C++ UnicodeString::isBogus方法代码示例

UVector& 
ICUService::getVisibleIDs(UVector& result, const UnicodeString* matchID, UErrorCode& status) const 
{
    result.removeAllElements();

    if (U_FAILURE(status)) {
        return result;
    }

    {
        Mutex mutex(&lock);
        const Hashtable* map = getVisibleIDMap(status);
        if (map != NULL) {
            ICUServiceKey* fallbackKey = createKey(matchID, status);

            for (int32_t pos = UHASH_FIRST;;) {
                const UHashElement* e = map->nextElement(pos);
                if (e == NULL) {
                    break;
                }

                const UnicodeString* id = (const UnicodeString*)e->key.pointer;
                if (fallbackKey != NULL) {
                    if (!fallbackKey->isFallbackOf(*id)) {
                        continue;
                    }
                }

                UnicodeString* idClone = new UnicodeString(*id);
                if (idClone == NULL || idClone->isBogus()) {
                    delete idClone;
                    status = U_MEMORY_ALLOCATION_ERROR;
                    break;
                }
                result.addElement(idClone, status);
                if (U_FAILURE(status)) {
                    delete idClone;
                    break;
                }
            }
            delete fallbackKey;
        }
    }
    if (U_FAILURE(status)) {
        result.removeAllElements();
    }
    return result;
}

void U_EXPORT2
Normalizer::normalize(const UnicodeString & source,
                      UNormalizationMode mode, int32_t options,
                      UnicodeString & result,
                      UErrorCode & status)
{
	if (source.isBogus() || U_FAILURE(status))
	{
		result.setToBogus();
		if (U_SUCCESS(status))
		{
			status = U_ILLEGAL_ARGUMENT_ERROR;
		}
	}
	else
	{
		UnicodeString localDest;
		UnicodeString * dest;

		if (&source != &result)
		{
			dest = &result;
		}
		else
		{
			// the source and result strings are the same object, use a temporary one
			dest = &localDest;
		}
		const Normalizer2 * n2 = Normalizer2Factory::getInstance(mode, status);
		if (U_SUCCESS(status))
		{
			if (options & UNORM_UNICODE_3_2)
			{
				FilteredNormalizer2(*n2, *uniset_getUnicode32Instance(status)).
				normalize(source, *dest, status);
			}
			else
			{
				n2->normalize(source, *dest, status);
			}
		}
		if (dest == &localDest && U_SUCCESS(status))
		{
			result = *dest;
		}
	}
}

UnicodeString& U_EXPORT2
ZoneMeta::getZoneIdByMetazone(const UnicodeString &mzid, const UnicodeString &region, UnicodeString &result) {
    UErrorCode status = U_ZERO_ERROR;
    const UChar *tzid = NULL;
    int32_t tzidLen = 0;
    char keyBuf[ZID_KEY_MAX + 1];
    int32_t keyLen = 0;

    if (mzid.isBogus() || mzid.length() > ZID_KEY_MAX) {
        result.setToBogus();
        return result;
    }

    keyLen = mzid.extract(0, mzid.length(), keyBuf, ZID_KEY_MAX + 1, US_INV);
    keyBuf[keyLen] = 0;

    UResourceBundle *rb = ures_openDirect(NULL, gMetaZones, &status);
    ures_getByKey(rb, gMapTimezonesTag, rb, &status);
    ures_getByKey(rb, keyBuf, rb, &status);

    if (U_SUCCESS(status)) {
        // check region mapping
        if (region.length() == 2 || region.length() == 3) {
            keyLen = region.extract(0, region.length(), keyBuf, ZID_KEY_MAX + 1, US_INV);
            keyBuf[keyLen] = 0;
            tzid = ures_getStringByKey(rb, keyBuf, &tzidLen, &status);
            if (status == U_MISSING_RESOURCE_ERROR) {
                status = U_ZERO_ERROR;
            }
        }
        if (U_SUCCESS(status) && tzid == NULL) {
            // try "001"
            tzid = ures_getStringByKey(rb, gWorldTag, &tzidLen, &status);
        }
    }
    ures_close(rb);

    if (tzid == NULL) {
        result.setToBogus();
    } else {
        result.setTo(tzid, tzidLen);
    }

    return result;
}

TransliteratorSpec::TransliteratorSpec(const UnicodeString& theSpec)
: top(theSpec),
  res(0)
{
    UErrorCode status = U_ZERO_ERROR;
    Locale topLoc("");
    LocaleUtility::initLocaleFromName(theSpec, topLoc);
    if (!topLoc.isBogus()) {
        res = new ResourceBundle(U_ICUDATA_TRANSLIT, topLoc, status);
        /* test for NULL */
        if (res == 0) {
            return;
        }
        if (U_FAILURE(status) || status == U_USING_DEFAULT_WARNING) {
            delete res;
            res = 0;
        }
    }

    // Canonicalize script name -or- do locale->script mapping
    status = U_ZERO_ERROR;
    static const int32_t capacity = 10;
    UScriptCode script[capacity]={USCRIPT_INVALID_CODE};
    int32_t num = uscript_getCode(CharString().appendInvariantChars(theSpec, status).data(),
                                  script, capacity, &status);
    if (num > 0 && script[0] != USCRIPT_INVALID_CODE) {
        scriptName = UnicodeString(uscript_getName(script[0]), -1, US_INV);
    }

    // Canonicalize top
    if (res != 0) {
        // Canonicalize locale name
        UnicodeString locStr;
        LocaleUtility::initNameFromLocale(topLoc, locStr);
        if (!locStr.isBogus()) {
            top = locStr;
        }
    } else if (scriptName.length() != 0) {
        // We are a script; use canonical name
        top = scriptName;
    }

    // assert(spec != top);
    reset();
}

//.........这里部分代码省略.........
            // first test of cache failed, so we'll have to update
            // the cache if we eventually succeed-- that is, if we're 
            // going to update the cache at all.
            putInCache = TRUE;

            int32_t index = startIndex;
            while (index < limit) {
                ICUServiceFactory* f = (ICUServiceFactory*)factories->elementAt(index++);
                UObject* service = f->create(key, this, status);
                if (U_FAILURE(status)) {
                    delete service;
                    return NULL;
                }
                if (service != NULL) {
                    result = new CacheEntry(currentDescriptor, service);
                    if (result == NULL) {
                        delete service;
                        status = U_MEMORY_ALLOCATION_ERROR;
                        return NULL;
                    }

                    goto outerEnd;
                }
            }

            // prepare to load the cache with all additional ids that 
            // will resolve to result, assuming we'll succeed.  We
            // don't want to keep querying on an id that's going to
            // fallback to the one that succeeded, we want to hit the
            // cache the first time next goaround.
            if (cacheDescriptorList._obj == NULL) {
                cacheDescriptorList._obj = new UVector(uprv_deleteUObject, NULL, 5, status);
                if (U_FAILURE(status)) {
                    return NULL;
                }
            }
            UnicodeString* idToCache = new UnicodeString(currentDescriptor);
            if (idToCache == NULL || idToCache->isBogus()) {
                status = U_MEMORY_ALLOCATION_ERROR;
                return NULL;
            }

            cacheDescriptorList._obj->addElement(idToCache, status);
            if (U_FAILURE(status)) {
                return NULL;
            }
        } while (key.fallback());
outerEnd:

        if (result != NULL) {
            if (putInCache && cacheResult) {
                serviceCache->put(result->actualDescriptor, result, status);
                if (U_FAILURE(status)) {
                    delete result;
                    return NULL;
                }

                if (cacheDescriptorList._obj != NULL) {
                    for (int32_t i = cacheDescriptorList._obj->size(); --i >= 0;) {
                        UnicodeString* desc = (UnicodeString*)cacheDescriptorList._obj->elementAt(i);
                        serviceCache->put(*desc, result, status);
                        if (U_FAILURE(status)) {
                            delete result;
                            return NULL;
                        }

                        result->ref();
                        cacheDescriptorList._obj->removeElementAt(i);
                    }
                }
            }

            if (actualReturn != NULL) {
                // strip null prefix
                if (result->actualDescriptor.indexOf((UChar)0x2f) == 0) { // U+002f=slash (/)
                    actualReturn->remove();
                    actualReturn->append(result->actualDescriptor, 
                        1, 
                        result->actualDescriptor.length() - 1);
                } else {
                    *actualReturn = result->actualDescriptor;
                }

                if (actualReturn->isBogus()) {
                    status = U_MEMORY_ALLOCATION_ERROR;
                    delete result;
                    return NULL;
                }
            }

            UObject* service = cloneInstance(result->service);
            if (putInCache && !cacheResult) {
                delete result;
            }
            return service;
        }
    }

    return handleDefault(key, actualReturn, status);
}

//.........这里部分代码省略.........
				"|z!s!", &tz_object, &locale, &locale_len) == FAILURE) {
			intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR,
				"intlgregcal_create_instance: bad arguments", 0 TSRMLS_CC);
			Z_OBJ_P(return_value) = NULL;
			return;
		}
	}
	if (variant > 2 && zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC,
			"lll|lll", &largs[0], &largs[1], &largs[2], &largs[3], &largs[4],
			&largs[5]) == FAILURE) {
		intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR,
			"intlgregcal_create_instance: bad arguments", 0 TSRMLS_CC);
		Z_OBJ_P(return_value) = NULL;
		return;
	}
	
	// instantion of ICU object
	GregorianCalendar *gcal = NULL;

	if (variant <= 2) {
		// From timezone and locale (0 to 2 arguments)
		TimeZone *tz = timezone_process_timezone_argument(tz_object, NULL,
			"intlgregcal_create_instance" TSRMLS_CC);
		if (tz == NULL) {
			Z_OBJ_P(return_value) = NULL;
			return;
		}
		if (!locale) {
			locale = const_cast<char*>(intl_locale_get_default(TSRMLS_C));
		}
		
		gcal = new GregorianCalendar(tz, Locale::createFromName(locale),
			status);
		if (U_FAILURE(status)) {
			intl_error_set(NULL, status, "intlgregcal_create_instance: error "
				"creating ICU GregorianCalendar from time zone and locale", 0 TSRMLS_CC);
			if (gcal) {
				delete gcal;
			}
			delete tz;
			Z_OBJ_P(return_value) = NULL;
			return;
		}
	} else {
		// From date/time (3, 5 or 6 arguments)
		for (int i = 0; i < variant; i++) {
			if (largs[i] < INT32_MIN || largs[i] > INT32_MAX) {
				intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR,
					"intlgregcal_create_instance: at least one of the arguments"
					" has an absolute value that is too large", 0 TSRMLS_CC);
				Z_OBJ_P(return_value) = NULL;
				return;
			}
		}

		if (variant == 3) {
			gcal = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
				(int32_t)largs[2], status);
		} else if (variant == 5) {
			gcal = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
				(int32_t)largs[2], (int32_t)largs[3], (int32_t)largs[4], status);
		} else if (variant == 6) {
			gcal = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
				(int32_t)largs[2], (int32_t)largs[3], (int32_t)largs[4], (int32_t)largs[5],
				status);
		}
		if (U_FAILURE(status)) {
			intl_error_set(NULL, status, "intlgregcal_create_instance: error "
				"creating ICU GregorianCalendar from date", 0 TSRMLS_CC);
			if (gcal) {
				delete gcal;
			}
			Z_OBJ_P(return_value) = NULL;
			return;
		}

		timelib_tzinfo *tzinfo = get_timezone_info(TSRMLS_C);
#if U_ICU_VERSION_MAJOR_NUM * 10 + U_ICU_VERSION_MINOR_NUM >= 42
		UnicodeString tzstr = UnicodeString::fromUTF8(StringPiece(tzinfo->name));
#else
		UnicodeString tzstr = UnicodeString(tzinfo->name,
			strlen(tzinfo->name), US_INV);
#endif
		if (tzstr.isBogus()) {
			intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR,
				"intlgregcal_create_instance: could not create UTF-8 string "
				"from PHP's default timezone name (see date_default_timezone_get())",
				0 TSRMLS_CC);
			delete gcal;
			Z_OBJ_P(return_value) = NULL;
			return;
		}

		TimeZone *tz = TimeZone::createTimeZone(tzstr);
		gcal->adoptTimeZone(tz);
	}
    
    Calendar_object *co = Z_INTL_CALENDAR_P(return_value);
    co->ucal = gcal;
}

void 
ICUServiceTest::testAPI_Two()
{
    UErrorCode status = U_ZERO_ERROR;
    TestStringService service;
    service.registerFactory(new AnonymousStringFactory(), status);

    // anonymous factory will still handle the id
    {
        UErrorCode status = U_ZERO_ERROR;
        const UnicodeString en_US = "en_US";
        UnicodeString* result = (UnicodeString*)service.get(en_US, status);
        confirmEqual("21) locale", result, &en_US);
        delete result;
    }

    // still normalizes id
    {
        UErrorCode status = U_ZERO_ERROR;
        const UnicodeString en_US_BAR = "en_US_BAR";
        UnicodeString resultID;
        UnicodeString* result = (UnicodeString*)service.get("EN_us_bar", &resultID, status);
        confirmEqual("22) locale", &resultID, &en_US_BAR);
        delete result;
    }

    // we can override for particular ids
    UnicodeString* singleton0 = new UnicodeString("Zero");
    service.registerInstance(singleton0, "en_US_BAR", status);
    {
        UErrorCode status = U_ZERO_ERROR;
        UnicodeString* result = (UnicodeString*)service.get("en_US_BAR", status);
        confirmEqual("23) override super", result, singleton0);
        delete result;
    }

    // empty service should not recognize anything 
    service.reset();
    {
        UErrorCode status = U_ZERO_ERROR;
        UnicodeString* result = (UnicodeString*)service.get("en_US", status);
        confirmIdentical("24) empty", result, NULL);
    }

    // create a custom multiple key factory
    {
        UnicodeString xids[] = {
            "en_US_VALLEY_GIRL", 
            "en_US_VALLEY_BOY",
            "en_US_SURFER_GAL",
            "en_US_SURFER_DUDE"
        };
        int32_t count = sizeof(xids)/sizeof(UnicodeString);

        ICUServiceFactory* f = new TestMultipleKeyStringFactory(xids, count, "Later");
        service.registerFactory(f, status);
    }

    // iterate over the visual ids returned by the multiple factory
    {
        UErrorCode status = U_ZERO_ERROR;
        UVector ids(uhash_deleteUnicodeString, uhash_compareUnicodeString, 0, status);
        service.getVisibleIDs(ids, status);
        for (int i = 0; i < ids.size(); ++i) {
            const UnicodeString* id = (const UnicodeString*)ids[i];
            UnicodeString* result = (UnicodeString*)service.get(*id, status);
            if (result) {
                logln("  " + *id + " --> " + *result);
                delete result;
            } else {
                errln("could not find " + *id);
            }
        }
        // four visible ids
        confirmIdentical("25) visible ids", ids.size(), 4);
    }

    // iterate over the display names
    {
        UErrorCode status = U_ZERO_ERROR;
        UVector names(status);
        service.getDisplayNames(names, status);
        for (int i = 0; i < names.size(); ++i) {
            const StringPair* pair = (const StringPair*)names[i];
            logln("  " + pair->displayName + " --> " + pair->id);
        }
        confirmIdentical("26) display names", names.size(), 4);
    }

    // no valid display name
    {
        UnicodeString name;
        service.getDisplayName("en_US_VALLEY_GEEK", name);
        confirmBoolean("27) get display name", name.isBogus());
    }

    {
        UnicodeString name;
        service.getDisplayName("en_US_SURFER_DUDE", name, Locale::getEnglish());
        confirmStringsEqual("28) get display name", name, "English (United States, SURFER_DUDE)");
//.........这里部分代码省略.........

void
LocaleDisplayNamesImpl::initialize(void) {
    LocaleDisplayNamesImpl *nonConstThis = (LocaleDisplayNamesImpl *)this;
    nonConstThis->locale = langData.getLocale() == Locale::getRoot()
        ? regionData.getLocale()
        : langData.getLocale();

    UnicodeString sep;
    langData.getNoFallback("localeDisplayPattern", "separator", sep);
    if (sep.isBogus()) {
        sep = UnicodeString("{0}, {1}", -1, US_INV);
    }
    UErrorCode status = U_ZERO_ERROR;
    separatorFormat.applyPatternMinMaxArguments(sep, 2, 2, status);

    UnicodeString pattern;
    langData.getNoFallback("localeDisplayPattern", "pattern", pattern);
    if (pattern.isBogus()) {
        pattern = UnicodeString("{0} ({1})", -1, US_INV);
    }
    format.applyPatternMinMaxArguments(pattern, 2, 2, status);
    if (pattern.indexOf((UChar)0xFF08) >= 0) {
        formatOpenParen.setTo((UChar)0xFF08);         // fullwidth (
        formatReplaceOpenParen.setTo((UChar)0xFF3B);  // fullwidth [
        formatCloseParen.setTo((UChar)0xFF09);        // fullwidth )
        formatReplaceCloseParen.setTo((UChar)0xFF3D); // fullwidth ]
    } else {
        formatOpenParen.setTo((UChar)0x0028);         // (
        formatReplaceOpenParen.setTo((UChar)0x005B);  // [
        formatCloseParen.setTo((UChar)0x0029);        // )
        formatReplaceCloseParen.setTo((UChar)0x005D); // ]
    }

    UnicodeString ktPattern;
    langData.get("localeDisplayPattern", "keyTypePattern", ktPattern);
    if (ktPattern.isBogus()) {
        ktPattern = UnicodeString("{0}={1}", -1, US_INV);
    }
    keyTypeFormat.applyPatternMinMaxArguments(ktPattern, 2, 2, status);

    uprv_memset(fCapitalization, 0, sizeof(fCapitalization));
#if !UCONFIG_NO_BREAK_ITERATION
    // Only get the context data if we need it! This is a const object so we know now...
    // Also check whether we will need a break iterator (depends on the data)
    UBool needBrkIter = FALSE;
    if (capitalizationContext == UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU || capitalizationContext == UDISPCTX_CAPITALIZATION_FOR_STANDALONE) {
        LocalUResourceBundlePointer resource(ures_open(NULL, locale.getName(), &status));
        if (U_FAILURE(status)) { return; }
        CapitalizationContextSink sink(*this);
        ures_getAllItemsWithFallback(resource.getAlias(), "contextTransforms", sink, status);
        if (status == U_MISSING_RESOURCE_ERROR) {
            // Silently ignore.  Not every locale has contextTransforms.
            status = U_ZERO_ERROR;
        } else if (U_FAILURE(status)) {
            return;
        }
        needBrkIter = sink.hasCapitalizationUsage;
    }
    // Get a sentence break iterator if we will need it
    if (needBrkIter || capitalizationContext == UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE) {
        status = U_ZERO_ERROR;
        capitalizationBrkIter = BreakIterator::createSentenceInstance(locale, status);
        if (U_FAILURE(status)) {
            delete capitalizationBrkIter;
            capitalizationBrkIter = NULL;
        }
    }
#endif
}

const UChar* U_EXPORT2
ZoneMeta::getCanonicalCLDRID(const UnicodeString &tzid, UErrorCode& status) {
    if (U_FAILURE(status)) {
        return NULL;
    }

    if (tzid.isBogus() || tzid.length() > ZID_KEY_MAX) {
        status = U_ILLEGAL_ARGUMENT_ERROR;
        return NULL;
    }

    // Checking the cached results
    umtx_initOnce(gCanonicalIDCacheInitOnce, &initCanonicalIDCache, status);
    if (U_FAILURE(status)) {
        return NULL;
    }

    const UChar *canonicalID = NULL;

    UErrorCode tmpStatus = U_ZERO_ERROR;
    UChar utzid[ZID_KEY_MAX + 1];
    tzid.extract(utzid, ZID_KEY_MAX + 1, tmpStatus);
    U_ASSERT(tmpStatus == U_ZERO_ERROR);    // we checked the length of tzid already

    if (!uprv_isInvariantUString(utzid, -1)) {
        // All of known tz IDs are only containing ASCII invariant characters.
        status = U_ILLEGAL_ARGUMENT_ERROR;
        return NULL;
    }

    // Check if it was already cached
    umtx_lock(gZoneMetaLock());
    {
        canonicalID = (const UChar *)uhash_get(gCanonicalIDCache, utzid);
    }
    umtx_unlock(gZoneMetaLock());

    if (canonicalID != NULL) {
        return canonicalID;
    }

    // If not, resolve CLDR canonical ID with resource data
    UBool isInputCanonical = FALSE;
    char id[ZID_KEY_MAX + 1];
    tzid.extract(0, 0x7fffffff, id, UPRV_LENGTHOF(id), US_INV);

    // replace '/' with ':'
    char *p = id;
    while (*p++) {
        if (*p == '/') {
            *p = ':';
        }
    }

    UResourceBundle *top = ures_openDirect(NULL, gKeyTypeData, &tmpStatus);
    UResourceBundle *rb = ures_getByKey(top, gTypeMapTag, NULL, &tmpStatus);
    ures_getByKey(rb, gTimezoneTag, rb, &tmpStatus);
    ures_getByKey(rb, id, rb, &tmpStatus);
    if (U_SUCCESS(tmpStatus)) {
        // type entry (canonical) found
        // the input is the canonical ID. resolve to const UChar*
        canonicalID = TimeZone::findID(tzid);
        isInputCanonical = TRUE;
    }

    if (canonicalID == NULL) {
        // If a map element not found, then look for an alias
        tmpStatus = U_ZERO_ERROR;
        ures_getByKey(top, gTypeAliasTag, rb, &tmpStatus);
        ures_getByKey(rb, gTimezoneTag, rb, &tmpStatus);
        const UChar *canonical = ures_getStringByKey(rb,id,NULL,&tmpStatus);
        if (U_SUCCESS(tmpStatus)) {
            // canonical map found
            canonicalID = canonical;
        }

        if (canonicalID == NULL) {
            // Dereference the input ID using the tz data
            const UChar *derefer = TimeZone::dereferOlsonLink(tzid);
            if (derefer == NULL) {
                status = U_ILLEGAL_ARGUMENT_ERROR;
            } else {
                int32_t len = u_strlen(derefer);
                u_UCharsToChars(derefer,id,len);
                id[len] = (char) 0; // Make sure it is null terminated.

                // replace '/' with ':'
                char *q = id;
                while (*q++) {
                    if (*q == '/') {
                        *q = ':';
                    }
                }

                // If a dereference turned something up then look for an alias.
                // rb still points to the alias table, so we don't have to go looking
                // for it.
                tmpStatus = U_ZERO_ERROR;
                canonical = ures_getStringByKey(rb,id,NULL,&tmpStatus);
                if (U_SUCCESS(tmpStatus)) {
//.........这里部分代码省略.........

UnicodeString&
LocaleDisplayNamesImpl::localeDisplayName(const Locale& locale,
                                          UnicodeString& result) const {
  if (locale.isBogus()) {
    result.setToBogus();
    return result;
  }
  UnicodeString resultName;

  const char* lang = locale.getLanguage();
  if (uprv_strlen(lang) == 0) {
    lang = "root";
  }
  const char* script = locale.getScript();
  const char* country = locale.getCountry();
  const char* variant = locale.getVariant();

  UBool hasScript = uprv_strlen(script) > 0;
  UBool hasCountry = uprv_strlen(country) > 0;
  UBool hasVariant = uprv_strlen(variant) > 0;

  if (dialectHandling == ULDN_DIALECT_NAMES) {
    char buffer[ULOC_FULLNAME_CAPACITY];
    do { // loop construct is so we can break early out of search
      if (hasScript && hasCountry) {
        ncat(buffer, ULOC_FULLNAME_CAPACITY, lang, "_", script, "_", country, (char *)0);
        localeIdName(buffer, resultName);
        if (!resultName.isBogus()) {
          hasScript = FALSE;
          hasCountry = FALSE;
          break;
        }
      }
      if (hasScript) {
        ncat(buffer, ULOC_FULLNAME_CAPACITY, lang, "_", script, (char *)0);
        localeIdName(buffer, resultName);
        if (!resultName.isBogus()) {
          hasScript = FALSE;
          break;
        }
      }
      if (hasCountry) {
        ncat(buffer, ULOC_FULLNAME_CAPACITY, lang, "_", country, (char*)0);
        localeIdName(buffer, resultName);
        if (!resultName.isBogus()) {
          hasCountry = FALSE;
          break;
        }
      }
    } while (FALSE);
  }
  if (resultName.isBogus() || resultName.isEmpty()) {
    localeIdName(lang, resultName);
  }

  UnicodeString resultRemainder;
  UnicodeString temp;
  UErrorCode status = U_ZERO_ERROR;

  if (hasScript) {
    resultRemainder.append(scriptDisplayName(script, temp, TRUE));
  }
  if (hasCountry) {
    appendWithSep(resultRemainder, regionDisplayName(country, temp, TRUE));
  }
  if (hasVariant) {
    appendWithSep(resultRemainder, variantDisplayName(variant, temp, TRUE));
  }
  resultRemainder.findAndReplace(formatOpenParen, formatReplaceOpenParen);
  resultRemainder.findAndReplace(formatCloseParen, formatReplaceCloseParen);

  LocalPointer<StringEnumeration> e(locale.createKeywords(status));
  if (e.isValid() && U_SUCCESS(status)) {
    UnicodeString temp2;
    char value[ULOC_KEYWORD_AND_VALUES_CAPACITY]; // sigh, no ULOC_VALUE_CAPACITY
    const char* key;
    while ((key = e->next((int32_t *)0, status)) != NULL) {
      locale.getKeywordValue(key, value, ULOC_KEYWORD_AND_VALUES_CAPACITY, status);
      if (U_FAILURE(status)) {
        return result;
      }
      keyDisplayName(key, temp, TRUE);
      temp.findAndReplace(formatOpenParen, formatReplaceOpenParen);
      temp.findAndReplace(formatCloseParen, formatReplaceCloseParen);
      keyValueDisplayName(key, value, temp2, TRUE);
      temp2.findAndReplace(formatOpenParen, formatReplaceOpenParen);
      temp2.findAndReplace(formatCloseParen, formatReplaceCloseParen);
      if (temp2 != UnicodeString(value, -1, US_INV)) {
        appendWithSep(resultRemainder, temp2);
      } else if (temp != UnicodeString(key, -1, US_INV)) {
        UnicodeString temp3;
        keyTypeFormat.format(temp, temp2, temp3, status);
        appendWithSep(resultRemainder, temp3);
      } else {
        appendWithSep(resultRemainder, temp)
          .append((UChar)0x3d /* = */)
          .append(temp2);
      }
    }
  }
//.........这里部分代码省略.........

void UTS46Test::TestAPI() {
    UErrorCode errorCode=U_ZERO_ERROR;
    UnicodeString result;
    IDNAInfo info;
    UnicodeString input=UNICODE_STRING_SIMPLE("www.eXample.cOm");
    UnicodeString expected=UNICODE_STRING_SIMPLE("www.example.com");
    trans->nameToASCII(input, result, info, errorCode);
    if(U_FAILURE(errorCode) || info.hasErrors() || result!=expected) {
        errln("T.nameToASCII(www.example.com) info.errors=%04lx result matches=%d %s",
              (long)info.getErrors(), result==expected, u_errorName(errorCode));
    }
    errorCode=U_USELESS_COLLATOR_ERROR;
    trans->nameToUnicode(input, result, info, errorCode);
    if(errorCode!=U_USELESS_COLLATOR_ERROR || !result.isBogus()) {
        errln("T.nameToUnicode(U_FAILURE) did not preserve the errorCode "
              "or not result.setToBogus() - %s",
              u_errorName(errorCode));
    }
    errorCode=U_ZERO_ERROR;
    input.setToBogus();
    result=UNICODE_STRING_SIMPLE("quatsch");
    nontrans->labelToASCII(input, result, info, errorCode);
    if(errorCode!=U_ILLEGAL_ARGUMENT_ERROR || !result.isBogus()) {
        errln("N.labelToASCII(bogus) did not set illegal-argument-error "
              "or not result.setToBogus() - %s",
              u_errorName(errorCode));
    }
    errorCode=U_ZERO_ERROR;
    input=UNICODE_STRING_SIMPLE("xn--bcher.de-65a");
    expected=UNICODE_STRING_SIMPLE("xn--bcher\\uFFFDde-65a").unescape();
    nontrans->labelToASCII(input, result, info, errorCode);
    if( U_FAILURE(errorCode) ||
        info.getErrors()!=(UIDNA_ERROR_LABEL_HAS_DOT|UIDNA_ERROR_INVALID_ACE_LABEL) ||
        result!=expected
    ) {
        errln("N.labelToASCII(label-with-dot) failed with errors %04lx - %s",
              info.getErrors(), u_errorName(errorCode));
    }
    // UTF-8
    char buffer[100];
    TestCheckedArrayByteSink sink(buffer, UPRV_LENGTHOF(buffer));
    errorCode=U_ZERO_ERROR;
    nontrans->labelToUnicodeUTF8(StringPiece(NULL, 5), sink, info, errorCode);
    if(errorCode!=U_ILLEGAL_ARGUMENT_ERROR || sink.NumberOfBytesWritten()!=0) {
        errln("N.labelToUnicodeUTF8(StringPiece(NULL, 5)) did not set illegal-argument-error ",
              "or did output something - %s",
              u_errorName(errorCode));
    }

    sink.Reset();
    errorCode=U_ZERO_ERROR;
    nontrans->nameToASCII_UTF8(StringPiece(), sink, info, errorCode);
    if(U_FAILURE(errorCode) || sink.NumberOfBytesWritten()!=0 || !sink.calledFlush) {
        errln("N.nameToASCII_UTF8(empty) failed - %s",
              u_errorName(errorCode));
    }

    static const char s[]={ 0x61, (char)0xc3, (char)0x9f };
    sink.Reset();
    errorCode=U_USELESS_COLLATOR_ERROR;
    nontrans->nameToUnicodeUTF8(StringPiece(s, 3), sink, info, errorCode);
    if(errorCode!=U_USELESS_COLLATOR_ERROR || sink.NumberOfBytesWritten()!=0) {
        errln("N.nameToUnicode_UTF8(U_FAILURE) did not preserve the errorCode "
              "or did output something - %s",
              u_errorName(errorCode));
    }

    sink.Reset();
    errorCode=U_ZERO_ERROR;
    trans->labelToUnicodeUTF8(StringPiece(s, 3), sink, info, errorCode);
    if( U_FAILURE(errorCode) || sink.NumberOfBytesWritten()!=3 ||
        buffer[0]!=0x61 || buffer[1]!=0x73 || buffer[2]!=0x73 ||
        !sink.calledFlush
    ) {
        errln("T.labelToUnicodeUTF8(a sharp-s) failed - %s",
              u_errorName(errorCode));
    }

    sink.Reset();
    errorCode=U_ZERO_ERROR;
    // "eXampLe.cOm"
    static const char eX[]={ 0x65, 0x58, 0x61, 0x6d, 0x70, 0x4c, 0x65, 0x2e, 0x63, 0x4f, 0x6d, 0 };
    // "example.com"
    static const char ex[]={ 0x65, 0x78, 0x61, 0x6d, 0x70, 0x6c, 0x65, 0x2e, 0x63, 0x6f, 0x6d };
    trans->nameToUnicodeUTF8(eX, sink, info, errorCode);
    if( U_FAILURE(errorCode) || sink.NumberOfBytesWritten()!=11 ||
        0!=memcmp(ex, buffer, 11) || !sink.calledFlush
    ) {
        errln("T.nameToUnicodeUTF8(eXampLe.cOm) failed - %s",
              u_errorName(errorCode));
    }
}

 inline UBool isSet() const {
   return !prefix.isBogus();
 }

bool DecimalMatcher::match(StringSegment& segment, ParsedNumber& result, int8_t exponentSign,
                           UErrorCode&) const {
    if (result.seenNumber() && exponentSign == 0) {
        // A number has already been consumed.
        return false;
    } else if (exponentSign != 0) {
        // scientific notation always comes after the number
        U_ASSERT(!result.quantity.bogus);
    }

    // Initial offset before any character consumption.
    int32_t initialOffset = segment.getOffset();

    // Return value: whether to ask for more characters.
    bool maybeMore = false;

    // All digits consumed so far.
    number::impl::DecimalQuantity digitsConsumed;
    digitsConsumed.bogus = true;

    // The total number of digits after the decimal place, used for scaling the result.
    int32_t digitsAfterDecimalPlace = 0;

    // The actual grouping and decimal separators used in the string.
    // If non-null, we have seen that token.
    UnicodeString actualGroupingString;
    UnicodeString actualDecimalString;
    actualGroupingString.setToBogus();
    actualDecimalString.setToBogus();

    // Information for two groups: the previous group and the current group.
    //
    // Each group has three pieces of information:
    //
    // Offset: the string position of the beginning of the group, including a leading separator
    // if there was a leading separator. This is needed in case we need to rewind the parse to
    // that position.
    //
    // Separator type:
    // 0 => beginning of string
    // 1 => lead separator is a grouping separator
    // 2 => lead separator is a decimal separator
    //
    // Count: the number of digits in the group. If -1, the group has been validated.
    int32_t currGroupOffset = 0;
    int32_t currGroupSepType = 0;
    int32_t currGroupCount = 0;
    int32_t prevGroupOffset = -1;
    int32_t prevGroupSepType = -1;
    int32_t prevGroupCount = -1;

    while (segment.length() > 0) {
        maybeMore = false;

        // Attempt to match a digit.
        int8_t digit = -1;

        // Try by code point digit value.
        UChar32 cp = segment.getCodePoint();
        if (u_isdigit(cp)) {
            segment.adjustOffset(U16_LENGTH(cp));
            digit = static_cast<int8_t>(u_digit(cp, 10));
        }

        // Try by digit string.
        if (digit == -1 && !fLocalDigitStrings.isNull()) {
            for (int32_t i = 0; i < 10; i++) {
                const UnicodeString& str = fLocalDigitStrings[i];
                if (str.isEmpty()) {
                    continue;
                }
                int32_t overlap = segment.getCommonPrefixLength(str);
                if (overlap == str.length()) {
                    segment.adjustOffset(overlap);
                    digit = static_cast<int8_t>(i);
                    break;
                }
                maybeMore = maybeMore || (overlap == segment.length());
            }
        }

        if (digit >= 0) {
            // Digit was found.
            if (digitsConsumed.bogus) {
                digitsConsumed.bogus = false;
                digitsConsumed.clear();
            }
            digitsConsumed.appendDigit(digit, 0, true);
            currGroupCount++;
            if (!actualDecimalString.isBogus()) {
                digitsAfterDecimalPlace++;
            }
            continue;
        }

        // Attempt to match a literal grouping or decimal separator.
        bool isDecimal = false;
        bool isGrouping = false;

        // 1) Attempt the decimal separator string literal.
//.........这里部分代码省略.........

/* {{{ timezone_convert_to_datetimezone
 *	   Convert from TimeZone to DateTimeZone object */
U_CFUNC zval *timezone_convert_to_datetimezone(const TimeZone *timeZone,
											   intl_error *outside_error,
											   const char *func, zval *ret)
{
	UnicodeString		id;
	char				*message = NULL;
	php_timezone_obj	*tzobj;
	zval				arg;

	timeZone->getID(id);
	if (id.isBogus()) {
		spprintf(&message, 0, "%s: could not obtain TimeZone id", func);
		intl_errors_set(outside_error, U_ILLEGAL_ARGUMENT_ERROR,
			message, 1);
		goto error;
	}

	object_init_ex(ret, php_date_get_timezone_ce());
	tzobj = Z_PHPTIMEZONE_P(ret);

	if (id.compare(0, 3, UnicodeString("GMT", sizeof("GMT")-1, US_INV)) == 0) {
		/* The DateTimeZone constructor doesn't support offset time zones,
		 * so we must mess with DateTimeZone structure ourselves */
		tzobj->initialized	  = 1;
		tzobj->type			  = TIMELIB_ZONETYPE_OFFSET;
		//convert offset from milliseconds to minutes
		tzobj->tzi.utc_offset = -1 * timeZone->getRawOffset() / (60 * 1000);
	} else {
		char *str;
		size_t str_len;
		/* Call the constructor! */
		if (intl_charFromString(id, &str, &str_len, &INTL_ERROR_CODE(*outside_error)) == FAILURE) {
			spprintf(&message, 0, "%s: could not convert id to UTF-8", func);
			intl_errors_set(outside_error, INTL_ERROR_CODE(*outside_error),
				message, 1);
			goto error;
		}
		ZVAL_STRINGL(&arg, str, str_len);
		//???
		efree(str);
		zend_call_method_with_1_params(ret, NULL, NULL, "__construct", NULL, &arg);
		if (EG(exception)) {
			spprintf(&message, 0,
				"%s: DateTimeZone constructor threw exception", func);
			intl_errors_set(outside_error, U_ILLEGAL_ARGUMENT_ERROR,
				message, 1);
			zend_object_store_ctor_failed(Z_OBJ_P(ret));
			zval_ptr_dtor(&arg);
			goto error;
		}
		zval_ptr_dtor(&arg);
	}

	if (0) {
error:
		if (ret) {
			zval_ptr_dtor(ret);
		}
		ret = NULL;
	}

	if (message) {
		efree(message);
	}
	return ret;
}

void
LocaleDisplayNamesImpl::initialize(void) {
    LocaleDisplayNamesImpl *nonConstThis = (LocaleDisplayNamesImpl *)this;
    nonConstThis->locale = langData.getLocale() == Locale::getRoot()
        ? regionData.getLocale()
        : langData.getLocale();

    UnicodeString sep;
    langData.getNoFallback("localeDisplayPattern", "separator", sep);
    if (sep.isBogus()) {
        sep = UnicodeString("{0}, {1}", -1, US_INV);
    }
    UErrorCode status = U_ZERO_ERROR;
    separatorFormat.applyPatternMinMaxArguments(sep, 2, 2, status);

    UnicodeString pattern;
    langData.getNoFallback("localeDisplayPattern", "pattern", pattern);
    if (pattern.isBogus()) {
        pattern = UnicodeString("{0} ({1})", -1, US_INV);
    }
    format.applyPatternMinMaxArguments(pattern, 2, 2, status);
    if (pattern.indexOf((UChar)0xFF08) >= 0) {
        formatOpenParen.setTo((UChar)0xFF08);         // fullwidth (
        formatReplaceOpenParen.setTo((UChar)0xFF3B);  // fullwidth [
        formatCloseParen.setTo((UChar)0xFF09);        // fullwidth )
        formatReplaceCloseParen.setTo((UChar)0xFF3D); // fullwidth ]
    } else {
        formatOpenParen.setTo((UChar)0x0028);         // (
        formatReplaceOpenParen.setTo((UChar)0x005B);  // [
        formatCloseParen.setTo((UChar)0x0029);        // )
        formatReplaceCloseParen.setTo((UChar)0x005D); // ]
    }

    UnicodeString ktPattern;
    langData.get("localeDisplayPattern", "keyTypePattern", ktPattern);
    if (ktPattern.isBogus()) {
        ktPattern = UnicodeString("{0}={1}", -1, US_INV);
    }
    keyTypeFormat.applyPatternMinMaxArguments(ktPattern, 2, 2, status);

    uprv_memset(fCapitalization, 0, sizeof(fCapitalization));
#if !UCONFIG_NO_BREAK_ITERATION
    // The following is basically copied from DateFormatSymbols::initializeData
    typedef struct {
        const char * usageName;
        LocaleDisplayNamesImpl::CapContextUsage usageEnum;
    } ContextUsageNameToEnum;
    const ContextUsageNameToEnum contextUsageTypeMap[] = {
       // Entries must be sorted by usageTypeName; entry with NULL name terminates list.
        { "key",        kCapContextUsageKey },
        { "keyValue",   kCapContextUsageKeyValue },
        { "languages",  kCapContextUsageLanguage },
        { "script",     kCapContextUsageScript },
        { "territory",  kCapContextUsageTerritory },
        { "variant",    kCapContextUsageVariant },
        { NULL,         (CapContextUsage)0 },
    };
    // Only get the context data if we need it! This is a const object so we know now...
    // Also check whether we will need a break iterator (depends on the data)
    UBool needBrkIter = FALSE;
    if (capitalizationContext == UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU || capitalizationContext == UDISPCTX_CAPITALIZATION_FOR_STANDALONE) {
        int32_t len = 0;
        UResourceBundle *localeBundle = ures_open(NULL, locale.getName(), &status);
        if (U_SUCCESS(status)) {
            UResourceBundle *contextTransforms = ures_getByKeyWithFallback(localeBundle, "contextTransforms", NULL, &status);
            if (U_SUCCESS(status)) {
                UResourceBundle *contextTransformUsage;
                while ( (contextTransformUsage = ures_getNextResource(contextTransforms, NULL, &status)) != NULL ) {
                    const int32_t * intVector = ures_getIntVector(contextTransformUsage, &len, &status);
                    if (U_SUCCESS(status) && intVector != NULL && len >= 2) {
                        const char* usageKey = ures_getKey(contextTransformUsage);
                        if (usageKey != NULL) {
                            const ContextUsageNameToEnum * typeMapPtr = contextUsageTypeMap;
                            int32_t compResult = 0;
                            // linear search; list is short and we cannot be sure that bsearch is available
                            while ( typeMapPtr->usageName != NULL && (compResult = uprv_strcmp(usageKey, typeMapPtr->usageName)) > 0 ) {
                                ++typeMapPtr;
                            }
                            if (typeMapPtr->usageName != NULL && compResult == 0) {
                                int32_t titlecaseInt = (capitalizationContext == UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU)? intVector[0]: intVector[1];
                                if (titlecaseInt != 0) {
                                    fCapitalization[typeMapPtr->usageEnum] = TRUE;;
                                    needBrkIter = TRUE;
                                }
                            }
                        }
                    }
                    status = U_ZERO_ERROR;
                    ures_close(contextTransformUsage);
                }
                ures_close(contextTransforms);
            }
            ures_close(localeBundle);
        }
    }
    // Get a sentence break iterator if we will need it
    if (needBrkIter || capitalizationContext == UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE) {
        status = U_ZERO_ERROR;
        capitalizationBrkIter = BreakIterator::createSentenceInstance(locale, status);
        if (U_FAILURE(status)) {
//.........这里部分代码省略.........