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

/*
    Copyright (C) 1997 Martin Jones ([e-mail])
              (C) 1997 Torben Weis ([e-mail])
              (C) 1999,2001 Lars Knoll ([e-mail])
              (C) 2000,2001 Dirk Mueller ([e-mail])
    Copyright (C) 2004, 2005, 2006, 2007 Apple Inc. All rights reserved.

    This library is free software; you can redistribute it and/or
    modify it under the terms of the GNU Library General Public
    License as published by the Free Software Foundation; either
    version 2 of the License, or (at your option) any later version.

    This library is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    Library General Public License for more details.

    You should have received a copy of the GNU Library General Public License
    along with this library; see the file COPYING.LIB.  If not, write to
    the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
    Boston, MA 02110-1301, USA.
*/

#include "config.h"
#include "HTMLParser.h"

#include "CharacterNames.h"
#include "CSSPropertyNames.h"
#include "CSSValueKeywords.h"
#include "Comment.h"
#include "Console.h"
#include "DOMWindow.h"
#include "DocumentFragment.h"
#include "DocumentType.h"
#include "Frame.h"
#include "HTMLBodyElement.h"
#include "HTMLDocument.h"
#include "HTMLDivElement.h"
#include "HTMLDListElement.h"
#include "HTMLElementFactory.h"
#include "HTMLFormElement.h"
#include "HTMLHeadElement.h"
#include "HTMLHRElement.h"
#include "HTMLHtmlElement.h"
#include "HTMLIsIndexElement.h"
#include "HTMLMapElement.h"
#include "HTMLNames.h"
#include "HTMLTableCellElement.h"
#include "HTMLTableRowElement.h"
#include "HTMLTableSectionElement.h"
#include "HTMLTokenizer.h"
#include "LocalizedStrings.h"
#include "Settings.h"
#include "Text.h"
#include <wtf/StdLibExtras.h>
    
namespace WebCore {

using namespace HTMLNames;

static const unsigned cMaxRedundantTagDepth = 20;
static const unsigned cResidualStyleMaxDepth = 200;

struct HTMLStackElem : Noncopyable {
    HTMLStackElem(const AtomicString& t, int lvl, Node* n, bool r, HTMLStackElem* nx)
        : tagName(t)
        , level(lvl)
        , strayTableContent(false)
        , node(n)
        , didRefNode(r)
        , next(nx)
    {
    }

    void derefNode()
    {
        if (didRefNode)
            node->deref();
    }

    AtomicString tagName;
    int level;
    bool strayTableContent;
    Node* node;
    bool didRefNode;
    HTMLStackElem* next;
};

/**
 * The parser parses tokenized input into the document, building up the
 * document tree. If the document is well-formed, parsing it is straightforward.
 *
 * Unfortunately, we have to handle many HTML documents that are not well-formed,
 * so the parser has to be tolerant about errors.
 *
 * We have to take care of at least the following error conditions:
 *
 * 1. The element being added is explicitly forbidden inside some outer tag.
 *    In this case we should close all tags up to the one, which forbids
 *    the element, and add it afterwards.
 *
 * 2. We are not allowed to add the element directly. It could be that
 *    the person writing the document forgot some tag in between (or that the
 *    tag in between is optional). This could be the case with the following
 *    tags: HTML HEAD BODY TBODY TR TD LI (did I forget any?).
 *
 * 3. We want to add a block element inside to an inline element. Close all
 *    inline elements up to the next higher block element.
 *
 * 4. If this doesn't help, close elements until we are allowed to add the
 *    element or ignore the tag.
 *
 */

HTMLParser::HTMLParser(HTMLDocument* doc, bool reportErrors)
    : m_document(doc)
    , m_current(doc)
    , m_didRefCurrent(false)
    , m_blockStack(0)
    , m_hasPElementInScope(NotInScope)
    , m_head(0)
    , m_inBody(false)
    , m_haveContent(false)
    , m_haveFrameSet(false)
    , m_isParsingFragment(false)
    , m_reportErrors(reportErrors)
    , m_handlingResidualStyleAcrossBlocks(false)
    , m_inStrayTableContent(0)
{
}

HTMLParser::HTMLParser(DocumentFragment* frag)
    : m_document(frag->document())
    , m_current(frag)
    , m_didRefCurrent(true)
    , m_blockStack(0)
    , m_hasPElementInScope(NotInScope)
    , m_head(0)
    , m_inBody(true)
    , m_haveContent(false)
    , m_haveFrameSet(false)
    , m_isParsingFragment(true)
    , m_reportErrors(false)
    , m_handlingResidualStyleAcrossBlocks(false)
    , m_inStrayTableContent(0)
{
    if (frag)
        frag->ref();
}

HTMLParser::~HTMLParser()
{
    freeBlock();
    if (m_didRefCurrent)
        m_current->deref(); 
}

void HTMLParser::reset()
{
    ASSERT(!m_isParsingFragment);

    setCurrent(m_document);

    freeBlock();

    m_inBody = false;
    m_haveFrameSet = false;
    m_haveContent = false;
    m_inStrayTableContent = 0;

    m_currentFormElement = 0;
    m_currentMapElement = 0;
    m_head = 0;
    m_isindexElement = 0;

    m_skipModeTag = nullAtom;
}

void HTMLParser::setCurrent(Node* newCurrent) 
{
    bool didRefNewCurrent = newCurrent && newCurrent != m_document;
    if (didRefNewCurrent) 
        newCurrent->ref(); 
    if (m_didRefCurrent) 
        m_current->deref();
    m_current = newCurrent;
    m_didRefCurrent = didRefNewCurrent;
}

PassRefPtr<Node> HTMLParser::parseToken(Token* t)
{
    if (!m_skipModeTag.isNull()) {
        if (!t->beginTag && t->tagName == m_skipModeTag)
            // Found the end tag for the current skip mode, so we're done skipping.
            m_skipModeTag = nullAtom;
        else if (m_current->localName() == t->tagName)
            // Do not skip </iframe>.
            // FIXME: What does that comment mean? How can it be right to parse a token without clearing m_skipModeTag?
            ;
        else
            return 0;
    }

    // Apparently some sites use </br> instead of <br>. Be compatible with IE and Firefox and treat this like <br>.
    if (t->isCloseTag(brTag) && m_document->inCompatMode()) {
        reportError(MalformedBRError);
        t->beginTag = true;
    }

    if (!t->beginTag) {
        processCloseTag(t);
        return 0;
    }

    // Ignore spaces, if we're not inside a paragraph or other inline code.
    // Do not alter the text if it is part of a scriptTag.
    if (t->tagName == textAtom && t->text && m_current->localName() != scriptTag) {
        if (m_inBody && !skipMode() && m_current->localName() != styleTag &&
            m_current->localName() != titleTag && !t->text->containsOnlyWhitespace())
            m_haveContent = true;
        
        RefPtr<Node> n;
        String text = t->text.get();
        unsigned charsLeft = text.length();
        while (charsLeft) {
            // split large blocks of text to nodes of manageable size
            n = Text::createWithLengthLimit(m_document, text, charsLeft);
            if (!insertNode(n.get(), t->selfClosingTag))
                return 0;
        }
        return n;
    }

    RefPtr<Node> n = getNode(t);
    // just to be sure, and to catch currently unimplemented stuff
    if (!n)
        return 0;

    // set attributes
    if (n->isHTMLElement()) {
        HTMLElement* e = static_cast<HTMLElement*>(n.get());
        e->setAttributeMap(t->attrs.get());

        // take care of optional close tags
        if (e->endTagRequirement() == TagStatusOptional)
            popBlock(t->tagName);
            
        // If the node does not have a forbidden end tag requirement, and if the broken XML self-closing
        // syntax was used, report an error.
        if (t->brokenXMLStyle && e->endTagRequirement() != TagStatusForbidden) {
            if (t->tagName == scriptTag)
                reportError(IncorrectXMLCloseScriptWarning);
            else
                reportError(IncorrectXMLSelfCloseError, &t->tagName);
        }
    }

    if (!insertNode(n.get(), t->selfClosingTag)) {
        // we couldn't insert the node

        if (n->isElementNode()) {
            Element* e = static_cast<Element*>(n.get());
            e->setAttributeMap(0);
        }

        if (m_currentMapElement == n)
            m_currentMapElement = 0;

        if (m_currentFormElement == n)
            m_currentFormElement = 0;

        if (m_head == n)
            m_head = 0;

        return 0;
    }
    return n;
}

void HTMLParser::parseDoctypeToken(DoctypeToken* t)
{
    // Ignore any doctype after the first.  Ignore doctypes in fragments.
    if (m_document->doctype() || m_isParsingFragment || m_current != m_document)
        return;
        
    // Make a new doctype node and set it as our doctype.
    m_document->addChild(DocumentType::create(m_document, String::adopt(t->m_name), String::adopt(t->m_publicID), String::adopt(t->m_systemID)));
}

static bool isTableSection(const Node* n)
{
    return n->hasTagName(tbodyTag) || n->hasTagName(tfootTag) || n->hasTagName(theadTag);
}

static bool isTablePart(const Node* n)
{
    return n->hasTagName(trTag) || n->hasTagName(tdTag) || n->hasTagName(thTag) ||
           isTableSection(n);
}

static bool isTableRelated(const Node* n)
{
    return n->hasTagName(tableTag) || isTablePart(n);
}

static bool isScopingTag(const AtomicString& tagName)
{
    return tagName == appletTag || tagName == captionTag || tagName == tdTag || tagName == thTag || tagName == buttonTag || tagName == marqueeTag || tagName == objectTag || tagName == tableTag || tagName == htmlTag;
}

bool HTMLParser::insertNode(Node* n, bool flat)
{
    RefPtr<Node> protectNode(n);

    const AtomicString& localName = n->localName();
    int tagPriority = n->isHTMLElement() ? static_cast<HTMLElement*>(n)->tagPriority() : 0;
    
    // <table> is never allowed inside stray table content.  Always pop out of the stray table content
    // and close up the first table, and then start the second table as a sibling.
    if (m_inStrayTableContent && localName == tableTag)
        popBlock(tableTag);