htmltokenizer.cpp

将konqueror浏览器移植到ARM9 2410中

            int ll = kMin(src.length(), CBUFLEN-cBufferPos);

            while(ll--) {
                curchar = *src;
                if(curchar <= '>') {
                    if(curchar <= ' ' || curchar == '=' || curchar == '>') {
                        unsigned int a;
                        cBuffer[cBufferPos] = '\0';
                        a = khtml::getAttrID(cBuffer, cBufferPos);
                        if ( !a )
                            attrName = QString::fromLatin1(QCString(cBuffer, cBufferPos+1).data());

                        dest = buffer;
                        *dest++ = a;
#ifdef TOKEN_DEBUG
                        if (!a || (cBufferPos && *cBuffer == '!'))
                            kdDebug( 6036 ) << "Unknown attribute: *" << QCString(cBuffer, cBufferPos+1).data() << "*" << endl;
                        else
                            kdDebug( 6036 ) << "Known attribute: " << QCString(cBuffer, cBufferPos+1).data() << endl;
#endif
                        tag = SearchEqual;
                        break;
                    }
                }
                cBuffer[cBufferPos++] = (char) curchar | 0x20;
                ++src;
            }
            if ( cBufferPos == CBUFLEN ) {
                cBuffer[cBufferPos] = '\0';
                attrName = QString::fromLatin1(QCString(cBuffer, cBufferPos+1).data());
                dest = buffer;
                *dest++ = 0;
                tag = SearchEqual;
            }
            break;
        }
        case SearchEqual:
        {
#if defined(TOKEN_DEBUG) && TOKEN_DEBUG > 1
                qDebug("SearchEqual");
#endif
            ushort curchar;
            bool atespace = false;
            while(src.length()) {
                curchar = src->unicode();
                if(curchar > ' ') {
                    if(curchar == '=') {
#ifdef TOKEN_DEBUG
                        kdDebug(6036) << "found equal" << endl;
#endif
                        tag = SearchValue;
                        ++src;
                    }
                    else if(atespace && (curchar == '\'' || curchar == '"'))
                    {
                        tag = SearchValue;
                        *dest++ = 0;
                        attrName = QString::null;
                    }
                    else {
                        AttrImpl* a = 0;
                        if(*buffer)
                            a = new AttrImpl(parser->docPtr(), (int)*buffer);
                        else if ( !attrName.isEmpty() && attrName != "/" )
                            a = new AttrImpl(parser->docPtr(), attrName);

                        if ( a ) {
                            a->setValue("");
                            currToken.insertAttr(a);
                        }

                        dest = buffer;
                        tag = SearchAttribute;
                    }
                    break;
                }
                atespace = true;
                ++src;
            }
            break;
        }
        case SearchValue:
        {
            ushort curchar;
            while(src.length()) {
                curchar = src->unicode();
                if(curchar > ' ') {
                    if(( curchar == '\'' || curchar == '\"' )) {
                        tquote = curchar == '\"' ? DoubleQuote : SingleQuote;
                        tag = QuotedValue;
                        ++src;
                    } else
                        tag = Value;

                    break;
                }
                ++src;
            }
            break;
        }
        case QuotedValue:
        {
#if defined(TOKEN_DEBUG) && TOKEN_DEBUG > 1
                qDebug("QuotedValue");
#endif
            ushort curchar;
            while(src.length()) {
                checkBuffer();

                curchar = src->unicode();
                if(curchar <= '\'' && !src.escaped()) {
                    // ### attributes like '&{blaa....};' are supposed to be treated as jscript.
                    if ( curchar == '&' )
                    {
                        ++src;
                        parseEntity(src, dest, true);
                        break;
                    }
                    else if ( (tquote == SingleQuote && curchar == '\'') ||
                              (tquote == DoubleQuote && curchar == '\"') )
                    {
                        // end of attribute
                        AttrImpl* a;

                        if(*buffer)
                            a = new AttrImpl(parser->docPtr(), (int)*buffer);
                        else
                            a = new AttrImpl(parser->docPtr(), DOMString(attrName));

                        if(a->attrId || !attrName.isNull())
                        {
                            // some <input type=hidden> rely on trailing spaces. argh
                            while(dest > buffer+1 && (*(dest-1) == '\n' || *(dest-1) == '\r'))
                                dest--; // remove trailing newlines
                            a->setValue(DOMString(buffer+1, dest-buffer-1));
                            currToken.insertAttr(a);
                        }
                        else {
                            // hmm, suboptimal, but happens seldom
                            delete a;
                            a = 0;
                        }

                        dest = buffer;
                        tag = SearchAttribute;
                        tquote = NoQuote;
                        ++src;
                        break;
                    }
                }
                *dest++ = *src;
                ++src;
            }
            break;
        }
        case Value:
        {
#if defined(TOKEN_DEBUG) && TOKEN_DEBUG > 1
            qDebug("Value");
#endif
            ushort curchar;
            while(src.length()) {
                checkBuffer();
                curchar = src->unicode();
                if(curchar <= '>' && !src.escaped()) {
                    // parse Entities
                    if ( curchar == '&' )
                    {
                        ++src;
                        parseEntity(src, dest, true);
                        break;
                    }
                    // '/' does not delimit in IE!
                    if ( curchar <= ' ' || curchar == '>' )
                    {
                        // no quotes. Every space means end of value
                        AttrImpl* a;
                        if(*buffer)
                            a = new AttrImpl(parser->docPtr(), (int)*buffer);
                        else
                            a = new AttrImpl(parser->docPtr(), DOMString(attrName));

                        a->setValue(DOMString(buffer+1, dest-buffer-1));
                        currToken.insertAttr(a);

                        dest = buffer;
                        tag = SearchAttribute;
                        break;
                    }
                }

                *dest++ = *src;
                ++src;
            }
            break;
        }
        case SearchEnd:
        {
#if defined(TOKEN_DEBUG) && TOKEN_DEBUG > 1
                qDebug("SearchEnd");
#endif
            while(src.length()) {
                if(*src == '>')
                    break;

                ++src;
            }
            if(!src.length() && *src != '>') break;

            searchCount = 0; // Stop looking for '<!--' sequence
            tag = NoTag;
            tquote = NoQuote;
            ++src;

            if ( !currToken.id ) //stop if tag is unknown
                return;

            uint tagID = currToken.id;
#if defined(TOKEN_DEBUG) && TOKEN_DEBUG > 0
            kdDebug( 6036 ) << "appending Tag: " << tagID << endl;
#endif
            bool beginTag = tagID < ID_CLOSE_TAG;

            if(!beginTag)
                tagID -= ID_CLOSE_TAG;
            else if ( beginTag && tagID == ID_SCRIPT ) {
                AttrImpl* a = 0;
                scriptSrc = scriptSrcCharset = "";
                if ( currToken.attrs && !parser->doc()->view()->part()->onlyLocalReferences()) {
                    if ( ( a = currToken.attrs->getIdItem( ATTR_SRC ) ) )
                        scriptSrc = parser->doc()->view()->part()->completeURL(khtml::parseURL( a->value() ).string() ).url();
                    if ( ( a = currToken.attrs->getIdItem( ATTR_CHARSET ) ) )
                        scriptSrcCharset = a->value().string().stripWhiteSpace();
                    if ( scriptSrcCharset.isEmpty() )
                        scriptSrcCharset = parser->doc()->view()->part()->encoding();
                    a = currToken.attrs->getIdItem( ATTR_LANGUAGE );
                }
                javascript = true;
                if( a ) {
                    QString lang = a->value().string();
                    lang = lang.lower();
                    if( !lang.contains("javascript") &&
                        !lang.contains("ecmascript") &&
                        !lang.contains("livescript") &&
                        !lang.contains("jscript") )
                        javascript = false;
                } else {
                    if( currToken.attrs )
                        a = currToken.attrs->getIdItem(ATTR_TYPE);
                    if( a ) {
                        QString lang = a->value().string();
                        lang = lang.lower();
                        if( !lang.contains("javascript") &&
                            !lang.contains("ecmascript") &&
                            !lang.contains("livescript") &&
                            !lang.contains("jscript") )
                            javascript = false;
                    }
                }
            }

            processToken();

            // we have to take care to close the pre block in
            // case we encounter an unallowed element....
            if(pre && beginTag && !DOM::checkChild(ID_PRE, tagID)) {
                kdDebug(6036) << " not allowed in <pre> " << (int)tagID << endl;
                pre = false;
            }

            switch( tagID ) {
            case ID_PRE:
                prePos = 0;
                pre = beginTag;
                break;
            case ID_TEXTAREA:
                if(beginTag)
                    parseSpecial(src, textarea = true );
                break;
            case ID_SCRIPT:
                if (beginTag)
                    parseSpecial(src, script = true);
                break;
            case ID_STYLE:
                if (beginTag)
                    parseSpecial(src, style = true);
                break;
            case ID_LISTING:
                if (beginTag)
                    parseSpecial(src, listing = true);
                break;
            case ID_SELECT:
                select = beginTag;
                break;
            case ID_PLAINTEXT:
                plaintext = beginTag;
                break;
            }
            return; // Finished parsing tag!
        }
        } // end switch
    }
    return;
}

void HTMLTokenizer::addPending()
{
    if ( select)
    {
        *dest++ = ' ';
    }
    else if ( textarea )
    {
        if (pending == LFPending)
            *dest++ = '\n';
        else
            *dest++ = ' ';
    }
    else if ( pre )
    {
        int p;

        switch (pending)
        {
        case SpacePending:
            // Insert a breaking space
            *dest++ = QChar(' ');
            prePos++;
            break;

        case LFPending:
            *dest = '\n';
            dest++;
            prePos = 0;
            break;

        case TabPending:
            p = TAB_SIZE - ( prePos % TAB_SIZE );
            for ( int x = 0; x < p; x++ )
            {
                *dest = QChar(' ');
                dest++;
            }
            prePos += p;
            break;

        default:
#if defined(TOKEN_DEBUG) && TOKEN_DEBUG > 1
            kdDebug( 6036 ) << "Assertion failed: pending = " << (int) pending << endl;
#endif
            break;
        }
    }
    else
    {
        *dest++ = ' ';
    }

    pending = NonePending;
}

void HTMLTokenizer::write( const QString &str, bool appendData )
{
#ifdef TOKEN_DEBUG
    kdDebug( 6036 ) << "Tokenizer::write(\"" << str << "\"," << appendData << ")" << endl;
#endif

    if ( !buffer )
        return;

    if ( ( m_executingScript && appendData ) ||
         ( !m_executingScript && loadingExtScript ) ) {
        // don't parse; we will do this later
        pendingSrc += str;
        return;
    }

    if ( onHold ) {
        QString rest = QString( src.current(), src.length() );
        rest += str;
        _src = rest;
        return;
    }
    else
        _src = str;

    src = DOMStringIt(_src);

    if (Entity)
        parseEntity(src, dest);

    if ( src.length() ) {
        if (plaintext)
            parseText(src);
        else if (script)
            parseSpecial(src, false);
        else if (style)