qxmlstream.cpp

奇趣公司比较新的qt/emd版本

  huge, or it is being delivered over a network connection

  \sa tokenType(), tokenString()
 */
QXmlStreamReader::TokenType QXmlStreamReader::readNext()
{
    Q_D(QXmlStreamReader);
    if (d->type != Invalid) {
        if (!d->hasCheckedStartDocument)
            if (!d->checkStartDocument())
                return d->type; // synthetic StartDocument or error
        d->parse();
        if (d->atEnd && d->type != EndDocument && d->type != Invalid)
            d->raiseError(PrematureEndOfDocumentError);
        else if (!d->atEnd && d->type == EndDocument)
            d->raiseWellFormedError(QXmlStream::tr("Extra content at end of document."));
    } else if (d->error == PrematureEndOfDocumentError) {
        // resume error
        d->type = NoToken;
        d->atEnd = false;
        d->token = -1;
        return readNext();
    }
    return d->type;
}


/*!
  Returns the type of the current token.

  The current token can also be queried with the convenience functions
  isStartDocument(), isEndDocument(), isStartElement(),
  isEndElement(), isCharacters(), isComment(), isDTD(),
  isEntityReference(), and isProcessingInstruction()

  \sa tokenString()
 */
QXmlStreamReader::TokenType QXmlStreamReader::tokenType() const
{
    Q_D(const QXmlStreamReader);
    return d->type;
}

static const char * QXmlStreamReader_tokenTypeString[] = {
    "NoToken",
    "Invalid",
    "StartDocument",
    "EndDocument",
    "StartElement",
    "EndElement",
    "Characters",
    "Comment",
    "DTD",
    "EntityReference",
    "ProcessingInstruction"
};


/*!
    \property  QXmlStreamReader::namespaceProcessing
    the namespace-processing flag of the stream reader

    This property controls whether or not the stream reader processes
    namespaces. If enabled, the reader processes namespaces, otherwise
    it does not.

    By default, namespace-processing is enabled.
*/


void QXmlStreamReader::setNamespaceProcessing(bool enable)
{
    Q_D(QXmlStreamReader);
    d->namespaceProcessing = enable;
}

bool QXmlStreamReader::namespaceProcessing() const
{
    Q_D(const QXmlStreamReader);
    return d->namespaceProcessing;
}

/*! Returns the reader's current token as string.

\sa tokenType()
*/
QString QXmlStreamReader::tokenString() const
{
    Q_D(const QXmlStreamReader);
    return QLatin1String(QXmlStreamReader_tokenTypeString[d->type]);
}


QXmlStreamPrivateTagStack::QXmlStreamPrivateTagStack()
{
    tagStack.reserve(16);
    tagStackStringStorage.reserve(32);
    tagStackStringStorageSize = 0;
    NamespaceDeclaration &namespaceDeclaration = namespaceDeclarations.push();
    namespaceDeclaration.prefix = addToStringStorage(QLatin1String("xml"));
    namespaceDeclaration.namespaceUri = addToStringStorage(QLatin1String("http://www.w3.org/XML/1998/namespace"));
    tagStackDefaultStringStorageSize = tagStackStringStorageSize;
}

QXmlStreamReaderPrivate::QXmlStreamReaderPrivate(QXmlStreamReader *q)
    :q_ptr(q)
{
    device = 0;
    deleteDevice = false;
#ifndef QT_NO_TEXTCODEC
    decoder = 0;
#endif
    stack_size = 64;
    sym_stack = 0;
    state_stack = 0;
    reallocateStack();
    init();
    entityHash.insert(QLatin1String("lt"), Entity::createLiteral(QLatin1String("<")));
    entityHash.insert(QLatin1String("gt"), Entity::createLiteral(QLatin1String(">")));
    entityHash.insert(QLatin1String("amp"), Entity::createLiteral(QLatin1String("&")));
    entityHash.insert(QLatin1String("apos"), Entity::createLiteral(QLatin1String("'")));
    entityHash.insert(QLatin1String("quot"), Entity::createLiteral(QLatin1String("\"")));
}

void QXmlStreamReaderPrivate::init()
{
    tos = 0;
    scanDtd = false;
    token = -1;
    token_char = 0;
    isEmptyElement = false;
    isWhitespace = false;
    isCDATA = false;
    standalone = false;
    tos = 0;
    resumeReduction = 0;
    state_stack[tos++] = 0;
    state_stack[tos] = 0;
    putStack.clear();
    putStack.reserve(32);
    textBuffer.clear();
    textBuffer.reserve(256);
    initTagStack();
    tagsDone = false;
    attributes.clear();
    attributes.reserve(16);
    lineNumber = lastLineStart = characterOffset = 0;
    readBufferPos = 0;
    nbytesread = 0;
#ifndef QT_NO_TEXTCODEC
    codec = QTextCodec::codecForMib(106); // utf8
    delete decoder;
    decoder = 0;
#endif
    attributeStack.clear();
    attributeStack.reserve(16);
    entityParser = 0;
    hasCheckedStartDocument = false;
    normalizeLiterals = false;
    hasSeenTag = false;
    atEnd = false;
    inParseEntity = false;
    referenceToUnparsedEntityDetected = false;
    referenceToParameterEntityDetected = false;
    hasExternalDtdSubset = false;
    lockEncoding = false;
    namespaceProcessing = true;
    rawReadBuffer.clear();
    dataBuffer.clear();
    readBuffer.clear();

    type = QXmlStreamReader::NoToken;
    error = QXmlStreamReader::NoError;
}

/*
  Well-formed requires that we verify entity values. We do this with a
  standard parser.
 */
void QXmlStreamReaderPrivate::parseEntity(const QString &value)
{
    Q_Q(QXmlStreamReader);

    if (value.isEmpty())
        return;


    if (!entityParser)
        entityParser = new QXmlStreamReaderPrivate(q);
    else
        entityParser->init();
    entityParser->inParseEntity = true;
    entityParser->readBuffer = value;
    entityParser->injectToken(PARSE_ENTITY);
    while (!entityParser->atEnd && entityParser->type != QXmlStreamReader::Invalid)
        entityParser->parse();
    if (entityParser->type == QXmlStreamReader::Invalid || entityParser->tagStack.size())
        raiseWellFormedError(QXmlStream::tr("Invalid entity value."));

}

inline void QXmlStreamReaderPrivate::reallocateStack()
{
    stack_size <<= 1;
    sym_stack = reinterpret_cast<Value*> (qRealloc(sym_stack, stack_size * sizeof(Value)));
    state_stack = reinterpret_cast<int*> (qRealloc(state_stack, stack_size * sizeof(int)));
}


QXmlStreamReaderPrivate::~QXmlStreamReaderPrivate()
{
#ifndef QT_NO_TEXTCODEC
    delete decoder;
#endif
    qFree(sym_stack);
    qFree(state_stack);
    delete entityParser;
}


inline uint QXmlStreamReaderPrivate::filterCarriageReturn()
{
    uint peekc = peekChar();
    if (peekc == '\n') {
        if (putStack.size())
            putStack.pop();
        else
            ++readBufferPos;
        return peekc;
    }
    if (peekc == 0) {
        putChar('\r');
        return 0;
    }
    return '\n';
}

/*!
 \internal
 If the end of the file is encountered, 0 is returned.
 */
inline uint QXmlStreamReaderPrivate::getChar()
{
    uint c;
    if (putStack.size()) {
        c = atEnd ? 0 : putStack.pop();
    } else {
        if (readBufferPos < readBuffer.size())
            c = readBuffer.at(readBufferPos++).unicode();
        else
            c = getChar_helper();
    }

    return c;
}

inline uint QXmlStreamReaderPrivate::peekChar()
{
    uint c;
    if (putStack.size()) {
        c = putStack.top();
    } else if (readBufferPos < readBuffer.size()) {
        c = readBuffer.at(readBufferPos).unicode();
    } else {
        if ((c = getChar_helper()))
            --readBufferPos;
    }

    return c;
}

/*!
  \internal

  Scans characters until \a str is encountered, and validates the characters
  as according to the Char[2] production and do the line-ending normalization.
  If any character is invalid, false is returned, otherwise true upon success.

  If \a tokenToInject is not less than zero, injectToken() is called with
  \a tokenToInject when \a str is found.

  If any error occurred, false is returned, otherwise true.
  */
bool QXmlStreamReaderPrivate::scanUntil(const char *str, short tokenToInject)
{
    int pos = textBuffer.size();
    int oldLineNumber = lineNumber;

    while (uint c = getChar()) {
        /* First, we do the validation & normalization. */
        switch (c) {
        case '\r':
            if ((c = filterCarriageReturn()) == 0)
                break;
            // fall through
        case '\n':
            ++lineNumber;
            lastLineStart = characterOffset + readBufferPos;
            // fall through
        case '\t':
            textBuffer.inline_append(c);
            continue;
        default:
            if(c < 0x20 || (c > 0xFFFD && c < 0x10000) || c > 0x10FFFF ) {
                raiseWellFormedError(QXmlStream::tr("Invalid XML character."));
                lineNumber = oldLineNumber;
                return false;
            }
            textBuffer.inline_append(c);
        }


        /* Second, attempt to lookup str. */
        if (c == uint(*str)) {
            if (!*(str + 1)) {
                if (tokenToInject >= 0)
                    injectToken(tokenToInject);
                return true;
            } else {
                if (scanString(str + 1, tokenToInject, false))
                    return true;
            }
        }
    }
    putString(textBuffer, pos);
    textBuffer.resize(pos);
    lineNumber = oldLineNumber;
    return false;
}

bool QXmlStreamReaderPrivate::scanString(const char *str, short tokenToInject, bool requireSpace)
{
    int n = 0;
    while (str[n]) {
        ushort c = getChar();
        if (c != ushort(str[n])) {
            if (c)
                putChar(c);
            while (n--) {
                putChar(ushort(str[n]));
            }
            return false;
        }
        ++n;
    }
    for (int i = 0; i < n; ++i)
        textBuffer.inline_append(ushort(str[i]));
    if (requireSpace) {
        int s = fastScanSpace();
        if (!s || atEnd) {
            int pos = textBuffer.size() - n - s;
            putString(textBuffer, pos);
            textBuffer.resize(pos);
            return false;
        }
    }
    if (tokenToInject >= 0)
        injectToken(tokenToInject);
    return true;
}

bool QXmlStreamReaderPrivate::scanAfterLangleBang()
{
    switch (peekChar()) {
    case '[':
        return scanString(spell[CDATA_START], CDATA_START, false);
    case 'D':
        return scanString(spell[DOCTYPE], DOCTYPE);
    case 'A':
        return scanString(spell[ATTLIST], ATTLIST);
    case 'N':
        return scanString(spell[NOTATION], NOTATION);
    case 'E':
        if (scanString(spell[ELEMENT], ELEMENT))
            return true;
        return scanString(spell[ENTITY], ENTITY);

    default:
        ;
    };
    return false;
}

bool QXmlStreamReaderPrivate::scanPublicOrSystem()
{
    switch (peekChar()) {
    case 'S':
        return scanString(spell[SYSTEM], SYSTEM);
    case 'P':
        return scanString(spell[PUBLIC], PUBLIC);
    default:
        ;
    }
    return false;
}

bool QXmlStreamReaderPrivate::scanNData()
{
    if (fastScanSpace()) {
        if (scanString(spell[NDATA], NDATA))
            return true;
        putChar(' ');
    }
    return false;
}

bool QXmlStreamReaderPrivate::scanAfterDefaultDecl()
{
    switch (peekChar()) {
    case 'R':
        return scanString(spell[REQUIRED], REQUIRED, false);
    case 'I':
        return scanString(spell[IMPLIED], IMPLIED, false);
    case 'F':
        return scanString(spell[FIXED], FIXED, false);
    default:
        ;
    }
    return false;
}

bool QXmlStreamReaderPrivate::scanAttType()
{
    switch (peekChar()) {
    case 'C':
        return scanString(spell[CDATA], CDATA);
    case 'I':
        if (scanString(spell[ID], ID))
            return true;
        if (scanString(spell[IDREF], IDREF))
            return true;
        return scanString(spell[IDREFS], IDREFS);
    case 'E':
        if (scanString(spell[ENTITY], ENTITY))
            return true;
        return scanString(spell[ENTITIES], ENTITIES);
    case 'N':
        if (scanString(spell[NOTATION], NOTATION))