C++ UnicodeSet::isEmpty方法代码示例

/**
 * Implement Transliterator framework
 */
void CompoundTransliterator::handleGetSourceSet(UnicodeSet& result) const {
    UnicodeSet set;
    result.clear();
    for (int32_t i=0; i<count; ++i) {
    result.addAll(trans[i]->getSourceSet(set));
    // Take the example of Hiragana-Latin.  This is really
    // Hiragana-Katakana; Katakana-Latin.  The source set of
    // these two is roughly [:Hiragana:] and [:Katakana:].
    // But the source set for the entire transliterator is
    // actually [:Hiragana:] ONLY -- that is, the first
    // non-empty source set.

    // This is a heuristic, and not 100% reliable.
    if (!result.isEmpty()) {
        break;
    }
    }
}

UBool AlphabeticIndex::addChineseIndexCharacters(UErrorCode &errorCode) {
    UnicodeSet contractions;
    collatorPrimaryOnly_->internalAddContractions(BASE[0], contractions, errorCode);
    if (U_FAILURE(errorCode) || contractions.isEmpty()) { return FALSE; }
    initialLabels_->addAll(contractions);
    UnicodeSetIterator iter(contractions);
    while (iter.next()) {
        const UnicodeString &s = iter.getString();
        U_ASSERT (s.startsWith(BASE, BASE_LENGTH));
        UChar c = s.charAt(s.length() - 1);
        if (0x41 <= c && c <= 0x5A) {  // A-Z
            // There are Pinyin labels, add ASCII A-Z labels as well.
            initialLabels_->add(0x41, 0x5A);  // A-Z
            break;
        }
    }
    return TRUE;
}

void TransliteratorErrorTest::TestUnicodeSetErrors() {
    UnicodeString badPattern="[[:L:]-[0x0300-0x0400]";
    UnicodeSet set;
    UErrorCode status = U_ZERO_ERROR;
    UnicodeString result;

    if (!set.isEmpty()) {
        errln("FAIL: The default ctor of UnicodeSet created a non-empty object.");
    }
    set.applyPattern(badPattern, status);
    if (U_SUCCESS(status)) {
        errln("FAIL: Applied a bad pattern to the UnicodeSet object okay.");
    }
    status = U_ZERO_ERROR;
    UnicodeSet *set1 = new UnicodeSet(badPattern, status);
    if (U_SUCCESS(status)) {
        errln("FAIL: Created a UnicodeSet based on bad patterns.");
    }
    delete set1;
}

void
PreparsedUCD::parseScriptExtensions(const char *s, UnicodeSet &scx, UErrorCode &errorCode) {
    if(U_FAILURE(errorCode)) { return; }
    scx.clear();
    CharString scString;
    for(;;) {
        const char *scs;
        const char *scLimit=strchr(s, ' ');
        if(scLimit!=NULL) {
            scs=scString.clear().append(s, (int32_t)(scLimit-s), errorCode).data();
            if(U_FAILURE(errorCode)) { return; }
        } else {
            scs=s;
        }
        int32_t script=pnames->getPropertyValueEnum(UCHAR_SCRIPT, scs);
        if(script==UCHAR_INVALID_CODE) {
            fprintf(stderr,
                    "error in preparsed UCD: '%s' is not a valid script code on line %ld\n",
                    scs, (long)lineNumber);
            errorCode=U_PARSE_ERROR;
            return;
        } else if(scx.contains(script)) {
            fprintf(stderr,
                    "error in preparsed UCD: scx has duplicate '%s' codes on line %ld\n",
                    scs, (long)lineNumber);
            errorCode=U_PARSE_ERROR;
            return;
        } else {
            scx.add(script);
        }
        if(scLimit!=NULL) {
            s=scLimit+1;
        } else {
            break;
        }
    }
    if(scx.isEmpty()) {
        fprintf(stderr, "error in preparsed UCD: empty scx= on line %ld\n", (long)lineNumber);
        errorCode=U_PARSE_ERROR;
    }
}

UVector *AlphabeticIndex::firstStringsInScript(UErrorCode &status) {
    if (U_FAILURE(status)) {
        return NULL;
    }
    LocalPointer<UVector> dest(new UVector(status), status);
    if (U_FAILURE(status)) {
        return NULL;
    }
    dest->setDeleter(uprv_deleteUObject);
    // Fetch the script-first-primary contractions which are defined in the root collator.
    // They all start with U+FDD1.
    UnicodeSet set;
    collatorPrimaryOnly_->internalAddContractions(0xFDD1, set, status);
    if (U_FAILURE(status)) {
        return NULL;
    }
    if (set.isEmpty()) {
        status = U_UNSUPPORTED_ERROR;
        return NULL;
    }
    UnicodeSetIterator iter(set);
    while (iter.next()) {
        const UnicodeString &boundary = iter.getString();
        uint32_t gcMask = U_GET_GC_MASK(boundary.char32At(1));
        if ((gcMask & (U_GC_L_MASK | U_GC_CN_MASK)) == 0) {
            // Ignore boundaries for the special reordering groups.
            // Take only those for "real scripts" (where the sample character is a Letter,
            // and the one for unassigned implicit weights (Cn).
            continue;
        }
        UnicodeString *s = new UnicodeString(boundary);
        if (s == NULL) {
            status = U_MEMORY_ALLOCATION_ERROR;
            return NULL;
        }
        dest->addElement(s, status);
    }
    return dest.orphan();
}

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;
    }

//.........这里部分代码省略.........

int32_t
CollationDataWriter::write(UBool isBase, const UVersionInfo dataVersion,
                           const CollationData &data, const CollationSettings &settings,
                           const void *rootElements, int32_t rootElementsLength,
                           int32_t indexes[], uint8_t *dest, int32_t capacity,
                           UErrorCode &errorCode) {
    if(U_FAILURE(errorCode)) { return 0; }
    if(capacity < 0 || (capacity > 0 && dest == NULL)) {
        errorCode = U_ILLEGAL_ARGUMENT_ERROR;
        return 0;
    }

    // Figure out which data items to write before settling on
    // the indexes length and writing offsets.
    // For any data item, we need to write the start and limit offsets,
    // so the indexes length must be at least index-of-start-offset + 2.
    int32_t indexesLength;
    UBool hasMappings;
    UnicodeSet unsafeBackwardSet;
    const CollationData *baseData = data.base;

    int32_t fastLatinVersion;
    if(data.fastLatinTable != NULL) {
        fastLatinVersion = (int32_t)CollationFastLatin::VERSION << 16;
    } else {
        fastLatinVersion = 0;
    }
    int32_t fastLatinTableLength = 0;

    if(isBase) {
        // For the root collator, we write an even number of indexes
        // so that we start with an 8-aligned offset.
        indexesLength = CollationDataReader::IX_TOTAL_SIZE + 1;
        U_ASSERT(settings.reorderCodesLength == 0);
        hasMappings = TRUE;
        unsafeBackwardSet = *data.unsafeBackwardSet;
        fastLatinTableLength = data.fastLatinTableLength;
    } else if(baseData == NULL) {
        hasMappings = FALSE;
        if(settings.reorderCodesLength == 0) {
            // only options
            indexesLength = CollationDataReader::IX_OPTIONS + 1;  // no limit offset here
        } else {
            // only options, reorder codes, and the reorder table
            indexesLength = CollationDataReader::IX_REORDER_TABLE_OFFSET + 2;
        }
    } else {
        hasMappings = TRUE;
        // Tailored mappings, and what else?
        // Check in ascending order of optional tailoring data items.
        indexesLength = CollationDataReader::IX_CE32S_OFFSET + 2;
        if(data.contextsLength != 0) {
            indexesLength = CollationDataReader::IX_CONTEXTS_OFFSET + 2;
        }
        unsafeBackwardSet.addAll(*data.unsafeBackwardSet).removeAll(*baseData->unsafeBackwardSet);
        if(!unsafeBackwardSet.isEmpty()) {
            indexesLength = CollationDataReader::IX_UNSAFE_BWD_OFFSET + 2;
        }
        if(data.fastLatinTable != baseData->fastLatinTable) {
            fastLatinTableLength = data.fastLatinTableLength;
            indexesLength = CollationDataReader::IX_FAST_LATIN_TABLE_OFFSET + 2;
        }
    }

    UVector32 codesAndRanges(errorCode);
    const int32_t *reorderCodes = settings.reorderCodes;
    int32_t reorderCodesLength = settings.reorderCodesLength;
    if(settings.hasReordering() &&
            CollationSettings::reorderTableHasSplitBytes(settings.reorderTable)) {
        // Rebuild the full list of reorder ranges.
        // The list in the settings is truncated for efficiency.
        data.makeReorderRanges(reorderCodes, reorderCodesLength, codesAndRanges, errorCode);
        // Write the codes, then the ranges.
        for(int32_t i = 0; i < reorderCodesLength; ++i) {
            codesAndRanges.insertElementAt(reorderCodes[i], i, errorCode);
        }
        if(U_FAILURE(errorCode)) { return 0; }
        reorderCodes = codesAndRanges.getBuffer();
        reorderCodesLength = codesAndRanges.size();
    }

    int32_t headerSize;
    if(isBase) {
        headerSize = 0;  // udata_create() writes the header
    } else {
        DataHeader header;
        header.dataHeader.magic1 = 0xda;
        header.dataHeader.magic2 = 0x27;
        uprv_memcpy(&header.info, &dataInfo, sizeof(UDataInfo));
        uprv_memcpy(header.info.dataVersion, dataVersion, sizeof(UVersionInfo));
        headerSize = (int32_t)sizeof(header);
        U_ASSERT((headerSize & 3) == 0);  // multiple of 4 bytes
        if(hasMappings && data.cesLength != 0) {
            // Sum of the sizes of the data items which are
            // not automatically multiples of 8 bytes and which are placed before the CEs.
            int32_t sum = headerSize + (indexesLength + reorderCodesLength) * 4;
            if((sum & 7) != 0) {
                // We need to add padding somewhere so that the 64-bit CEs are 8-aligned.
                // We add to the header size here.
                // Alternatively, we could increment the indexesLength
//.........这里部分代码省略.........