htmlparser.cpp

monqueror一个很具有参考价值的源玛

            break;
        n = new HTMLFrameSetElementImpl(document);
        haveFrameSet = true;
        inBody = true;
        break;
        // a bit a special case, since the frame is inlined...
    case ID_IFRAME:
        n = new HTMLIFrameElementImpl(document);
        break;
		
// form elements
    case ID_FORM:
        // close all open forms...
        popBlock(ID_FORM);
        form = new HTMLFormElementImpl(document);
        n = form;
        break;

    case ID_BUTTON:
        n = new HTMLButtonElementImpl(document, form);
        break;
    case ID_FIELDSET:
        n = new HTMLFieldSetElementImpl(document, form);
        break;

    case ID_INPUT:
        n = new HTMLInputElementImpl(document, form);
        break;
    case ID_OPTION:
        n = new HTMLOptionElementImpl(document, form);
        break;
    case ID_SELECT:
        n = new HTMLSelectElementImpl(document, form);
        break;
    case ID_TEXTAREA:
        n = new HTMLTextAreaElementImpl(document, form);
        break;
    case ID_LABEL:
        n = new HTMLLabelElementImpl(document, form);
        break;
    case ID_LEGEND:
        n = new HTMLLegendElementImpl(document, form);
        break;

    case ID_OPTGROUP:
        n = new HTMLOptGroupElementImpl(document, form);
        break;


// lists
    case ID_DL:
        n = new HTMLDListElementImpl(document);
        break;
    case ID_DD:
        n = new HTMLGenericElementImpl(document, t->id);
        popBlock(ID_DT);
        popBlock(ID_DD);
        break;
    case ID_DT:
        n = new HTMLGenericElementImpl(document, t->id);
        popBlock(ID_DD);
        popBlock(ID_DT);
        break;
    case ID_UL:
    {
        n = new HTMLUListElementImpl(document);
        break;
    }
    case ID_OL:
    {
        n = new HTMLOListElementImpl(document);
        break;
    }
    case ID_DIR:
        n = new HTMLDirectoryElementImpl(document);
        break;
    case ID_MENU:
        n = new HTMLMenuElementImpl(document);
        break;
    case ID_LI:
    {
        popBlock(ID_LI);
	HTMLElementImpl *e = new HTMLLIElementImpl(document);
        n = e;
	if( current->id() != ID_UL && current->id() != ID_OL )
		e->addCSSProperty(CSS_PROP_LIST_STYLE_POSITION, DOMString("inside"), false);
        break;
    }
// formatting elements (block)
    case ID_BLOCKQUOTE:
        n = new HTMLBlockquoteElementImpl(document);
        break;
    case ID_DIV:
        n = new HTMLDivElementImpl(document);
        break;
    case ID_H1:
    case ID_H2:
    case ID_H3:
    case ID_H4:
    case ID_H5:
    case ID_H6:
        n = new HTMLHeadingElementImpl(document, t->id);
        break;
    case ID_HR:
        n = new HTMLHRElementImpl(document);
        break;
    case ID_P:
        n = new HTMLParagraphElementImpl(document);
        break;
    case ID_PRE:
        n = new HTMLPreElementImpl(document);
        break;

// font stuff
    case ID_BASEFONT:
        n = new HTMLBaseFontElementImpl(document);
        break;
    case ID_FONT:
        n = new HTMLFontElementImpl(document);
        break;

// ins/del
    case ID_DEL:
    case ID_INS:
        n = new HTMLModElementImpl(document, t->id);
        break;

// anchor
    case ID_A:
        n = new HTMLAnchorElementImpl(document);
        break;

// images
    case ID_IMG:
        n = new HTMLImageElementImpl(document);
        break;
    case ID_MAP:
        map = new HTMLMapElementImpl(document);
        n = map;
        break;
    case ID_AREA:
        n = new HTMLAreaElementImpl(document);
        break;

// objects, applets and scripts
// TODO
	/*
    case ID_APPLET:
        n = new HTMLAppletElementImpl(document);
        break;
    case ID_EMBED:
        n = new HTMLEmbedElementImpl(document);
        break;
    case ID_OBJECT:
        n = new HTMLObjectElementImpl(document);
        break;
    case ID_PARAM:
        n = new HTMLParamElementImpl(document);
        break;
 	*/
// tables
   case ID_SCRIPT:
        n = new HTMLScriptElementImpl(document);
        break;
   case ID_TABLE:
        n = new HTMLTableElementImpl(document);
        break;
    case ID_CAPTION:
        n = new HTMLTableCaptionElementImpl(document);
        break;
    case ID_COLGROUP:
    case ID_COL:
        n = new HTMLTableColElementImpl(document, t->id);
        break;
    case ID_TR:
	popBlock(ID_TR);
        n = new HTMLTableRowElementImpl(document);
        break;
    case ID_TD:
    case ID_TH:
	popBlock(ID_TH);
	popBlock(ID_TD);
        n = new HTMLTableCellElementImpl(document, t->id);
        break;
    case ID_THEAD:
    case ID_TBODY:
    case ID_TFOOT:
        n = new HTMLTableSectionElementImpl(document, t->id);
        break;

// inline elements
    case ID_BR:
        n = new HTMLBRElementImpl(document);
        break;
    case ID_Q:
        n = new HTMLQuoteElementImpl(document);
        break;

// elements with no special representation in the DOM

// block:
    case ID_ADDRESS:
    case ID_CENTER:
        n = new HTMLGenericElementImpl(document, t->id);
        break;
// inline
        // %fontstyle
    case ID_TT:
    case ID_U:
    case ID_B:
    case ID_I:
    case ID_S:
    case ID_STRIKE:
    case ID_BIG:
    case ID_SMALL:

        // %phrase
    case ID_EM:
    case ID_STRONG:
    case ID_DFN:
    case ID_CODE:
    case ID_SAMP:
    case ID_KBD:
    case ID_VAR:
    case ID_CITE:
    case ID_ABBR:
    case ID_ACRONYM:

        // %special
    case ID_SUB:
    case ID_SUP:
    case ID_SPAN:
        n = new HTMLGenericElementImpl(document, t->id);
        break;

    case ID_BDO:
        break;

        // these are special, and normally not rendered
    case ID_NOEMBED:
        discard_until = ID_NOEMBED + ID_CLOSE_TAG;
        return 0;
    case ID_NOFRAMES:
        discard_until = ID_NOFRAMES + ID_CLOSE_TAG;
        return 0;
    case ID_NOSCRIPT:
        if(HTMLWidget->part()->jScriptEnabled())
            discard_until = ID_NOSCRIPT + ID_CLOSE_TAG;
        return 0;
        // Waldo's plaintext stuff
    case ID_PLAIN:
        return 0;
        break;
// text
    case ID_TEXT:
        n = new TextImpl(document, t->text);
        if (t->complexText )
            n->setComplexText(true);
        break;
    case ID_COMMENT:
#ifdef COMMENTS_IN_DOM
        n = new CommentImpl(document, t->text);
#endif
        break;
    default:
        //kdDebug( 6035 ) << "Unknown tag " << t->id << "!" << endl;
        break;
    }
    return n;
}

void KHTMLParser::processCloseTag(Token *t)
{
    // support for really broken html. Can't believe I'm supporting such crap (lars)
    switch(t->id)
    {
    case ID_HTML+ID_CLOSE_TAG:
    case ID_BODY+ID_CLOSE_TAG:
        // 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.
        return;
    case ID_FORM+ID_CLOSE_TAG:
        form = 0;
        // this one is to get the right style on the body element
        break;
    case ID_MAP+ID_CLOSE_TAG:
        map = 0;
        break;
    case ID_HEAD+ID_CLOSE_TAG:
        //inBody = true;
        // don't close head neither. the creation of body will do it for us.
        // fixes some sites, that define stylesheets after </head>
        return;
    case ID_TITLE+ID_CLOSE_TAG:
        if ( current->id() == ID_TITLE )
          static_cast<HTMLTitleElementImpl *>(current)->setTitle();
        break;
    default:
        break;
    }

#ifdef PARSER_DEBUG
    //kdDebug( 6035 ) << "added the following childs to " << current->nodeName().string() << endl;
    NodeImpl *child = current->firstChild();
    while(child != 0)
    {
        //kdDebug( 6035 ) << "    " << child->nodeName().string() << endl;
        child = child->nextSibling();
    }
#endif
    popBlock(t->id-ID_CLOSE_TAG);
#ifdef PARSER_DEBUG
    //kdDebug( 6035 ) << "closeTag --> current = " << current->nodeName().string() << endl;
#endif
}


void KHTMLParser::pushBlock(int _id, int _level)
{
    HTMLStackElem *Elem = new HTMLStackElem(_id, _level, current, blockStack);

    blockStack = Elem;
    addForbidden(_id, forbiddenTag);
}

void KHTMLParser::popBlock( int _id )
{
    HTMLStackElem *Elem = blockStack;
    int maxLevel = 0;

    while( Elem && (Elem->id != _id))
    {
        if (maxLevel < Elem->level)
        {
            maxLevel = Elem->level;
        }
        Elem = Elem->next;
    }
    if (!Elem || maxLevel > Elem->level)
        return;

    Elem = blockStack;

    while (Elem)
    {
        if (Elem->id == _id)
        {
            popOneBlock();
            Elem = 0;
        }
        else
        {
            popOneBlock();
            Elem = blockStack;
        }
    }
}

void KHTMLParser::popOneBlock()
{
    HTMLStackElem *Elem = blockStack;

    // we should never get here, but some bad html might cause it.
#ifndef PARSER_DEBUG
    if(!Elem) return;
#else
    //kdDebug( 6035 ) << "popping block: " << Elem->id << endl;
#endif

    if(Elem->node != current)
        if(current->renderer()) current->renderer()->close();

    removeForbidden(Elem->id, forbiddenTag);

    blockStack = Elem->next;
    // we only set inline to false, if the element we close is a block level element.
    // This helps getting cases as <p><b>bla</b> <b>bla</b> right.
    if(!current->isInline())
        _inline = false;
    current = Elem->node;

    delete Elem;
}

void KHTMLParser::popInlineBlocks()
{
    while(current->isInline())
        popOneBlock();
}

void KHTMLParser::freeBlock()
{
    while (blockStack)
        popOneBlock();
    blockStack = 0;
}

void KHTMLParser::createHead()
{
    if(head || !document->firstChild())
        return;

    head = new HTMLHeadElementImpl(document);
    try
    {
        HTMLElementImpl *body = document->body();
        document->firstChild()->insertBefore(head, body);
    }
    catch(DOMException e)
    {
#ifdef PARSER_DEBUG
        //kdDebug( 6035 ) << "adding form before of table failed!!!!" << endl;
#endif
        delete head;
        head = 0;
    }
}