htmltokenizer.cpp

手机浏览器源码程序,功能强大

            skipLF = false;
        }

        if (skipLF)
        {
            skipLF = false;
            ++list;
        }
        else if (( *list == '\n' ) || ( *list == '\r' ))
        {
            if (discard == LFDiscard)
            {
                // Ignore this LF
                discard = NoneDiscard; // We have discarded 1 LF
            }
            else
            {
                // Process this LF
                if (pending)
                    addPending();
                pending = LFPending;
            }
            /* Check for MS-DOS CRLF sequence */
            if (*list == '\r')
            {
                skipLF = true;
            }
            ++list;
        }
        else if (( *list == ' ' ) || ( *list == '\t'))
        {
            if (pending)
                addPending();
            if (*list == ' ')
                pending = SpacePending;
            else
                pending = TabPending;

            ++list;
        }
        else
        {
            discard = NoneDiscard;
            if (pending)
                addPending();

            prePos++;
            *dest++ = *list;
            ++list;
        }

    }

    if (pending)
        addPending();

    prePos = 0;

    pre = old_pre;
}

void HTMLTokenizer::parseSpecial(TokenizerString &src)
{
    assert( textarea || title || !Entity );
    assert( !tag );
    assert( xmp+textarea+title+style+script == 1 );
    if (script)
        scriptStartLineno = lineno+src.lineCount();

    if ( comment ) parseComment( src );

    while ( !src.isEmpty() ) {
        RETURN_IF_OOM( checkScriptBuffer( ) );
        unsigned char ch = src->latin1();
        if ( !scriptCodeResync && !brokenComments && !textarea && !xmp && !title && ch == '-' && scriptCodeSize >= 3 && !src.escaped() && scriptCode[scriptCodeSize-3] == '<' && scriptCode[scriptCodeSize-2] == '!' && scriptCode[scriptCodeSize-1] == '-' ) {
            comment = true;
            parseComment( src );
            continue;
        }
        if ( scriptCodeResync && !tquote && ( ch == '>' ) ) {
            ++src;
            scriptCodeSize = scriptCodeResync-1;
            scriptCodeResync = 0;
            scriptCode[ scriptCodeSize ] = scriptCode[ scriptCodeSize + 1 ] = 0;
            if ( script )
                scriptHandler();
            else {
                processListing(TokenizerString(scriptCode, scriptCodeSize));
                processToken();
                if ( style )         { currToken.id = ID_STYLE + ID_CLOSE_TAG; }
                else if ( textarea ) { currToken.id = ID_TEXTAREA + ID_CLOSE_TAG; }
                else if ( title ) { currToken.id = ID_TITLE + ID_CLOSE_TAG; }
                else if ( xmp )  { currToken.id = ID_XMP + ID_CLOSE_TAG; }
                processToken();
                style = script = style = textarea = title = xmp = false;
                tquote = NoQuote;
                scriptCodeSize = scriptCodeResync = 0;
            }
            return;
        }
        // possible end of tagname, lets check.
        if ( !scriptCodeResync && !escaped && !src.escaped() && ( ch == '>' || ch == '/' || ch <= ' ' ) && ch &&
             scriptCodeSize >= searchStopperLen &&
             tagMatch( searchStopper, scriptCode+scriptCodeSize-searchStopperLen, searchStopperLen )) {
            scriptCodeResync = scriptCodeSize-searchStopperLen+1;
            tquote = NoQuote;
            continue;
        }
        if ( scriptCodeResync && !escaped ) {
            if(ch == '\"')
                tquote = (tquote == NoQuote) ? DoubleQuote : ((tquote == SingleQuote) ? SingleQuote : NoQuote);
            else if(ch == '\'')
                tquote = (tquote == NoQuote) ? SingleQuote : (tquote == DoubleQuote) ? DoubleQuote : NoQuote;
            else if (tquote != NoQuote && (ch == '\r' || ch == '\n'))
                tquote = NoQuote;
        }
        escaped = ( !escaped && ch == '\\' );
        if (!scriptCodeResync && (textarea||title) && !src.escaped() && ch == '&') {
            QChar *scriptCodeDest = scriptCode+scriptCodeSize;
            ++src;
            parseEntity(src,scriptCodeDest,true);
            scriptCodeSize = scriptCodeDest-scriptCode;
        }
        else {
            scriptCode[scriptCodeSize] = *src;
            fixUpChar(scriptCode[scriptCodeSize]);
            ++scriptCodeSize;
            ++src;
        }
    }
}

void HTMLTokenizer::scriptHandler()
{
    // We are inside a <script>
    bool doScriptExec = false;
    CachedScript* cs = 0;
    // don't load external scripts for standalone documents (for now)
    if (!scriptSrc.isEmpty() && parser->doc()->part()) {
        // forget what we just got; load from src url instead
        if ( !parser->skipMode() ) {
#ifdef INSTRUMENT_LAYOUT_SCHEDULING
            if (!parser->doc()->ownerElement())
                printf("Requesting script at time %d\n", parser->doc()->elapsedTime());
#endif
            if ( (cs = parser->doc()->docLoader()->requestScript(scriptSrc, scriptSrcCharset) )
#if NOKIA_CHANGES
                && !parser->doc()->fastDisplayMode()
#endif
                )
                cachedScript.enqueue(cs);
        }
        scriptSrc=QString::null;
    }
    else {
#ifdef TOKEN_DEBUG
        kdDebug( 6036 ) << "---START SCRIPT---" << endl;
        kdDebug( 6036 ) << QString(scriptCode, scriptCodeSize) << endl;
        kdDebug( 6036 ) << "---END SCRIPT---" << endl;
#endif
        // Parse scriptCode containing <script> info
#if NOKIA_CHANGES
         if (!parser->doc()->fastDisplayMode())
#endif
            doScriptExec = true;
    }
    processListing(TokenizerString(scriptCode, scriptCodeSize));
    QString exScript( buffer, dest-buffer );
    processToken();
    currToken.id = ID_SCRIPT + ID_CLOSE_TAG;
    processToken();

    TokenizerString *savedPrependingSrc = currentPrependingSrc;
    TokenizerString prependingSrc;
    currentPrependingSrc = &prependingSrc;
    if ( !parser->skipMode() ) {
        if (cs) {
             //kdDebug( 6036 ) << "cachedscript extern!" << endl;
             //kdDebug( 6036 ) << "src: *" << QString( src.current(), src.length() ).latin1() << "*" << endl;
             //kdDebug( 6036 ) << "pending: *" << pendingSrc.latin1() << "*" << endl;
#if NOKIA_CHANGES
             if (!parser->doc()->fastDisplayMode()) {
#endif
                if (savedPrependingSrc) {
                    savedPrependingSrc->append(src);
                } else {
                    pendingSrc.prepend(src);
                }
                setSrc(TokenizerString());
                scriptCodeSize = scriptCodeResync = 0;
                cs->ref(this);
                // will be 0 if script was already loaded and ref() executed it
                if (!cachedScript.isEmpty())
                    loadingExtScript = true;
#if NOKIA_CHANGES
            }
#endif
        }
        else if (view && doScriptExec && javascript ) {
            if (!m_executingScript)
                pendingSrc.prepend(src);
            else
                prependingSrc = src;
            setSrc(TokenizerString());
            scriptCodeSize = scriptCodeResync = 0;
            //QTime dt;
            //dt.start();
            scriptExecution( exScript, QString::null, scriptStartLineno );
	    //kdDebug( 6036 ) << "script execution time:" << dt.elapsed() << endl;
        }
    }

    script = false;
    scriptCodeSize = scriptCodeResync = 0;

    if ( !m_executingScript && !loadingExtScript ) {
	// kdDebug( 6036 ) << "adding pending Output to parsed string" << endl;
	src.append(pendingSrc);
	pendingSrc.clear();
    } else if (!prependingSrc.isEmpty()) {
	// restore first so that the write appends in the right place
	// (does not hurt to do it again below)
	currentPrependingSrc = savedPrependingSrc;

	// we need to do this slightly modified bit of one of the write() cases
	// because we want to prepend to pendingSrc rather than appending
	// if there's no previous prependingSrc
	if (loadingExtScript) {
	    if (currentPrependingSrc) {
		currentPrependingSrc->append(prependingSrc);
	    } else {
		pendingSrc.prepend(prependingSrc);
	    }
	} else {
	    write(prependingSrc, false);
	}
    }

    currentPrependingSrc = savedPrependingSrc;
}

void HTMLTokenizer::scriptExecution( const QString& str, QString scriptURL,
                                     int baseLine)
{
#if APPLE_CHANGES
    if (!view || !view->part())
        return;
#endif
    bool oldscript = script;
    m_executingScript++;
    script = false;
    QString url;
    if (scriptURL.isNull())
      url = static_cast<DocumentImpl*>(view->part()->document().handle())->URL();
    else
      url = scriptURL;

    TokenizerString *savedPrependingSrc = currentPrependingSrc;
    TokenizerString prependingSrc;
    currentPrependingSrc = &prependingSrc;

#ifdef INSTRUMENT_LAYOUT_SCHEDULING
    if (!parser->doc()->ownerElement())
        printf("beginning script execution at %d\n", parser->doc()->elapsedTime());
#endif

    view->part()->executeScript(url,baseLine,Node(),str);

    allowYield = true;

#ifdef INSTRUMENT_LAYOUT_SCHEDULING
    if (!parser->doc()->ownerElement())
        printf("ending script execution at %d\n", parser->doc()->elapsedTime());
#endif

    m_executingScript--;
    script = oldscript;

    if ( !m_executingScript && !loadingExtScript ) {
	// kdDebug( 6036 ) << "adding pending Output to parsed string" << endl;
	src.append(pendingSrc);
	pendingSrc.clear();
    } else if (!prependingSrc.isEmpty()) {
	// restore first so that the write appends in the right place
	// (does not hurt to do it again below)
	currentPrependingSrc = savedPrependingSrc;

	// we need to do this slightly modified bit of one of the write() cases
	// because we want to prepend to pendingSrc rather than appending
	// if there's no previous prependingSrc
	if (loadingExtScript) {
	    if (currentPrependingSrc) {
		currentPrependingSrc->append(prependingSrc);
	    } else {
		pendingSrc.prepend(prependingSrc);
	    }
	} else {
	    write(prependingSrc, false);
	}
    }

    currentPrependingSrc = savedPrependingSrc;
}

void HTMLTokenizer::parseComment(TokenizerString &src)
{
    RETURN_IF_OOM( checkScriptBuffer(src.length()) );
    while ( !src.isEmpty() ) {
        scriptCode[ scriptCodeSize++ ] = *src;
#if defined(TOKEN_DEBUG) && TOKEN_DEBUG > 1
        qDebug("comment is now: *%s*",
               QConstString((QChar*)src.current(), QMIN(16, src.length())).string().latin1());
#endif
        if (src->unicode() == '>') {
            bool handleBrokenComments = brokenComments && !(script || style);
            int endCharsCount = 1; // start off with one for the '>' character
            if (scriptCodeSize > 2 && scriptCode[scriptCodeSize-3] == '-' && scriptCode[scriptCodeSize-2] == '-') {
                endCharsCount = 3;
            }
            else if (scriptCodeSize > 3 && scriptCode[scriptCodeSize-4] == '-' && scriptCode[scriptCodeSize-3] == '-' &&
                scriptCode[scriptCodeSize-2] == '!') {
                // Other browsers will accept --!> as a close comment, even though it's
                // not technically valid.
                endCharsCount = 4;
            }
            if (handleBrokenComments || endCharsCount > 1) {
                ++src;
                if (!( script || xmp || textarea || style)) {
                    if (includesCommentsInDOM) {
                        RETURN_IF_OOM( checkScriptBuffer() );
                        scriptCode[ scriptCodeSize ] = 0;
                        scriptCode[ scriptCodeSize + 1 ] = 0;
                        currToken.id = ID_COMMENT;
                        processListing(TokenizerString(scriptCode, scriptCodeSize - endCharsCount));
                        processToken();
                        currToken.id = ID_COMMENT + ID_CLOSE_TAG;
                        processToken();
                    }
                    scriptCodeSize = 0;
                }
                comment = false;
                return; // Finished parsing comment
            }
        }
        ++src;
    }
}

void HTMLTokenizer::parseServer(TokenizerString &src)
{
    RETURN_IF_OOM( checkScriptBuffer(src.length()) );
    while ( !src.isEmpty() ) {
        scriptCode[ scriptCodeSize++ ] = *src;
        if (src->unicode() == '>' &&
            scriptCodeSize > 1 && scriptCode[scriptCodeSize-2] == '%') {
            ++src;
            server = false;
            scriptCodeSize = 0;
            return; // Finished parsing server include
        }
        ++src;
    }
}

void HTMLTokenizer::parseProcessingInstruction(TokenizerString &src)
{
    char oldchar = 0;
    while ( !src.isEmpty() )
    {
        unsigned char chbegin = src->latin1();
        if(chbegin == '\'') {
            tquote = tquote == SingleQuote ? NoQuote : SingleQuote;
        }
        else if(chbegin == '\"') {
            tquote = tquote == DoubleQuote ? NoQuote : DoubleQuote;
        }
        // Look for '?>'
        // some crappy sites omit the "?" before it, so
        // we look for an unquoted '>' instead. (IE compatible)
        else if ( chbegin == '>' && ( !tquote || oldchar == '?' ) )