UBool U_EXPORT2
-BreakIterator::unregister(URegistryKey key, UErrorCode& status)
+BreakIterator::unregister(URegistryKey key, UErrorCode& status)
{
if (U_SUCCESS(status)) {
if (hasService()) {
return gService->unregister(key, status);
}
- status = U_ILLEGAL_ARGUMENT_ERROR;
+ status = U_MEMORY_ALLOCATION_ERROR;
}
return FALSE;
}
StringEnumeration* U_EXPORT2
BreakIterator::getAvailableLocales(void)
{
- return getService()->getAvailableLocales();
+ ICULocaleService *service = getService();
+ if (service == NULL) {
+ return NULL;
+ }
+ return service->getAvailableLocales();
}
#endif /* UCONFIG_NO_SERVICE */
// -------------------------------------
BreakIterator*
-BreakIterator::createInstance(const Locale& loc, UBreakIteratorType kind, UErrorCode& status)
+BreakIterator::createInstance(const Locale& loc, int32_t kind, UErrorCode& status)
{
if (U_FAILURE(status)) {
return NULL;
}
-
- u_init(&status);
+
#if !UCONFIG_NO_SERVICE
if (hasService()) {
- Locale actualLoc;
+ Locale actualLoc("");
BreakIterator *result = (BreakIterator*)gService->get(loc, kind, &actualLoc, status);
// TODO: The way the service code works in ICU 2.8 is that if
// there is a real registered break iterator, the actualLoc
}
// -------------------------------------
+enum { kKeyValueLenMax = 32 };
-BreakIterator*
+BreakIterator*
BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
{
if (U_FAILURE(status)) {
return NULL;
}
+ char lbType[kKeyValueLenMax];
BreakIterator *result = NULL;
switch (kind) {
- case UBRK_CHARACTER:
- result = BreakIterator::buildInstance(loc, "grapheme", FALSE, status);
+ case UBRK_CHARACTER:
+ result = BreakIterator::buildInstance(loc, "grapheme", kind, status);
break;
case UBRK_WORD:
- result = BreakIterator::buildInstance(loc, "word", TRUE, status);
+ result = BreakIterator::buildInstance(loc, "word", kind, status);
break;
case UBRK_LINE:
- result = BreakIterator::buildInstance(loc, "line", TRUE, status);
+ uprv_strcpy(lbType, "line");
+ {
+ char lbKeyValue[kKeyValueLenMax] = {0};
+ UErrorCode kvStatus = U_ZERO_ERROR;
+ int32_t kLen = loc.getKeywordValue("lb", lbKeyValue, kKeyValueLenMax, kvStatus);
+ if (U_SUCCESS(kvStatus) && kLen > 0 && (uprv_strcmp(lbKeyValue,"strict")==0 || uprv_strcmp(lbKeyValue,"normal")==0 || uprv_strcmp(lbKeyValue,"loose")==0)) {
+ uprv_strcat(lbType, "_");
+ uprv_strcat(lbType, lbKeyValue);
+ }
+ }
+ result = BreakIterator::buildInstance(loc, lbType, kind, status);
+ if (U_SUCCESS(status) && result != NULL) {
+ char lwKeyValue[kKeyValueLenMax] = {0};
+ UErrorCode kvStatus = U_ZERO_ERROR;
+ int32_t kLen = loc.getKeywordValue("lw", lwKeyValue, kKeyValueLenMax, kvStatus);
+ result->setKeepAll(U_SUCCESS(kvStatus) && kLen > 0 && uprv_strcmp(lwKeyValue,"keepall")==0);
+ }
break;
case UBRK_SENTENCE:
- result = BreakIterator::buildInstance(loc, "sentence", FALSE, status);
+ result = BreakIterator::buildInstance(loc, "sentence", kind, status);
+ {
+ char ssKeyValue[kKeyValueLenMax] = {0};
+ UErrorCode kvStatus = U_ZERO_ERROR;
+ int32_t kLen = loc.getKeywordValue("ss", ssKeyValue, kKeyValueLenMax, kvStatus);
+ if (U_SUCCESS(kvStatus) && kLen > 0 && uprv_strcmp(ssKeyValue,"standard")==0) {
+ FilteredBreakIteratorBuilder* fbiBuilder = FilteredBreakIteratorBuilder::createInstance(loc, kvStatus);
+ if (U_SUCCESS(kvStatus)) {
+ result = fbiBuilder->build(result, status);
+ delete fbiBuilder;
+ }
+ }
+ }
break;
case UBRK_TITLE:
- result = BreakIterator::buildInstance(loc, "title", FALSE, status);
+ result = BreakIterator::buildInstance(loc, "title", kind, status);
break;
default:
status = U_ILLEGAL_ARGUMENT_ERROR;
return result;
}
-Locale
+Locale
BreakIterator::getLocale(ULocDataLocaleType type, UErrorCode& status) const {
U_LOCALE_BASED(locBased, *this);
return locBased.getLocale(type, status);
return locBased.getLocaleID(type, status);
}
+
+// This implementation of getRuleStatus is a do-nothing stub, here to
+// provide a default implementation for any derived BreakIterator classes that
+// do not implement it themselves.
+int32_t BreakIterator::getRuleStatus() const {
+ return 0;
+}
+
+// This implementation of getRuleStatusVec is a do-nothing stub, here to
+// provide a default implementation for any derived BreakIterator classes that
+// do not implement it themselves.
+int32_t BreakIterator::getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status) {
+ if (U_FAILURE(status)) {
+ return 0;
+ }
+ if (capacity < 1) {
+ status = U_BUFFER_OVERFLOW_ERROR;
+ return 1;
+ }
+ *fillInVec = 0;
+ return 1;
+}
+
+BreakIterator::BreakIterator (const Locale& valid, const Locale& actual) {
+ U_LOCALE_BASED(locBased, (*this));
+ locBased.setLocaleIDs(valid, actual);
+}
+
U_NAMESPACE_END
#endif /* #if !UCONFIG_NO_BREAK_ITERATION */