htmltokenizer.cpp

monqueror一个很具有参考价值的源玛

                       kdDebug( 6036 ) << "Known attribute: \"" << attrName << "\"" << endl;
#endif
                    }

                    tag = SearchEqual;
                    discard = SpaceDiscard; // discard spaces before '='
                }
                break;
            }
            case SearchEqual:
            {
                if( tquote )
                {
#ifdef TOKEN_DEBUG
                      kdDebug( 6036 ) << "bad HTML in parseTag: SearchEqual" << endl;
#endif
                      // this is moslty due to a missing '"' somewhere before..
                      // so let's start searching for a new tag
                      tquote = NoQuote;

                      Attribute a;
                      a.id = *buffer;
                      if(a.id==0) a.setName( attrName );
                      a.setValue(0, 0);
                      currToken->attrs.add(a);

                      dest = buffer;
                      tag = SearchAttribute;
                      discard = SpaceDiscard;
                      pending = NonePending;
                }
                else if( curchar == '=' )
                {
#ifdef TOKEN_DEBUG
                    kdDebug(6036) << "found equal" << endl;
#endif
                    tag = SearchValue;
                    pending = NonePending; // ignore spaces before '='
                    discard = SpaceDiscard; // discard spaces after '='
                    ++src;
                }
                else if( curchar == '>' )
                    tag = SearchEnd;
                else // other chars indicate a new attribte
                {
                    Attribute a;
                    a.id = *buffer;
                    if(a.id==0) a.setName( attrName );
                    a.setValue(0, 0);
                    currToken->attrs.add(a);

                    dest = buffer;
                    tag = SearchAttribute;
                    discard = SpaceDiscard;
                    pending = NonePending;
                }
                break;
            }
            case SearchValue:
            {
                if(tquote)
                {
                    tag = QuotedValue;
                }
                else
                {
                    tag = Value;
                }
                pending = NonePending;
                break;
            }
            case QuotedValue:
            {
                // ### attributes like '&{blaa....};' are supposed to be treated as jscript.
                if ( curchar == '&' )
                {
                    ++src;

                    discard = NoneDiscard;
                    if (pending)
                        addPending();

                    charEntity = true;
                    parseEntity(src, dest, true);
                    break;
                }
                else if ( !tquote )
                {
                    // end of attribute
                    Attribute a;
                    a.id = *buffer;
                    if(a.id==0) a.setName( attrName );
                    while(*(dest-1) == ' ' && dest>buffer+1) dest--; // remove trailing spaces
                    a.setValue(buffer+1, dest-buffer-1);
#ifdef TOKEN_DEBUG
                    kdDebug() << "adding value: *" << QConstString(buffer+1, dest-buffer-1).string() << "*" << endl;
#endif

                    currToken->attrs.add(a);

                    dest = buffer;
                    tag = SearchAttribute;
                    discard = SpaceDiscard;
                    pending = NonePending;
                    break;
                }
                if( pending ) addPending();

                discard = NoneDiscard;
                *dest++ = src[0];
                ++src;
                break;
            }
            case Value:
            {
                if( tquote )
                {
                  // additional quote. discard it, and define as end of
                  // attribute
#ifdef TOKEN_DEBUG
                    kdDebug( 6036 ) << "bad HTML in parseTag: Value" << endl;
#endif
                    tquote = NoQuote;
                }
                // if discard==NoneDiscard at this point, it means
                // that we passed an empty "" pair. bit hacky, but...
                // helps with <tag attr=""otherattr="something">
                if ( pending || src[0].latin1() == '>' || discard==NoneDiscard)
                {
                    // no quotes. Every space means end of value
                    Attribute a;
                    a.id = *buffer;
                    if(a.id==0) a.setName( attrName );
                    a.setValue(buffer+1, dest-buffer-1);
#ifdef TOKEN_DEBUG
                    kdDebug() << "adding value: *" << QConstString(buffer+1, dest-buffer-1).string() << "*" << endl;
#endif
                    currToken->attrs.add(a);

                    dest = buffer;
                    tag = SearchAttribute;
                    discard = SpaceDiscard;
                    pending = NonePending;
                    break;
                }
                *dest++ = src[0];
                ++src;
                break;
            }
            case SearchEnd:
            {
                if ( curchar != '>')
                {
                    ++src; // discard everything, until we found the end
                    break;
                }

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

                if ( currToken->id == 0 ) //stop if tag is unknown
                {
                    discard = NoneDiscard;
                    *dest = QChar::null;
                    return;
                }

                if(dest>buffer)
                {
                    // add the last attribute
                    Attribute a;
                    a.id = *buffer;
                    if(a.id==0) a.setName( attrName );
                    a.setValue(buffer+1, dest-buffer-1);
                    currToken->attrs.add(a);
                    dest = buffer;
                }
                uint tagID = currToken->id;
#ifdef TOKEN_DEBUG
                kdDebug( 6036 ) << "appending Tag: " << tagID << endl;
#endif
                bool beginTag = (tagID < ID_CLOSE_TAG);
                if( beginTag && tagID != ID_IMG && tagID != ID_INPUT ) {
                    // Ignore Space/LF's after a start tag
                    discard = LFDiscard;
                } else if (!beginTag) {
                    // Don't ignore CR/LF's after a close tag
                    discard = NoneDiscard;
                    tagID -= ID_CLOSE_TAG;
                }

                if ( tagID == ID_SCRIPT  && beginTag ) {
                    int attrIndex = currToken->attrs.find(ATTR_SRC);
                    scriptSrc = (attrIndex == -1 ? (QString)"" : currToken->attrs[attrIndex]->value().string());
                    attrIndex = currToken->attrs.find(ATTR_LANGUAGE);
                    javascript = true;
                    if( attrIndex != -1 ) {
                        QString lang = currToken->attrs[attrIndex]->value().string();
                        lang = lang.lower();
                        if( !lang.contains("javascript") &&
                            !lang.contains("ecmascript") &&
                            !lang.contains("jscript") )
                           javascript = false;
                    } else {
                        attrIndex = currToken->attrs.find(ATTR_TYPE);
                        if( attrIndex != -1 ) {
                            QString lang = currToken->attrs[attrIndex]->value().string();
                            lang = lang.lower();
                            if( !lang.contains("javascript") &&
                                !lang.contains("ecmascript") &&
                                !lang.contains("jscript") )
                                javascript = false;
                        }
                    }
                }

                processToken();

                if(pre)
                {
                    // we have to take care to close the pre block in
                    // case we encounter an unallowed element....
                    if(!DOM::checkChild(ID_PRE, tagID)) {
                        //kdDebug(0) << " not allowed in <pre> " << (int)tagID << endl;
                        pre = false;
                    }
                }

                if ( tagID == ID_PRE )
                {
                    prePos = 0;
                    pre = beginTag;
                }
                else if ( tagID == ID_TEXTAREA )
                {
                    if(beginTag) {
                        listing = true;
			textarea = true;
                        searchCount = 0;
                        searchFor = textareaEnd;
                        scriptCode = QT_ALLOC_QCHAR_VEC( 1024 );
                        scriptCodeSize = 0;
                        scriptCodeMaxSize = 1024;
                        parseListing(src);
                    }
                }
                else if ( tagID == ID_SCRIPT )
                {
                    if (beginTag)
                    {
#ifdef TOKEN_DEBUG
                        kdDebug( 6036 ) << "start of script, token->id = " << currToken->id << endl;
#endif
                        script = true;
                        searchCount = 0;
                        searchFor = scriptEnd;
                        scriptCode = QT_ALLOC_QCHAR_VEC( 1024 );
                        scriptCodeSize = 0;
                        scriptCodeMaxSize = 1024;
                        parseScript(src);
#ifdef TOKEN_DEBUG
                        kdDebug( 6036 ) << "end of script" << endl;
#endif
                    }
                }
                else if ( tagID == ID_STYLE )
                {
                    if (beginTag)
                    {
                        style = true;
                        searchCount = 0;
                        searchFor = styleEnd;
                        scriptCode = QT_ALLOC_QCHAR_VEC( 1024 );
                        scriptCodeSize = 0;
                        scriptCodeMaxSize = 1024;
                        parseStyle(src);
                    }
                }
                else if ( tagID == ID_LISTING )
                {
                    if (beginTag)
                    {
                        listing = true;
                        searchCount = 0;
                        searchFor = listingEnd;
                        scriptCode = QT_ALLOC_QCHAR_VEC( 1024 );
                        scriptCodeSize = 0;
                        scriptCodeMaxSize = 1024;
                        parseListing(src);
                    }
                }
                else if ( tagID == ID_SELECT )
                {
                    select = beginTag;
                }
                return; // Finished parsing tag!
            }
            default:
            {
#ifdef TOKEN_DEBUG
                kdDebug( 6036 ) << "error in parseTag! " << __LINE__ << endl;
#endif
                return;
            }

            } // end switch
        }
    }
    return;
}

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

        switch (pending)
        {
        case SpacePending:
            // Insert a non-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:
#ifdef TOKEN_DEBUG
            kdDebug( 6036 ) << "Assertion failed: pending = " << (int) pending << endl;
#endif
            break;
        }
    }
    else
    {
        *dest++ = ' ';
    }

    pending = NonePending;
}

void HTMLTokenizer::setPlainText()
{
    if (!plaintext)
    {
       // Do this only once!
       plaintext = true;
       currToken->id = ID_PLAIN;
       processToken();
       dest = buffer;
    }
}

void HTMLTokenizer::write( const QString &str )
{
    // we have to make this function reentrant. This is needed, because some
    // script code could call document.write(), which would add something here.
#ifdef TOKEN_DEBUG
    kdDebug( 6036 ) << "Tokenizer::write(\"" << str << "\")" << endl;
#endif
#if DEBUG_BY_XHTANG
	fprintf(stderr,"HTMLTokenizer::write strEmpty:%d buffer:%p\n",str.isEmpty(),buffer);
	fprintf(stderr,"HTMLTokenizer::write str:%s\n",(const char*)str);
#endif
	if ( str.isEmpty() || buffer == 0L )
        return;
    // reentrant...
    // we just insert the code at the tokenizers current position. Parsing will continue once
    // we return from the script stuff
    // (this won't happen if we're in the middle of loading an external script)
    if(executingScript) {
#if DEBUG_BY_XHTANG
	fprintf(stderr,"HTMLTokenizer::write executingScript\n");
#endif
#ifdef TOKEN_DEBUG
        kdDebug( 6036 ) << "adding to scriptOutput" << endl;
#endif
        scriptOutput += str;
        return;
    }

    if (loadingExtScript) {
#if DEBUG_BY_XHTANG
	fprintf(stderr,"HTMLTokenizer::write loadingExtScript\n");			
#endif	
        // don't parse; we will do this later
        pendingSrc += str;
        return;
    }

    _src = str;
    src = DOMStringIt(_src);
    if(!currToken) currToken = new Token;

    if (plaintext)
        parseText(src);
    else if (comment)