htmlparser.cpp

将konqueror浏览器移植到ARM9 2410中

/*
    This file is part of the KDE libraries

    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])

    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., 59 Temple Place - Suite 330,
    Boston, MA 02111-1307, USA.
*/
//----------------------------------------------------------------------------
//
// KDE HTML Widget -- HTML Parser
//#define PARSER_DEBUG

#include "htmlparser.h"

#include "dom_nodeimpl.h"
#include "dom_exception.h"
#include "html_baseimpl.h"
#include "html_blockimpl.h"
#include "html_documentimpl.h"
#include "html_elementimpl.h"
#include "html_formimpl.h"
#include "html_headimpl.h"
#include "html_imageimpl.h"
#include "html_inlineimpl.h"
#include "html_listimpl.h"
#include "html_miscimpl.h"
#include "html_tableimpl.h"
#include "html_objectimpl.h"
#include "dom_textimpl.h"
#include "htmlhashes.h"
#include "htmltokenizer.h"
#include "khtmlview.h"
#include "khtml_part.h"
#include "cssproperties.h"
#include "cssvalues.h"

#include "rendering/render_object.h"

#include <kdebug.h>
#include <klocale.h>

using namespace DOM;
using namespace khtml;

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

/**
 * @internal
 */
class HTMLStackElem
{
public:
    HTMLStackElem( int _id,
                   int _level,
                   DOM::NodeImpl *_node,
                   HTMLStackElem * _next
        )
        :
        id(_id),
        level(_level),
        node(_node),
        next(_next)
        { }

    int       id;
    int       level;
    NodeImpl *node;
    HTMLStackElem *next;
};

/**
 * @internal
 *
 * The parser parses tokenized input into the document, building up the
 * document tree. If the document is wellformed, parsing it is
 * straightforward.
 * Unfortunately, people can't write wellformed HTML documents, so the parser
 * has to be tolerant about errors.
 *
 * We have to take care of 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 inbetween (or that the
 *    tag inbetween is optional...) This could be the case with the following
 *    tags: HTML HEAD BODY TBODY TR TD LI (did I forget any?)
 * 3. We wan't 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.
 *
 */
KHTMLParser::KHTMLParser( KHTMLView *_parent, DocumentPtr *doc)
{
    //kdDebug( 6035 ) << "parser constructor" << endl;
#if SPEED_DEBUG > 0
    qt.start();
#endif

    HTMLWidget    = _parent;
    document      = doc;
    document->ref();

    blockStack = 0;

    // ID_CLOSE_TAG == Num of tags
    forbiddenTag = new ushort[ID_CLOSE_TAG+1];

    reset();
}

KHTMLParser::KHTMLParser( DOM::DocumentFragmentImpl *i, DocumentPtr *doc )
{
    HTMLWidget = 0;
    document = doc;
    document->ref();

    forbiddenTag = new ushort[ID_CLOSE_TAG+1];

    blockStack = 0;

    reset();
    current = i;
    inBody = true;
    inSelect = false;
}

KHTMLParser::~KHTMLParser()
{
#if SPEED_DEBUG > 0
    kdDebug( ) << "TIME: parsing time was = " << qt.elapsed() << endl;
#endif


    document->deref();

    freeBlock();

    delete [] forbiddenTag;
    delete isindex;
}

void KHTMLParser::reset()
{
    current = document->document();

    freeBlock();

    // before parsing no tags are forbidden...
    memset(forbiddenTag, 0, (ID_CLOSE_TAG+1)*sizeof(ushort));

    inBody = false;
    noRealBody = true;
    haveFrameSet = false;
    inSelect = false;
    _inline = false;

    form = 0;
    map = 0;
    head = 0;
    end = false;
    isindex = 0;
    flat = false;
    haveKonqBlock = false;

    discard_until = 0;
}

void KHTMLParser::parseToken(Token *t)
{
    if (t->id > 2*ID_CLOSE_TAG)
    {
      kdDebug( 6035 ) << "Unknown tag!! tagID = " << t->id << endl;
      return;
    }
    if(discard_until) {
        if(t->id == discard_until)
            discard_until = 0;

        // do not skip </iframe>
        if ( discard_until || current->id() + ID_CLOSE_TAG != t->id )
            return;
    }

#ifdef PARSER_DEBUG
    kdDebug( 6035 ) << "\n\n==> parser: processing token " << getTagName(t->id).string() << "(" << t->id << ")"
                    << " current = " << getTagName(current->id()).string() << "(" << current->id() << ")" << endl;
    kdDebug(6035) << "inline=" << _inline << " inBody=" << inBody << " noRealBody=" << noRealBody << " haveFrameSet=" << haveFrameSet << endl;
#endif

    // holy shit. apparently some sites use </br> instead of <br>
    // be compatible with IE and NS
    if(t->id == ID_BR+ID_CLOSE_TAG && document->document()->parseMode() != DocumentImpl::Strict)
        t->id -= ID_CLOSE_TAG;

    if(t->id > ID_CLOSE_TAG)
    {
        processCloseTag(t);
        return;
    }

    // ignore spaces, if we're not inside a paragraph or other inline code
    if( t->id == ID_TEXT ) {
#ifdef PARSER_DEBUG
        if(t->text)
            kdDebug(6035) << "length="<< t->text->l << " text='" << QConstString(t->text->s, t->text->l).string() << "'" << endl;
#endif

        if ( inBody ) noRealBody = false;
    }

    NodeImpl *n = getElement(t);
    // just to be sure, and to catch currently unimplemented stuff
    if(!n)
        return;

    // set attributes
    if(n->isElementNode())
    {
        ElementImpl *e = static_cast<ElementImpl *>(n);
        e->setAttributeMap(t->attrs);

        // take care of optional close tags
        if(endTag[e->id()] == DOM::OPTIONAL)
            popBlock(t->id);
    }

    // if this tag is forbidden inside the current context, pop
    // blocks until we are allowed to add it...
    while(forbiddenTag[t->id]) popOneBlock();

    if ( !insertNode(n) ) {
        // we couldn't insert the node...
#ifdef PARSER_DEBUG
        kdDebug( 6035 ) << "insertNode failed current=" << current->id() << ", new=" << n->id() << "!" << endl;
#endif
        if (map == n)
        {
#ifdef PARSER_DEBUG
            kdDebug( 6035 ) << "  --> resetting map!" << endl;
#endif
            map = 0;
        }
        if (form == n)
        {
#ifdef PARSER_DEBUG
            kdDebug( 6035 ) << "   --> resetting form!" << endl;
#endif
            form = 0;
        }
        delete n;
    }
}

bool KHTMLParser::insertNode(NodeImpl *n)
{
    int id = n->id();

    // let's be stupid and just try to insert it.
    // this should work if the document is wellformed
#ifdef PARSER_DEBUG
    NodeImpl *tmp = current;
#endif
    NodeImpl *newNode = current->addChild(n);
    if ( newNode ) {
#ifdef PARSER_DEBUG
        kdDebug( 6035 ) << "added " << n->nodeName().string() << " to " << tmp->nodeName().string() << ", new current=" << newNode->nodeName().string() << endl;
#endif
        // don't push elements without end tag on the stack
        if(tagPriority[id] != 0 && !flat)
        {
            pushBlock(id, tagPriority[id]);
            current = newNode;
#if SPEED_DEBUG < 2
            if(!n->attached() && HTMLWidget )  n->attach();
#endif
            //_inline = current->isInline();
            if(current->isInline()) _inline = true;
        }
        else {
#if SPEED_DEBUG < 2
            if(!n->attached() && HTMLWidget)  n->attach();
	    if(n->renderer()) n->renderer()->close();
#endif
            flat = false;
        }

#if SPEED_DEBUG < 1
        if(tagPriority[id] == 0 && n->renderer())
            n->renderer()->calcMinMaxWidth();
#endif
        return true;
    } else {
#ifdef PARSER_DEBUG
        kdDebug( 6035 ) << "ADDING NODE FAILED!!!! current = " << current->nodeName().string() << ", new = " << n->nodeName().string() << endl;
#endif
        // error handling...
        HTMLElementImpl *e;
        bool handled = false;

        // switch according to the element to insert
        switch(id)
        {
        case ID_COMMENT:
            break;
        case ID_HEAD:
            // ### alllow not having <HTML> in at all, as per HTML spec
            if (!current->isDocumentNode() && !current->id() == ID_HTML )
                return false;
            break;
            // We can deal with a base, meta and link element in the body, by just adding the element to head.
        case ID_META:
        case ID_LINK:
        case ID_BASE:
            if( !head )
                createHead();
            if( head ) {
                head->addChild(n);
#if SPEED_DEBUG < 2
                if(!n->attached() && HTMLWidget)
                    n->attach();
#endif
                return true;
            }
            break;
        case ID_HTML:
            if (!current->isDocumentNode())
                return false;
            break;
        case ID_TITLE:
        case ID_STYLE:
            if ( !head )
                createHead();
            if ( head ) {
                DOM::NodeImpl *newNode = head->addChild(n);
                if ( newNode ) {
                    pushBlock(id, tagPriority[id]);
                    current = newNode;
#if SPEED_DEBUG < 2
                    if(!n->attached() && HTMLWidget)
                        n->attach();
#endif
                } else {
#ifdef PARSER_DEBUG
                    kdDebug( 6035 ) << "adding style before to body failed!!!!" << endl;
#endif
                    discard_until = ID_STYLE + ID_CLOSE_TAG;
                    return false;
                }
                return true;
            } else if(inBody) {
                discard_until = ID_STYLE + ID_CLOSE_TAG;
                return false;
            }
            break;
            // SCRIPT and OBJECT are allowd in the body.
        case ID_BODY:
            if(inBody && doc()->body()) {
                // we have another <BODY> element.... apply attributes to existing one
                // make sure we don't overwrite already existing attributes
                // some sites use <body bgcolor=rightcolor>...<body bgcolor=wrongcolor>
                NamedAttrMapImpl *map = static_cast<NamedAttrMapImpl*>(n->attributes());
                NamedAttrMapImpl *bodymap = static_cast<NamedAttrMapImpl*>(doc()->body()->attributes());
                unsigned long attrNo;
                int exceptioncode;
                bool changed = false;
                for (attrNo = 0; map && attrNo < map->length(); attrNo++)
                    if(!bodymap->getNamedItem(static_cast<AttrImpl*>(map->item(attrNo))->name())) {
                        doc()->body()->setAttributeNode(static_cast<AttrImpl*>(map->item(attrNo)->cloneNode(false,exceptioncode)), exceptioncode);
                        changed = true;
                    }
                if ( changed )
                    doc()->applyChanges();
                noRealBody = false;
            } else if ( current->isDocumentNode() )
                break;
            return false;
            break;

            // the following is a hack to move non rendered elements
            // outside of tables.
            // needed for broken constructs like <table><form ...><tr>....
        case ID_INPUT:
        {
            ElementImpl *e = static_cast<ElementImpl *>(n);
            DOMString type = e->getAttribute(ATTR_TYPE);

            if ( strcasecmp( type, "hidden" ) != 0 )
                break;
            // Fall through!
        }
        case ID_TEXT: