xmltokenizerlibxml2.cpp

linux下开源浏览器WebKit的源码,市面上的很多商用浏览器都是移植自WebKit

void setLoaderForLibXMLCallbacks(DocLoader* docLoader)
{
    globalDocLoader = docLoader;
}

static bool didInit = false;

static xmlParserCtxtPtr createStringParser(xmlSAXHandlerPtr handlers, void* userData)
{
    if (!didInit) {
        xmlInitParser();
        xmlRegisterInputCallbacks(matchFunc, openFunc, readFunc, closeFunc);
        xmlRegisterOutputCallbacks(matchFunc, openFunc, writeFunc, closeFunc);
        libxmlLoaderThread = currentThread();
        didInit = true;
    }

    xmlParserCtxtPtr parser = xmlCreatePushParserCtxt(handlers, 0, 0, 0, 0);
    parser->_private = userData;
    parser->replaceEntities = true;
    const UChar BOM = 0xFEFF;
    const unsigned char BOMHighByte = *reinterpret_cast<const unsigned char*>(&BOM);
    xmlSwitchEncoding(parser, BOMHighByte == 0xFF ? XML_CHAR_ENCODING_UTF16LE : XML_CHAR_ENCODING_UTF16BE);

    return parser;
}


// Chunk should be encoded in UTF-8
static xmlParserCtxtPtr createMemoryParser(xmlSAXHandlerPtr handlers, void* userData, const char* chunk)
{
    if (!didInit) {
        xmlInitParser();
        xmlRegisterInputCallbacks(matchFunc, openFunc, readFunc, closeFunc);
        xmlRegisterOutputCallbacks(matchFunc, openFunc, writeFunc, closeFunc);
        libxmlLoaderThread = currentThread();
        didInit = true;
    }

    xmlParserCtxtPtr parser = xmlCreateMemoryParserCtxt(chunk, xmlStrlen((const xmlChar*)chunk));

    if (!parser)
        return 0;

    // Copy the sax handler
    memcpy(parser->sax, handlers, sizeof(xmlSAXHandler));

    // Set parser options.
    // XML_PARSE_NODICT: default dictionary option.
    // XML_PARSE_NOENT: force entities substitutions.
    xmlCtxtUseOptions(parser, XML_PARSE_NODICT | XML_PARSE_NOENT);

    // Internal initialization
    parser->sax2 = 1;
    parser->instate = XML_PARSER_CONTENT; // We are parsing a CONTENT
    parser->depth = 0;
    parser->str_xml = xmlDictLookup(parser->dict, BAD_CAST "xml", 3);
    parser->str_xmlns = xmlDictLookup(parser->dict, BAD_CAST "xmlns", 5);
    parser->str_xml_ns = xmlDictLookup(parser->dict, XML_XML_NAMESPACE, 36);
    parser->_private = userData;

    return parser;
}

// --------------------------------

XMLTokenizer::XMLTokenizer(Document* _doc, FrameView* _view)
    : m_doc(_doc)
    , m_view(_view)
    , m_context(0)
    , m_pendingCallbacks(new PendingCallbacks)
    , m_currentNode(_doc)
    , m_currentNodeIsReferenced(false)
    , m_sawError(false)
    , m_sawXSLTransform(false)
    , m_sawFirstElement(false)
    , m_isXHTMLDocument(false)
    , m_parserPaused(false)
    , m_requestingScript(false)
    , m_finishCalled(false)
    , m_errorCount(0)
    , m_lastErrorLine(0)
    , m_lastErrorColumn(0)
    , m_pendingScript(0)
    , m_scriptStartLine(0)
    , m_parsingFragment(false)
{
}

XMLTokenizer::XMLTokenizer(DocumentFragment* fragment, Element* parentElement)
    : m_doc(fragment->document())
    , m_view(0)
    , m_context(0)
    , m_pendingCallbacks(new PendingCallbacks)
    , m_currentNode(fragment)
    , m_currentNodeIsReferenced(fragment)
    , m_sawError(false)
    , m_sawXSLTransform(false)
    , m_sawFirstElement(false)
    , m_isXHTMLDocument(false)
    , m_parserPaused(false)
    , m_requestingScript(false)
    , m_finishCalled(false)
    , m_errorCount(0)
    , m_lastErrorLine(0)
    , m_lastErrorColumn(0)
    , m_pendingScript(0)
    , m_scriptStartLine(0)
    , m_parsingFragment(true)
{
    if (fragment)
        fragment->ref();
    if (m_doc)
        m_doc->ref();
          
    // Add namespaces based on the parent node
    Vector<Element*> elemStack;
    while (parentElement) {
        elemStack.append(parentElement);
        
        Node* n = parentElement->parentNode();
        if (!n || !n->isElementNode())
            break;
        parentElement = static_cast<Element*>(n);
    }
    
    if (elemStack.isEmpty())
        return;
    
    for (Element* element = elemStack.last(); !elemStack.isEmpty(); elemStack.removeLast()) {
        if (NamedAttrMap* attrs = element->attributes()) {
            for (unsigned i = 0; i < attrs->length(); i++) {
                Attribute* attr = attrs->attributeItem(i);
                if (attr->localName() == "xmlns")
                    m_defaultNamespaceURI = attr->value();
                else if (attr->prefix() == "xmlns")
                    m_prefixToNamespaceMap.set(attr->localName(), attr->value());
            }
        }
    }

    // If the parent element is not in document tree, there may be no xmlns attribute; just default to the parent's namespace.
    if (m_defaultNamespaceURI.isNull() && !parentElement->inDocument())
        m_defaultNamespaceURI = parentElement->namespaceURI();
}

XMLTokenizer::~XMLTokenizer()
{
    setCurrentNode(0);
    if (m_parsingFragment && m_doc)
        m_doc->deref();
    if (m_pendingScript)
        m_pendingScript->removeClient(this);
    if (m_context)
        xmlFreeParserCtxt(m_context);
}

void XMLTokenizer::doWrite(const String& parseString)
{
    if (!m_context)
        initializeParserContext();
    
    // libXML throws an error if you try to switch the encoding for an empty string.
    if (parseString.length()) {
        // Hack around libxml2's lack of encoding overide support by manually
        // resetting the encoding to UTF-16 before every chunk.  Otherwise libxml
        // will detect <?xml version="1.0" encoding="<encoding name>"?> blocks 
        // and switch encodings, causing the parse to fail.
        const UChar BOM = 0xFEFF;
        const unsigned char BOMHighByte = *reinterpret_cast<const unsigned char*>(&BOM);
        xmlSwitchEncoding(m_context, BOMHighByte == 0xFF ? XML_CHAR_ENCODING_UTF16LE : XML_CHAR_ENCODING_UTF16BE);

        xmlParseChunk(m_context, reinterpret_cast<const char*>(parseString.characters()), sizeof(UChar) * parseString.length(), 0);
    }
    
    if (m_doc->decoder() && m_doc->decoder()->sawError()) {
        // If the decoder saw an error, report it as fatal (stops parsing)
        handleError(fatal, "Encoding error", lineNumber(), columnNumber());
    }

    return;
}

static inline String toString(const xmlChar* str, unsigned len)
{
    return UTF8Encoding().decode(reinterpret_cast<const char*>(str), len);
}

static inline String toString(const xmlChar* str)
{
    if (!str)
        return String();
    
    return UTF8Encoding().decode(reinterpret_cast<const char*>(str), strlen(reinterpret_cast<const char*>(str)));
}

struct _xmlSAX2Namespace {
    const xmlChar* prefix;
    const xmlChar* uri;
};
typedef struct _xmlSAX2Namespace xmlSAX2Namespace;

static inline void handleElementNamespaces(Element* newElement, const xmlChar** libxmlNamespaces, int nb_namespaces, ExceptionCode& ec)
{
    xmlSAX2Namespace* namespaces = reinterpret_cast<xmlSAX2Namespace*>(libxmlNamespaces);
    for(int i = 0; i < nb_namespaces; i++) {
        String namespaceQName = "xmlns";
        String namespaceURI = toString(namespaces[i].uri);
        if (namespaces[i].prefix)
            namespaceQName = "xmlns:" + toString(namespaces[i].prefix);
        newElement->setAttributeNS("http://www.w3.org/2000/xmlns/", namespaceQName, namespaceURI, ec);
        if (ec) // exception setting attributes
            return;
    }
}

struct _xmlSAX2Attributes {
    const xmlChar* localname;
    const xmlChar* prefix;
    const xmlChar* uri;
    const xmlChar* value;
    const xmlChar* end;
};
typedef struct _xmlSAX2Attributes xmlSAX2Attributes;

static inline void handleElementAttributes(Element* newElement, const xmlChar** libxmlAttributes, int nb_attributes, ExceptionCode& ec)
{
    xmlSAX2Attributes* attributes = reinterpret_cast<xmlSAX2Attributes*>(libxmlAttributes);
    for(int i = 0; i < nb_attributes; i++) {
        String attrLocalName = toString(attributes[i].localname);
        int valueLength = (int) (attributes[i].end - attributes[i].value);
        String attrValue = toString(attributes[i].value, valueLength);
        String attrPrefix = toString(attributes[i].prefix);
        String attrURI = attrPrefix.isEmpty() ? String() : toString(attributes[i].uri);
        String attrQName = attrPrefix.isEmpty() ? attrLocalName : attrPrefix + ":" + attrLocalName;
        
        newElement->setAttributeNS(attrURI, attrQName, attrValue, ec);
        if (ec) // exception setting attributes
            return;
    }
}

void XMLTokenizer::startElementNs(const xmlChar* xmlLocalName, const xmlChar* xmlPrefix, const xmlChar* xmlURI, int nb_namespaces,
                                  const xmlChar** libxmlNamespaces, int nb_attributes, int nb_defaulted, const xmlChar** libxmlAttributes)
{
    if (m_parserStopped)
        return;
    
    if (m_parserPaused) {
        m_pendingCallbacks->appendStartElementNSCallback(xmlLocalName, xmlPrefix, xmlURI, nb_namespaces, libxmlNamespaces,
                                                         nb_attributes, nb_defaulted, libxmlAttributes);
        return;
    }

    bool isFirstElement = !m_sawFirstElement;
    m_sawFirstElement = true;

    exitText();

    String localName = toString(xmlLocalName);
    String uri = toString(xmlURI);
    String prefix = toString(xmlPrefix);

    if (m_parsingFragment && uri.isNull()) {
        if (!prefix.isNull())
            uri = m_prefixToNamespaceMap.get(prefix);
        else
            uri = m_defaultNamespaceURI;
    }

    QualifiedName qName(prefix, localName, uri);
    RefPtr<Element> newElement = m_doc->createElement(qName, true);
    if (!newElement) {
        stopParsing();
        return;
    }
    
    ExceptionCode ec = 0;
    handleElementNamespaces(newElement.get(), libxmlNamespaces, nb_namespaces, ec);
    if (ec) {
        stopParsing();
        return;
    }

    ScriptController* jsProxy = m_doc->frame() ? m_doc->frame()->script() : 0;
    if (jsProxy && m_doc->frame()->script()->isEnabled())
        jsProxy->setEventHandlerLineno(lineNumber());

    handleElementAttributes(newElement.get(), libxmlAttributes, nb_attributes, ec);
    if (ec) {
        stopParsing();
        return;
    }

    if (jsProxy)
        jsProxy->setEventHandlerLineno(0);

    newElement->beginParsingChildren();

    ScriptElement* scriptElement = toScriptElement(newElement.get());
    if (scriptElement)
        m_scriptStartLine = lineNumber();

    if (!m_currentNode->addChild(newElement.get())) {
        stopParsing();
        return;
    }

    setCurrentNode(newElement.get());
    if (m_view && !newElement->attached())
        newElement->attach();

    if (isFirstElement && m_doc->frame())
        m_doc->frame()->loader()->dispatchDocumentElementAvailable();
}

void XMLTokenizer::endElementNs()
{
    if (m_parserStopped)
        return;

    if (m_parserPaused) {
        m_pendingCallbacks->appendEndElementNSCallback();
        return;
    }
    
    exitText();

    Node* n = m_currentNode;
    RefPtr<Node> parent = n->parentNode();
    n->finishParsingChildren();

    if (!n->isElementNode() || !m_view) {
        setCurrentNode(parent.get());
        return;
    }

    Element* element = static_cast<Element*>(n);
    ScriptElement* scriptElement = toScriptElement(element);
    if (!scriptElement) {
        setCurrentNode(parent.get());
        return;
    }

    // don't load external scripts for standalone documents (for now)
    ASSERT(!m_pendingScript);
    m_requestingScript = true;

    String scriptHref = scriptElement->sourceAttributeValue();
    if (!scriptHref.isEmpty()) {
        // we have a src attribute 
        String scriptCharset = scriptElement->scriptCharset();
        if ((m_pendingScript = m_doc->docLoader()->requestScript(scriptHref, scriptCharset))) {
            m_scriptElement = element;
            m_pendingScript->addClient(this);

            // m_pendingScript will be 0 if script was already loaded and ref() executed it
            if (m_pendingScript)
                pauseParsing();
        } else 
            m_scriptElement = 0;
    } else
        m_view->frame()->loader()->executeScript(ScriptSourceCode(scriptElement->scriptContent(), m_doc->url(), m_scriptStartLine));

    m_requestingScript = false;
    setCurrentNode(parent.get());
}

void XMLTokenizer::characters(const xmlChar* s, int len)
{
    if (m_parserStopped)
        return;
    
    if (m_parserPaused) {
        m_pendingCallbacks->appendCharactersCallback(s, len);
        return;
    }
    
    if (m_currentNode->isTextNode() || enterText())
        m_bufferedText.append(s, len);
}

void XMLTokenizer::error(ErrorType type, const char* message, va_list args)
{
    if (m_parserStopped)
        return;

#if PLATFORM(WIN_OS)
    char m[1024];
    vsnprintf(m, sizeof(m) - 1, message, args);
#else
    char* m;
    vasprintf(&m, message, args);
#endif
    
    if (m_parserPaused)
        m_pendingCallbacks->appendErrorCallback(type, m, lineNumber(), columnNumber());
    else
        handleError(type, m, lineNumber(), columnNumber());

#if !PLATFORM(WIN_OS)
    free(m);
#endif
}

void XMLTokenizer::processingInstruction(const xmlChar* target, const xmlChar* data)
{
    if (m_parserStopped)
        return;

    if (m_parserPaused) {
        m_pendingCallbacks->appendProcessingInstructionCallback(target, data);
        return;
    }
    
    exitText();

    // ### handle exceptions
    int exception = 0;
    RefPtr<ProcessingInstruction> pi = m_doc->createProcessingInstruction(
        toString(target), toString(data), exception);
    if (exception)
        return;

    pi->setCreatedByParser(true);
    
    if (!m_currentNode->addChild(pi.get()))
        return;
    if (m_view && !pi->attached())
        pi->attach();

    pi->finishParsingChildren();

#if ENABLE(XSLT)
    m_sawXSLTransform = !m_sawFirstElement && pi->isXSL();
    if (m_sawXSLTransform && !m_doc->transformSourceDocument())
        stopParsing();
#endif
}

void XMLTokenizer::cdataBlock(const xmlChar* s, int len)
{
    if (m_parserStopped)
        return;

    if (m_parserPaused) {
        m_pendingCallbacks->appendCDATABlockCallback(s, len);
        return;
    }
    
    exitText();

    RefPtr<Node> newNode = new CDATASection(m_doc, toString(s, len));
    if (!m_currentNode->addChild(newNode.get()))
        return;