htmltokenizer.cpp

linux下开源浏览器WebKit的源码,市面上的很多商用浏览器都是移植自WebKit

            return state; // Finished parsing server include
        }
        src.advance(m_lineNumber);
    }
    return state;
}

HTMLTokenizer::State HTMLTokenizer::parseProcessingInstruction(SegmentedString& src, State state)
{
    UChar oldchar = 0;
    while (!src.isEmpty()) {
        UChar chbegin = *src;
        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 == '?')) {
            // We got a '?>' sequence
            state.setInProcessingInstruction(false);
            src.advancePastNonNewline();
            state.setDiscardLF(true);
            return state; // Finished parsing comment!
        }
        src.advance(m_lineNumber);
        oldchar = chbegin;
    }
    
    return state;
}

HTMLTokenizer::State HTMLTokenizer::parseText(SegmentedString& src, State state)
{
    while (!src.isEmpty()) {
        UChar cc = *src;

        if (state.skipLF()) {
            state.setSkipLF(false);
            if (cc == '\n') {
                src.advancePastNewline(m_lineNumber);
                continue;
            }
        }

        // do we need to enlarge the buffer?
        checkBuffer();

        if (cc == '\r') {
            state.setSkipLF(true);
            *m_dest++ = '\n';
        } else
            *m_dest++ = cc;
        src.advance(m_lineNumber);
    }

    return state;
}


HTMLTokenizer::State HTMLTokenizer::parseEntity(SegmentedString& src, UChar*& dest, State state, unsigned& cBufferPos, bool start, bool parsingTag)
{
    if (start) {
        cBufferPos = 0;
        state.setEntityState(SearchEntity);
        EntityUnicodeValue = 0;
    }

    while(!src.isEmpty()) {
        UChar cc = *src;
        switch(state.entityState()) {
        case NoEntity:
            ASSERT(state.entityState() != NoEntity);
            return state;
        
        case SearchEntity:
            if (cc == '#') {
                m_cBuffer[cBufferPos++] = cc;
                src.advancePastNonNewline();
                state.setEntityState(NumericSearch);
            } else
                state.setEntityState(EntityName);
            break;

        case NumericSearch:
            if (cc == 'x' || cc == 'X') {
                m_cBuffer[cBufferPos++] = cc;
                src.advancePastNonNewline();
                state.setEntityState(Hexadecimal);
            } else if (cc >= '0' && cc <= '9')
                state.setEntityState(Decimal);
            else
                state.setEntityState(SearchSemicolon);
            break;

        case Hexadecimal: {
            int ll = min(src.length(), 10 - cBufferPos);
            while (ll--) {
                cc = *src;
                if (!((cc >= '0' && cc <= '9') || (cc >= 'a' && cc <= 'f') || (cc >= 'A' && cc <= 'F'))) {
                    state.setEntityState(SearchSemicolon);
                    break;
                }
                int digit;
                if (cc < 'A')
                    digit = cc - '0';
                else
                    digit = (cc - 'A' + 10) & 0xF; // handle both upper and lower case without a branch
                EntityUnicodeValue = EntityUnicodeValue * 16 + digit;
                m_cBuffer[cBufferPos++] = cc;
                src.advancePastNonNewline();
            }
            if (cBufferPos == 10)  
                state.setEntityState(SearchSemicolon);
            break;
        }
        case Decimal:
        {
            int ll = min(src.length(), 9-cBufferPos);
            while(ll--) {
                cc = *src;

                if (!(cc >= '0' && cc <= '9')) {
                    state.setEntityState(SearchSemicolon);
                    break;
                }

                EntityUnicodeValue = EntityUnicodeValue * 10 + (cc - '0');
                m_cBuffer[cBufferPos++] = cc;
                src.advancePastNonNewline();
            }
            if (cBufferPos == 9)  
                state.setEntityState(SearchSemicolon);
            break;
        }
        case EntityName:
        {
            int ll = min(src.length(), 9-cBufferPos);
            while(ll--) {
                cc = *src;

                if (!((cc >= 'a' && cc <= 'z') || (cc >= '0' && cc <= '9') || (cc >= 'A' && cc <= 'Z'))) {
                    state.setEntityState(SearchSemicolon);
                    break;
                }

                m_cBuffer[cBufferPos++] = cc;
                src.advancePastNonNewline();
            }
            if (cBufferPos == 9) 
                state.setEntityState(SearchSemicolon);
            if (state.entityState() == SearchSemicolon) {
                if(cBufferPos > 1) {
                    // Since the maximum length of entity name is 9,
                    // so a single char array which is allocated on
                    // the stack, its length is 10, should be OK.
                    // Also if we have an illegal character, we treat it
                    // as illegal entity name.
                    unsigned testedEntityNameLen = 0;
                    char tmpEntityNameBuffer[10];

                    ASSERT(cBufferPos < 10);
                    for (; testedEntityNameLen < cBufferPos; ++testedEntityNameLen) {
                        if (m_cBuffer[testedEntityNameLen] > 0x7e)
                            break;
                        tmpEntityNameBuffer[testedEntityNameLen] = m_cBuffer[testedEntityNameLen];
                    }

                    const Entity *e;

                    if (testedEntityNameLen == cBufferPos)
                        e = findEntity(tmpEntityNameBuffer, cBufferPos);
                    else
                        e = 0;

                    if(e)
                        EntityUnicodeValue = e->code;

                    // be IE compatible
                    if(parsingTag && EntityUnicodeValue > 255 && *src != ';')
                        EntityUnicodeValue = 0;
                }
            }
            else
                break;
        }
        case SearchSemicolon:
            // Don't allow values that are more than 21 bits.
            if (EntityUnicodeValue > 0 && EntityUnicodeValue <= 0x10FFFF) {
                if (!inViewSourceMode()) {
                    if (*src == ';')
                        src.advancePastNonNewline();
                    if (EntityUnicodeValue <= 0xFFFF) {
                        checkBuffer();
                        src.push(fixUpChar(EntityUnicodeValue));
                    } else {
                        // Convert to UTF-16, using surrogate code points.
                        checkBuffer(2);
                        src.push(U16_LEAD(EntityUnicodeValue));
                        src.push(U16_TRAIL(EntityUnicodeValue));
                    }
                } else {
                    // FIXME: We should eventually colorize entities by sending them as a special token.
                    checkBuffer(11);
                    *dest++ = '&';
                    for (unsigned i = 0; i < cBufferPos; i++)
                        dest[i] = m_cBuffer[i];
                    dest += cBufferPos;
                    if (*src == ';') {
                        *dest++ = ';';
                        src.advancePastNonNewline();
                    }
                }
            } else {
                checkBuffer(10);
                // ignore the sequence, add it to the buffer as plaintext
                *dest++ = '&';
                for (unsigned i = 0; i < cBufferPos; i++)
                    dest[i] = m_cBuffer[i];
                dest += cBufferPos;
            }

            state.setEntityState(NoEntity);
            return state;
        }
    }

    return state;
}

HTMLTokenizer::State HTMLTokenizer::parseDoctype(SegmentedString& src, State state)
{
    ASSERT(state.inDoctype());
    while (!src.isEmpty() && state.inDoctype()) {
        UChar c = *src;
        bool isWhitespace = c == '\r' || c == '\n' || c == '\t' || c == ' ';
        switch (m_doctypeToken.state()) {
            case DoctypeBegin: {
                m_doctypeToken.setState(DoctypeBeforeName);
                if (isWhitespace) {
                    src.advance(m_lineNumber);
                    if (inViewSourceMode())
                        m_doctypeToken.m_source.append(c);
                }
                break;
            }
            case DoctypeBeforeName: {
                if (c == '>') {
                    // Malformed.  Just exit.
                    src.advancePastNonNewline();
                    state.setInDoctype(false);
                    if (inViewSourceMode())
                        processDoctypeToken();
                } else if (isWhitespace) {
                    src.advance(m_lineNumber);
                    if (inViewSourceMode())
                        m_doctypeToken.m_source.append(c);
                } else
                    m_doctypeToken.setState(DoctypeName);
                break;
            }
            case DoctypeName: {
                if (c == '>') {
                    // Valid doctype. Emit it.
                    src.advancePastNonNewline();
                    state.setInDoctype(false);
                    processDoctypeToken();
                } else if (isWhitespace) {
                    m_doctypeSearchCount = 0; // Used now to scan for PUBLIC
                    m_doctypeSecondarySearchCount = 0; // Used now to scan for SYSTEM
                    m_doctypeToken.setState(DoctypeAfterName);
                    src.advance(m_lineNumber);
                    if (inViewSourceMode())
                        m_doctypeToken.m_source.append(c);
                } else {
                    src.advancePastNonNewline();
                    m_doctypeToken.m_name.append(c);
                    if (inViewSourceMode())
                        m_doctypeToken.m_source.append(c);
                }
                break;
            }
            case DoctypeAfterName: {
                if (c == '>') {
                    // Valid doctype. Emit it.
                    src.advancePastNonNewline();
                    state.setInDoctype(false);
                    processDoctypeToken();
                } else if (!isWhitespace) {
                    src.advancePastNonNewline();
                    if (toASCIILower(c) == publicStart[m_doctypeSearchCount]) {
                        m_doctypeSearchCount++;
                        if (m_doctypeSearchCount == 6)
                            // Found 'PUBLIC' sequence
                            m_doctypeToken.setState(DoctypeBeforePublicID);
                    } else if (m_doctypeSearchCount > 0) {
                        m_doctypeSearchCount = 0;
                        m_doctypeToken.setState(DoctypeBogus);
                    } else if (toASCIILower(c) == systemStart[m_doctypeSecondarySearchCount]) {
                        m_doctypeSecondarySearchCount++;
                        if (m_doctypeSecondarySearchCount == 6)
                            // Found 'SYSTEM' sequence
                            m_doctypeToken.setState(DoctypeBeforeSystemID);
                    } else {
                        m_doctypeSecondarySearchCount = 0;
                        m_doctypeToken.setState(DoctypeBogus);
                    }
                    if (inViewSourceMode())
                        m_doctypeToken.m_source.append(c);
                } else {
                    src.advance(m_lineNumber); // Whitespace keeps us in the after name state.
                    if (inViewSourceMode())
                        m_doctypeToken.m_source.append(c);
                }
                break;
            }
            case DoctypeBeforePublicID: {
                if (c == '\"' || c == '\'') {
                    tquote = c == '\"' ? DoubleQuote : SingleQuote;
                    m_doctypeToken.setState(DoctypePublicID);
                    src.advancePastNonNewline();
                    if (inViewSourceMode())
                        m_doctypeToken.m_source.append(c);
                } else if (c == '>') {
                    // Considered bogus.  Don't process the doctype.
                    src.advancePastNonNewline();
                    state.setInDoctype(false);