htmlparser.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

            insertNode(e);
            handled = true;
        }
    }

    // 3. If we couldn't handle the error, just return false and attempt to error-correct again.
    if (!handled) {
        reportError(IgnoredContentError, &localName, &m_current->localName());
        return false;
    }
    return insertNode(n);
}

typedef bool (HTMLParser::*CreateErrorCheckFunc)(Token* t, RefPtr<Node>&);
typedef HashMap<AtomicStringImpl*, CreateErrorCheckFunc> FunctionMap;

bool HTMLParser::textCreateErrorCheck(Token* t, RefPtr<Node>& result)
{
    result = new Text(m_document, t->text.get());
    return false;
}

bool HTMLParser::commentCreateErrorCheck(Token* t, RefPtr<Node>& result)
{
    result = new Comment(m_document, t->text.get());
    return false;
}

bool HTMLParser::headCreateErrorCheck(Token*, RefPtr<Node>& result)
{
    if (!m_head || m_current->localName() == htmlTag) {
        m_head = new HTMLHeadElement(headTag, m_document);
        result = m_head;
    } else
        reportError(MisplacedHeadError);
    return false;
}

bool HTMLParser::bodyCreateErrorCheck(Token*, RefPtr<Node>&)
{
    // body no longer allowed if we have a frameset
    if (m_haveFrameSet)
        return false;
    popBlock(headTag);
    startBody();
    return true;
}

bool HTMLParser::framesetCreateErrorCheck(Token*, RefPtr<Node>&)
{
    popBlock(headTag);
    if (m_inBody && !m_haveFrameSet && !m_haveContent) {
        popBlock(bodyTag);
        // ### actually for IE document.body returns the now hidden "body" element
        // we can't implement that behaviour now because it could cause too many
        // regressions and the headaches are not worth the work as long as there is
        // no site actually relying on that detail (Dirk)
        if (m_document->body())
            m_document->body()->setAttribute(styleAttr, "display:none");
        m_inBody = false;
    }
    if ((m_haveContent || m_haveFrameSet) && m_current->localName() == htmlTag)
        return false;
    m_haveFrameSet = true;
    startBody();
    return true;
}

bool HTMLParser::formCreateErrorCheck(Token* t, RefPtr<Node>& result)
{
    // Only create a new form if we're not already inside one.
    // This is consistent with other browsers' behavior.
    if (!m_currentFormElement) {
        m_currentFormElement = new HTMLFormElement(formTag, m_document);
        result = m_currentFormElement;
        pCloserCreateErrorCheck(t, result);
    }
    return false;
}

bool HTMLParser::isindexCreateErrorCheck(Token* t, RefPtr<Node>& result)
{
    RefPtr<Node> n = handleIsindex(t);
    if (!m_inBody)
        m_isindexElement = n.release();
    else {
        t->selfClosingTag = true;
        result = n.release();
    }
    return false;
}

bool HTMLParser::selectCreateErrorCheck(Token*, RefPtr<Node>&)
{
    return true;
}

bool HTMLParser::ddCreateErrorCheck(Token* t, RefPtr<Node>& result)
{
    pCloserCreateErrorCheck(t, result);
    popBlock(dtTag);
    popBlock(ddTag);
    return true;
}

bool HTMLParser::dtCreateErrorCheck(Token* t, RefPtr<Node>& result)
{
    pCloserCreateErrorCheck(t, result);
    popBlock(ddTag);
    popBlock(dtTag);
    return true;
}

bool HTMLParser::nestedCreateErrorCheck(Token* t, RefPtr<Node>&)
{
    popBlock(t->tagName);
    return true;
}

bool HTMLParser::nestedPCloserCreateErrorCheck(Token* t, RefPtr<Node>& result)
{
    pCloserCreateErrorCheck(t, result);
    popBlock(t->tagName);
    return true;
}

bool HTMLParser::nestedStyleCreateErrorCheck(Token* t, RefPtr<Node>&)
{
    return allowNestedRedundantTag(t->tagName);
}

bool HTMLParser::tableCellCreateErrorCheck(Token*, RefPtr<Node>&)
{
    popBlock(tdTag);
    popBlock(thTag);
    return true;
}

bool HTMLParser::tableSectionCreateErrorCheck(Token*, RefPtr<Node>&)
{
    popBlock(theadTag);
    popBlock(tbodyTag);
    popBlock(tfootTag);
    return true;
}

bool HTMLParser::noembedCreateErrorCheck(Token*, RefPtr<Node>&)
{
    setSkipMode(noembedTag);
    return true;
}

bool HTMLParser::noframesCreateErrorCheck(Token*, RefPtr<Node>&)
{
    setSkipMode(noframesTag);
    return true;
}

bool HTMLParser::noscriptCreateErrorCheck(Token*, RefPtr<Node>&)
{
    if (!m_isParsingFragment) {
        Settings* settings = m_document->settings();
        if (settings && settings->isJavaScriptEnabled())
            setSkipMode(noscriptTag);
    }
    return true;
}

bool HTMLParser::pCloserCreateErrorCheck(Token*, RefPtr<Node>&)
{
    if (hasPElementInScope())
        popBlock(pTag);
    return true;
}

bool HTMLParser::pCloserStrictCreateErrorCheck(Token*, RefPtr<Node>&)
{
    if (m_document->inCompatMode())
        return true;
    if (hasPElementInScope())
        popBlock(pTag);
    return true;
}

bool HTMLParser::mapCreateErrorCheck(Token*, RefPtr<Node>& result)
{
    m_currentMapElement = new HTMLMapElement(mapTag, m_document);
    result = m_currentMapElement;
    return false;
}

PassRefPtr<Node> HTMLParser::getNode(Token* t)
{
    // Init our error handling table.
    DEFINE_STATIC_LOCAL(FunctionMap, gFunctionMap, ());
    if (gFunctionMap.isEmpty()) {
        gFunctionMap.set(aTag.localName().impl(), &HTMLParser::nestedCreateErrorCheck);
        gFunctionMap.set(addressTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(bTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(bigTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(blockquoteTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(bodyTag.localName().impl(), &HTMLParser::bodyCreateErrorCheck);
        gFunctionMap.set(buttonTag.localName().impl(), &HTMLParser::nestedCreateErrorCheck);
        gFunctionMap.set(centerTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(commentAtom.impl(), &HTMLParser::commentCreateErrorCheck);
        gFunctionMap.set(ddTag.localName().impl(), &HTMLParser::ddCreateErrorCheck);
        gFunctionMap.set(dirTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(divTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(dlTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(dtTag.localName().impl(), &HTMLParser::dtCreateErrorCheck);
        gFunctionMap.set(formTag.localName().impl(), &HTMLParser::formCreateErrorCheck);
        gFunctionMap.set(fieldsetTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(framesetTag.localName().impl(), &HTMLParser::framesetCreateErrorCheck);
        gFunctionMap.set(h1Tag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(h2Tag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(h3Tag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(h4Tag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(h5Tag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(h6Tag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(headTag.localName().impl(), &HTMLParser::headCreateErrorCheck);
        gFunctionMap.set(hrTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(iTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(isindexTag.localName().impl(), &HTMLParser::isindexCreateErrorCheck);
        gFunctionMap.set(liTag.localName().impl(), &HTMLParser::nestedPCloserCreateErrorCheck);
        gFunctionMap.set(listingTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(mapTag.localName().impl(), &HTMLParser::mapCreateErrorCheck);
        gFunctionMap.set(menuTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(nobrTag.localName().impl(), &HTMLParser::nestedCreateErrorCheck);
        gFunctionMap.set(noembedTag.localName().impl(), &HTMLParser::noembedCreateErrorCheck);
        gFunctionMap.set(noframesTag.localName().impl(), &HTMLParser::noframesCreateErrorCheck);
        gFunctionMap.set(noscriptTag.localName().impl(), &HTMLParser::noscriptCreateErrorCheck);
        gFunctionMap.set(olTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(pTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(plaintextTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(preTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
        gFunctionMap.set(sTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(selectTag.localName().impl(), &HTMLParser::selectCreateErrorCheck);
        gFunctionMap.set(smallTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(strikeTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(tableTag.localName().impl(), &HTMLParser::pCloserStrictCreateErrorCheck);
        gFunctionMap.set(tbodyTag.localName().impl(), &HTMLParser::tableSectionCreateErrorCheck);
        gFunctionMap.set(tdTag.localName().impl(), &HTMLParser::tableCellCreateErrorCheck);
        gFunctionMap.set(textAtom.impl(), &HTMLParser::textCreateErrorCheck);
        gFunctionMap.set(tfootTag.localName().impl(), &HTMLParser::tableSectionCreateErrorCheck);
        gFunctionMap.set(thTag.localName().impl(), &HTMLParser::tableCellCreateErrorCheck);
        gFunctionMap.set(theadTag.localName().impl(), &HTMLParser::tableSectionCreateErrorCheck);
        gFunctionMap.set(trTag.localName().impl(), &HTMLParser::nestedCreateErrorCheck);
        gFunctionMap.set(ttTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(uTag.localName().impl(), &HTMLParser::nestedStyleCreateErrorCheck);
        gFunctionMap.set(ulTag.localName().impl(), &HTMLParser::pCloserCreateErrorCheck);
    }

    bool proceed = true;
    RefPtr<Node> result;
    if (CreateErrorCheckFunc errorCheckFunc = gFunctionMap.get(t->tagName.impl()))
        proceed = (this->*errorCheckFunc)(t, result);
    if (proceed)
        result = HTMLElementFactory::createHTMLElement(QualifiedName(nullAtom, t->tagName, xhtmlNamespaceURI), m_document, m_currentFormElement.get());
    return result.release();
}

bool HTMLParser::allowNestedRedundantTag(const AtomicString& tagName)
{
    // www.liceo.edu.mx is an example of a site that achieves a level of nesting of
    // about 1500 tags, all from a bunch of <b>s.  We will only allow at most 20
    // nested tags of the same type before just ignoring them all together.
    unsigned i = 0;
    for (HTMLStackElem* curr = m_blockStack;
         i < cMaxRedundantTagDepth && curr && curr->tagName == tagName;
         curr = curr->next, i++) { }
    return i != cMaxRedundantTagDepth;
}

void HTMLParser::processCloseTag(Token* t)
{
    // Support for really broken html.
    // we never close the body tag, since some stupid web pages close it before the actual end of the doc.
    // let's rely on the end() call to close things.
    if (t->tagName == htmlTag || t->tagName == bodyTag || t->tagName == commentAtom)
        return;
    
    bool checkForCloseTagErrors = true;
    if (t->tagName == formTag && m_currentFormElement) {
        m_currentFormElement = 0;
        checkForCloseTagErrors = false;
    } else if (t->tagName == mapTag)
        m_currentMapElement = 0;
    else if (t->tagName == pTag)
        checkForCloseTagErrors = false;
        
    HTMLStackElem* oldElem = m_blockStack;
    popBlock(t->tagName, checkForCloseTagErrors);
    if (oldElem == m_blockStack && t->tagName == pTag) {
        // We encountered a stray </p>.  Amazingly Gecko, WinIE, and MacIE all treat
        // this as a valid break, i.e., <p></p>.  So go ahead and make the empty
        // paragraph.
        t->beginTag = true;
        parseToken(t);
        popBlock(t->tagName);
        reportError(StrayParagraphCloseError);
    }
}

bool HTMLParser::isHeaderTag(const AtomicString& tagName)
{
    DEFINE_STATIC_LOCAL(HashSet<AtomicStringImpl*>, headerTags, ());
    if (headerTags.isEmpty()) {
        headerTags.add(h1Tag.localName().impl());
        headerTags.add(h2Tag.localName().impl());
        headerTags.add(h3Tag.localName().impl());
        headerTags.add(h4Tag.localName().impl());
        headerTags.add(h5Tag.localName().impl());
        headerTags.add(h6Tag.localName().impl());
    }
    
    return headerTags.contains(tagName.impl());
}

bool HTMLParser::isInline(Node* node) const
{
    if (node->isTextNode())
        return true;