C++ UnicodeString::getCapacity方法代码示例

/* {{{ intl_stringFromChar */
int intl_stringFromChar(UnicodeString &ret, char *str, size_t str_len, UErrorCode *status)
{
	if(str_len > INT32_MAX) {
		*status = U_BUFFER_OVERFLOW_ERROR;
		ret.setToBogus();
		return FAILURE;
	}
	//the number of UTF-16 code units is not larger than that of UTF-8 code
	//units, + 1 for the terminator
	int32_t capacity = (int32_t)str_len + 1;

	//no check necessary -- if NULL will fail ahead
	UChar	*utf16 = ret.getBuffer(capacity);
	int32_t utf16_len = 0;
	*status = U_ZERO_ERROR;
	u_strFromUTF8WithSub(utf16, ret.getCapacity(), &utf16_len,
		str, str_len, U_SENTINEL /* no substitution */, NULL,
		status);
	ret.releaseBuffer(utf16_len);
	if (U_FAILURE(*status)) {
		ret.setToBogus();
		return FAILURE;
	}
	return SUCCESS;
}

void
PreparsedUCD::parseString(const char *s, UnicodeString &uni, UErrorCode &errorCode) {
    UChar *buffer=uni.getBuffer(-1);
    int32_t length=u_parseString(s, buffer, uni.getCapacity(), NULL, &errorCode);
    if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
        errorCode=U_ZERO_ERROR;
        uni.releaseBuffer(0);
        buffer=uni.getBuffer(length);
        length=u_parseString(s, buffer, uni.getCapacity(), NULL, &errorCode);
    }
    uni.releaseBuffer(length);
    if(U_FAILURE(errorCode)) {
        fprintf(stderr,
                "error in preparsed UCD: '%s' is not a valid Unicode string on line %ld\n",
                s, (long)lineNumber);
    }
}

void
StringCaseTest::TestCasingImpl(const UnicodeString &input,
                               const UnicodeString &output,
                               int32_t whichCase,
                               void *iter, const char *localeID, uint32_t options) {
    // UnicodeString
    UnicodeString result;
    const char *name;
    Locale locale(localeID);

    result=input;
    switch(whichCase) {
    case TEST_LOWER:
        name="toLower";
        result.toLower(locale);
        break;
    case TEST_UPPER:
        name="toUpper";
        result.toUpper(locale);
        break;
#if !UCONFIG_NO_BREAK_ITERATION
    case TEST_TITLE:
        name="toTitle";
        result.toTitle((BreakIterator *)iter, locale, options);
        break;
#endif
    case TEST_FOLD:
        name="foldCase";
        result.foldCase(options);
        break;
    default:
        name="";
        break; // won't happen
    }
    if(result!=output) {
        dataerrln("error: UnicodeString.%s() got a wrong result for a test case from casing.res", name);
    }
#if !UCONFIG_NO_BREAK_ITERATION
    if(whichCase==TEST_TITLE && options==0) {
        result=input;
        result.toTitle((BreakIterator *)iter, locale);
        if(result!=output) {
            dataerrln("error: UnicodeString.toTitle(options=0) got a wrong result for a test case from casing.res");
        }
    }
#endif

    // UTF-8
    char utf8In[100], utf8Out[100];
    int32_t utf8InLength, utf8OutLength, resultLength;
    UChar *buffer;

    IcuTestErrorCode errorCode(*this, "TestCasingImpl");
    LocalUCaseMapPointer csm(ucasemap_open(localeID, options, errorCode));
#if !UCONFIG_NO_BREAK_ITERATION
    if(iter!=NULL) {
        // Clone the break iterator so that the UCaseMap can safely adopt it.
        UBreakIterator *clone=ubrk_safeClone((UBreakIterator *)iter, NULL, NULL, errorCode);
        ucasemap_setBreakIterator(csm.getAlias(), clone, errorCode);
    }
#endif

    u_strToUTF8(utf8In, (int32_t)sizeof(utf8In), &utf8InLength, input.getBuffer(), input.length(), errorCode);
    switch(whichCase) {
    case TEST_LOWER:
        name="ucasemap_utf8ToLower";
        utf8OutLength=ucasemap_utf8ToLower(csm.getAlias(),
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, errorCode);
        break;
    case TEST_UPPER:
        name="ucasemap_utf8ToUpper";
        utf8OutLength=ucasemap_utf8ToUpper(csm.getAlias(),
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, errorCode);
        break;
#if !UCONFIG_NO_BREAK_ITERATION
    case TEST_TITLE:
        name="ucasemap_utf8ToTitle";
        utf8OutLength=ucasemap_utf8ToTitle(csm.getAlias(),
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, errorCode);
        break;
#endif
    case TEST_FOLD:
        name="ucasemap_utf8FoldCase";
        utf8OutLength=ucasemap_utf8FoldCase(csm.getAlias(),
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, errorCode);
        break;
    default:
        name="";
        utf8OutLength=0;
        break; // won't happen
    }
    buffer=result.getBuffer(utf8OutLength);
    u_strFromUTF8(buffer, result.getCapacity(), &resultLength, utf8Out, utf8OutLength, errorCode);
    result.releaseBuffer(errorCode.isSuccess() ? resultLength : 0);

    if(errorCode.isFailure()) {
//.........这里部分代码省略.........

//.........这里部分代码省略.........
    result=input;
    switch(whichCase) {
    case TEST_LOWER:
        name="toLower";
        result.toLower(locale);
        break;
    case TEST_UPPER:
        name="toUpper";
        result.toUpper(locale);
        break;
#if !UCONFIG_NO_BREAK_ITERATION
    case TEST_TITLE:
        name="toTitle";
        result.toTitle((BreakIterator *)iter, locale, options);
        break;
#endif
    case TEST_FOLD:
        name="foldCase";
        result.foldCase(options);
        break;
    default:
        name="";
        break; // won't happen
    }
    if(result!=output) {
        errln("error: UnicodeString.%s() got a wrong result for a test case from casing.res", name);
    }
#if !UCONFIG_NO_BREAK_ITERATION
    if(whichCase==TEST_TITLE && options==0) {
        result=input;
        result.toTitle((BreakIterator *)iter, locale);
        if(result!=output) {
            errln("error: UnicodeString.toTitle(options=0) got a wrong result for a test case from casing.res");
        }
    }
#endif

    // UTF-8
    char utf8In[100], utf8Out[100];
    int32_t utf8InLength, utf8OutLength, resultLength;
    UChar *buffer;

    UCaseMap *csm;
    UErrorCode errorCode;

    errorCode=U_ZERO_ERROR;
    csm=ucasemap_open(localeID, options, &errorCode);
#if !UCONFIG_NO_BREAK_ITERATION
    if(iter!=NULL) {
        // Clone the break iterator so that the UCaseMap can safely adopt it.
        int32_t size=1;  // Not 0 because that only gives preflighting.
        UBreakIterator *clone=ubrk_safeClone((UBreakIterator *)iter, NULL, &size, &errorCode);
        ucasemap_setBreakIterator(csm, clone, &errorCode);
    }
#endif

    u_strToUTF8(utf8In, (int32_t)sizeof(utf8In), &utf8InLength, input.getBuffer(), input.length(), &errorCode);
    switch(whichCase) {
    case TEST_LOWER:
        name="ucasemap_utf8ToLower";
        utf8OutLength=ucasemap_utf8ToLower(csm,
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, &errorCode);
        break;
    case TEST_UPPER:
        name="ucasemap_utf8ToUpper";
        utf8OutLength=ucasemap_utf8ToUpper(csm,
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, &errorCode);
        break;
#if !UCONFIG_NO_BREAK_ITERATION
    case TEST_TITLE:
        name="ucasemap_utf8ToTitle";
        utf8OutLength=ucasemap_utf8ToTitle(csm,
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, &errorCode);
        break;
#endif
    case TEST_FOLD:
        name="ucasemap_utf8FoldCase";
        utf8OutLength=ucasemap_utf8FoldCase(csm,
                    utf8Out, (int32_t)sizeof(utf8Out),
                    utf8In, utf8InLength, &errorCode);
        break;
    default:
        name="";
        utf8OutLength=0;
        break; // won't happen
    }
    buffer=result.getBuffer(utf8OutLength);
    u_strFromUTF8(buffer, result.getCapacity(), &resultLength, utf8Out, utf8OutLength, &errorCode);
    result.releaseBuffer(U_SUCCESS(errorCode) ? resultLength : 0);

    if(U_FAILURE(errorCode)) {
        errln("error: %s() got an error for a test case from casing.res - %s", name, u_errorName(errorCode));
    } else if(result!=output) {
        errln("error: %s() got a wrong result for a test case from casing.res", name);
    }
    ucasemap_close(csm);
}

UXMLElement *
UXMLParser::parseFile(const char *filename, UErrorCode &errorCode) {
    char bytes[4096], charsetBuffer[100];
    FileStream *f;
    const char *charset, *pb;
    UnicodeString src;
    UConverter *cnv;
    UChar *buffer, *pu;
    int32_t fileLength, bytesLength, length, capacity;
    UBool flush;

    if(U_FAILURE(errorCode)) {
        return NULL;
    }

    f=T_FileStream_open(filename, "rb");
    if(f==NULL) {
        errorCode=U_FILE_ACCESS_ERROR;
        return NULL;
    }

    bytesLength=T_FileStream_read(f, bytes, (int32_t)sizeof(bytes));
    if(bytesLength<(int32_t)sizeof(bytes)) {
        // we have already read the entire file
        fileLength=bytesLength;
    } else {
        // get the file length
        fileLength=T_FileStream_size(f);
    }

    /*
     * get the charset:
     * 1. Unicode signature
     * 2. treat as ISO-8859-1 and read XML encoding="charser"
     * 3. default to UTF-8
     */
    charset=ucnv_detectUnicodeSignature(bytes, bytesLength, NULL, &errorCode);
    if(U_SUCCESS(errorCode) && charset!=NULL) {
        // open converter according to Unicode signature
        cnv=ucnv_open(charset, &errorCode);
    } else {
        // read as Latin-1 and parse the XML declaration and encoding
        cnv=ucnv_open("ISO-8859-1", &errorCode);
        if(U_FAILURE(errorCode)) {
            // unexpected error opening Latin-1 converter
            goto exit;
        }

        buffer=src.getBuffer(bytesLength);
        if(buffer==NULL) {
            // unexpected failure to reserve some string capacity
            errorCode=U_MEMORY_ALLOCATION_ERROR;
            goto exit;
        }
        pb=bytes;
        pu=buffer;
        ucnv_toUnicode(
            cnv,
            &pu, buffer+src.getCapacity(),
            &pb, bytes+bytesLength,
            NULL, TRUE, &errorCode);
        src.releaseBuffer(U_SUCCESS(errorCode) ? (int32_t)(pu-buffer) : 0);
        ucnv_close(cnv);
        cnv=NULL;
        if(U_FAILURE(errorCode)) {
            // unexpected error in conversion from Latin-1
            src.remove();
            goto exit;
        }

        // parse XML declaration
        if(mXMLDecl.reset(src).lookingAt(0, errorCode)) {
            int32_t declEnd=mXMLDecl.end(errorCode);
            // go beyond <?xml
            int32_t pos=src.indexOf((UChar)x_l)+1;

            mAttrValue.reset(src);
            while(pos<declEnd && mAttrValue.lookingAt(pos, errorCode)) {  // loop runs once per attribute on this element.
                UnicodeString attName  = mAttrValue.group(1, errorCode);
                UnicodeString attValue = mAttrValue.group(2, errorCode);

                // Trim the quotes from the att value.  These are left over from the original regex
                //   that parsed the attribue, which couldn't conveniently strip them.
                attValue.remove(0,1);                    // one char from the beginning
                attValue.truncate(attValue.length()-1);  // and one from the end.

                if(attName==UNICODE_STRING("encoding", 8)) {
                    length=attValue.extract(0, 0x7fffffff, charsetBuffer, (int32_t)sizeof(charsetBuffer));
                    charset=charsetBuffer;
                    break;
                }
                pos = mAttrValue.end(2, errorCode);
            }

            if(charset==NULL) {
                // default to UTF-8
                charset="UTF-8";
            }
            cnv=ucnv_open(charset, &errorCode);
        }
//.........这里部分代码省略.........

void NormalizationTransliterator::handleTransliterate(Replaceable& text, UTransPosition& offsets,
                                                      UBool isIncremental) const {
    // start and limit of the input range
    int32_t start = offsets.start;
    int32_t limit = offsets.limit;
    int32_t length, delta;

    if(start >= limit) {
        return;
    }

    // a C code unit iterator, implemented around the Replaceable
    UCharIterator iter;
    uiter_setReplaceable(&iter, &text);

    // the output string and buffer pointer
    UnicodeString output;
    UChar *buffer;
    UBool neededToNormalize;

    UErrorCode errorCode;

    /*
     * Normalize as short chunks at a time as possible even in
     * bulk mode, so that styled text is minimally disrupted.
     * In incremental mode, a chunk that ends with offsets.limit
     * must not be normalized.
     *
     * If it was known that the input text is not styled, then
     * a bulk mode normalization could look like this:
     *

    UChar staticChars[256];
    UnicodeString input;

    length = limit - start;
    input.setTo(staticChars, 0, sizeof(staticChars)/U_SIZEOF_UCHAR); // writable alias

    _Replaceable_extractBetween(text, start, limit, input.getBuffer(length));
    input.releaseBuffer(length);

    UErrorCode status = U_ZERO_ERROR;
    Normalizer::normalize(input, fMode, options, output, status);

    text.handleReplaceBetween(start, limit, output);

    int32_t delta = output.length() - length;
    offsets.contextLimit += delta;
    offsets.limit += delta;
    offsets.start = limit + delta;

     *
     */
    while(start < limit) {
        // set the iterator limits for the remaining input range
        // this is a moving target because of the replacements in the text object
        iter.start = iter.index = start;
        iter.limit = limit;

        // incrementally normalize a small chunk of the input
        buffer = output.getBuffer(-1);
        errorCode = U_ZERO_ERROR;
        length = unorm_next(&iter, buffer, output.getCapacity(),
                            fMode, 0,
                            TRUE, &neededToNormalize,
                            &errorCode);
        output.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);

        if(errorCode == U_BUFFER_OVERFLOW_ERROR) {
            // use a larger output string buffer and do it again from the start
            iter.index = start;
            buffer = output.getBuffer(length);
            errorCode = U_ZERO_ERROR;
            length = unorm_next(&iter, buffer, output.getCapacity(),
                                fMode, 0,
                                TRUE, &neededToNormalize,
                                &errorCode);
            output.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
        }

        if(U_FAILURE(errorCode)) {
            break;
        }

        limit = iter.index;
        if(isIncremental && limit == iter.limit) {
            // stop in incremental mode when we reach the input limit
            // in case there are additional characters that could change the
            // normalization result

            // UNLESS all characters in the result of the normalization of
            // the last run are in the skippable set
            const UChar *s=output.getBuffer();
            int32_t i=0, outLength=output.length();
            UChar32 c;

            while(i<outLength) {
                U16_NEXT(s, i, outLength, c);
                if(!unorm_isNFSkippable(c, fMode)) {
                    outLength=-1; // I wish C++ had labeled loops and break outer; ...
//.........这里部分代码省略.........

//.........这里部分代码省略.........
    if(errorCode.logIfFailureAndReset("unable to find the source/test/testdata "
                                      "folder (getSourceTestData())")) {
        return;
    }
    char bidiTestPath[400];
    strcpy(bidiTestPath, sourceTestDataPath);
    strcat(bidiTestPath, "BidiCharacterTest.txt");
    LocalStdioFilePointer bidiTestFile(fopen(bidiTestPath, "r"));
    if(bidiTestFile.isNull()) {
        errln("unable to open %s", bidiTestPath);
        return;
    }
    LocalUBiDiPointer ubidi(ubidi_open());
    lineNumber=0;
    levelsCount=0;
    orderingCount=0;
    errorCount=0;
    while(errorCount<20 && fgets(line, (int)sizeof(line), bidiTestFile.getAlias())!=NULL) {
        ++lineNumber;
        paraLevelName="N/A";
        inputString="N/A";
        // Remove trailing comments and whitespace.
        char *commentStart=strchr(line, '#');
        if(commentStart!=NULL) {
            *commentStart=0;
        }
        u_rtrim(line);
        const char *start=u_skipWhitespace(line);
        if(*start==0) {
            continue;  // Skip empty and comment-only lines.
        }
        // Parse the code point string in field 0.
        UChar *buffer=inputString.getBuffer(200);
        int32_t length=u_parseString(start, buffer, inputString.getCapacity(), NULL, errorCode);
        if(errorCode.logIfFailureAndReset("Invalid string in field 0")) {
            errln("Input line %d: %s", (int)lineNumber, line);
            inputString.remove();
            continue;
        }
        inputString.releaseBuffer(length);
        start=strchr(start, ';');
        if(start==NULL) {
            errorCount++;
            errln("\nError on line %d: Missing ; separator on line: %s", (int)lineNumber, line);
            continue;
        }
        start=u_skipWhitespace(start+1);
        char *end;
        int32_t paraDirection=(int32_t)strtol(start, &end, 10);
        UBiDiLevel paraLevel=UBIDI_MAX_EXPLICIT_LEVEL+2;
        if(paraDirection==0) {
            paraLevel=0;
            paraLevelName="LTR";
        }
        else if(paraDirection==1) {
            paraLevel=1;
            paraLevelName="RTL";
        }
        else if(paraDirection==2) {
            paraLevel=UBIDI_DEFAULT_LTR;
            paraLevelName="Auto/LTR";
        }
        else if(paraDirection==3) {
            paraLevel=UBIDI_DEFAULT_RTL;
            paraLevelName="Auto/RTL";
        }