parser.java

gcc的组建

            (
              getTokenAhead(1).kind == END ||
              (getTokenAhead(1).kind == WS && getTokenAhead(2).kind == END)
            )
           )
          {
            buffer.append(t.getImage().substring(0, t.getImage().length() - 2));

            /* Skip the closing > that we have already checked. */
            last = mustBe(t.kind);
            break comment;
          }
        else
          append(t);
        mustBe(t.kind);
      }
    hTag = new Token(start, last);
    handleComment();
  }

  /**
  * Read a script. The text, returned without any changes,
  * is terminated only by the closing tag SCRIPT.
  */
  protected void Script()
                 throws ParseException
  {
    Token name;

    Token start = hTag = mustBe(BEGIN);
    optional(WS);

    name = mustBe(SCRIPT);

    optional(WS);

    restOfTag(false, name, start);

    buffer.setLength(0);

    script: 
    while (!SCRIPT_CLOSE.matches(this))
      {
        append(getNextToken());
      }

    consume(SCRIPT_CLOSE);

    _handleText();

    endTag(false);
    _handleEndTag(makeTagElement(name.getImage(), false));
  }

  /**
  * Process SGML insertion that is not a comment.
  */
  protected void Sgml()
               throws ParseException
  {
    if (COMMENT_OPEN.matches(this))
      Comment();
    else // skip till ">"
      {
        Token start = hTag = mustBe(BEGIN);
        optional(WS);
        mustBe(EXCLAMATION);

        buffer.setLength(0);
        read: 
        while (true)
          {
            t = getNextToken();
            if (t.kind == Constants.ENTITY)
              {
                resolveAndAppendEntity(t);
              }
            else if (t.kind == EOF)
              {
                error("unexpected eof", t);
                break read;
              }
            else if (t.kind == END)
              break read;
            else
              append(t);
          }

        try
          {
            parseMarkupDeclarations(buffer);
          }
        catch (IOException ex)
          {
            error("Unable to parse SGML insertion: '" + buffer + "'",
                  new Token(start, t)
                 );
          }
      }
  }

  /**
  * Read a style definition. The text, returned without any changes,
  * is terminated only by the closing tag STYLE.
  */
  protected void Style()
                throws ParseException
  {
    Token name;

    Token start = hTag = mustBe(BEGIN);
    optional(WS);

    name = mustBe(STYLE);

    optional(WS);

    restOfTag(false, name, start);

    buffer.setLength(0);

    style: 
    while (!STYLE_CLOSE.matches(this))
      {
        append(getNextToken());
      }

    consume(STYLE_CLOSE);

    _handleText();

    endTag(false);
    _handleEndTag(makeTagElement(name.getImage(), false));
  }

  /**
   * Read a html tag.
   */
  protected void Tag()
              throws ParseException
  {
    mark(true);

    boolean closing = false;
    Token name;
    Token start = hTag = mustBe(BEGIN);

    optional(WS);
    name = getNextToken();
    optional(WS);

    if (name.kind == SLASH)
      {
        closing = true;
        name = getNextToken();
      }

    restOfTag(closing, name, start);
  }

  /**
   * A hook, for operations, preceeding call to handleText.
   * Handle text in a string buffer.
   * In non - preformatted mode, all line breaks immediately following the
   * start tag and immediately before an end tag is discarded,
   * \r, \n and \t are replaced by spaces, multiple space are replaced
   * by the single one and the result is  moved into array,
   * passing it  to handleText().
   */
  protected void _handleText()
  {
    char[] text;

    if (preformatted > 0)
      text = textProcessor.preprocessPreformatted(buffer);
    else
      text = textProcessor.preprocess(buffer);

    if (text != null && text.length > 0)
      {
        TagElement pcdata = new TagElement(dtd.getElement("#pcdata"));
        attributes = htmlAttributeSet.EMPTY_HTML_ATTRIBUTE_SET;
        _handleEmptyTag(pcdata);

        handleText(text);
        if (titleOpen)
          title.append(text);
      }
  }

  /**
   * Add the image of this token to the buffer.
   * @param t A token to append.
   */
  protected final void append(Token t)
  {
    if (t.kind != EOF)
      t.appendTo(buffer);
  }

  /**
   * Consume pattern that must match.
   * @param p A pattern to consume.
   */
  protected final void consume(pattern p)
  {
    node n;
    for (int i = 0; i < p.nodes.length; i++)
      {
        n = p.nodes [ i ];
        if (n.optional)
          optional(n.kind);
        else
          mustBe(n.kind);
      }
  }

  /**
   * The method is called when the HTML end (closing) tag is found or if
   * the parser concludes that the one should be present in the
   * current position. The method is called immediatly
   * before calling the handleEndTag().
   * @param omitted True if the tag is no actually present in the document,
   * but is supposed by the parser (like &lt;/html&gt; at the end of the
   * document).
   */
  protected void endTag(boolean omitted)
  {
  }

  /**
   * Handle HTML comment. The default method returns without action.
   * @param comment
   */
  protected void handleComment(char[] comment)
  {
  }

  /**
   * This is additionally called in when the HTML content terminates
   * without closing the HTML comment. This can only happen if the
   * HTML document contains errors (for example, the closing --;gt is
   * missing.
   */
  protected void handleEOFInComment()
  {
    error("Unclosed comment");
  }

  /**
   * Handle the tag with no content, like &lt;br&gt;. The method is
   * called for the elements that, in accordance with the current DTD,
   * has an empty content.
   * @param The tag being handled.
   * @throws javax.swing.text.ChangedCharSetException
   */
  protected void handleEmptyTag(TagElement tag)
                         throws javax.swing.text.ChangedCharSetException
  {
  }

  /**
   * The method is called when the HTML closing tag ((like &lt;/table&gt;)
   * is found or if the parser concludes that the one should be present
   * in the current position.
   * @param The tag
   */
  protected void handleEndTag(TagElement tag)
  {
  }

  /* Handle error that has occured in the given line. */
  protected void handleError(int line, String message)
  {
  }

  /**
   * The method is called when the HTML opening tag ((like &lt;table&gt;)
   * is found or if the parser concludes that the one should be present
   * in the current position.
   * @param The tag
   */
  protected void handleStartTag(TagElement tag)
  {
  }

  /**
   * Handle the text section.
   * <p> For non-preformatted section, the parser replaces
   * \t, \r and \n by spaces and then multiple spaces
   * by a single space. Additionaly, all whitespace around
   * tags is discarded.
   * </p>
   * <p> For pre-formatted text (inside TEXAREA and PRE), the parser preserves
   * all tabs and spaces, but removes <b>one</b>  bounding \r, \n or \r\n,
   * if it is present. Additionally, it replaces each occurence of \r or \r\n
   * by a single \n.</p>
   *
   * @param text A section text.
   */
  protected void handleText(char[] text)
  {
  }

  /**
   * Handle HTML &lt;title&gt; tag. This method is invoked when
   * both title starting and closing tags are already behind.
   * The passed argument contains the concatenation of all
   * title text sections.
   * @param The title text.
   */
  protected void handleTitle(char[] title)
  {
  }

  /**
   * Constructs the tag from the given element. In this implementation,
   * this is defined, but never called.
   * @return the tag
   */
  protected TagElement makeTag(Element element)
  {
    return makeTag(element, false);
  }

  /**
   * Constructs the tag from the given element.
   * @param the tag base {@link javax.swing.text.html.parser.Element}
   * @param isSupposed true if the tag is not actually present in the
   * html input, but the parser supposes that it should to occur in
   * the current location.
   * @return the tag
   */
  protected TagElement makeTag(Element element, boolean isSupposed)
  {
    return new TagElement(element, isSupposed);
  }

  /**
   * This is called when the tag, representing the given element,
   * occurs first time in the document.
   * @param element
   */
  protected void markFirstTime(Element element)
  {
  }

  /**
   * Consume the token that was checked before and hence MUST be present.
   * @param kind The kind of token to consume.
   */
  protected Token mustBe(int kind)
  {
    if (getTokenAhead().kind == kind)
      return getNextToken();
    else
      {
        String ei = "";
        if (kind < 1000)
          ei = " ('" + (char) kind + "') ";
        throw new AssertionError("The token of kind " + kind + ei +
                                 " MUST be here,"
                                );
      }
  }

  /**
   * Handle attribute without value. The default method uses
   * the only allowed attribute value from DTD.
   * If the attribute is unknown or allows several values,
   * the HTML.NULL_ATTRIBUTE_VALUE is used. The attribute with
   * this value is added to the attribute set.
   * @param element The name of element.
   * @param attribute The name of attribute without value.
   */
  protected void noValueAttribute(String element, String attribute)
  {
    Object value = HTML.NULL_ATTRIBUTE_VALUE;

    Element e = (Element) dtd.elementHash.get(element.toLowerCase());
    if (e != null)
      {
        AttributeList attr = e.getAttribute(attribute);
        if (attr != null)
          {
            Vector values = attr.values;
            if (values != null && values.size() == 1)
              value = values.get(0);
          }
      }
    attributes.addAttribute(attribute, value);
  }

  /**
   * Consume the optional token, if present.
   * @param kind The kind of token to consume.
   */
  protected Token optional(int kind)
  {
    if (getTokenAhead().kind == kind)
      return getNextToken();
    else
      return null;
  }

  /** Parse the html document. */
  protected void parseDocument()
                        throws ParseException
  {
    while (getTokenAhead().kind != EOF)
      {
        advanced = false;
        if (TAG.matches(this))
          Tag();
        else if (COMMENT_OPEN.matches(this))
          Comment();
        else if (STYLE_OPEN.matches(this))
          Style();
        else if (SCRIPT_OPEN.matches(this))
          Script();
        else if (SGML.matches(this))
          Sgml();
        else
          CDATA(true);

        // Surely HTML error, treat as a text.
        if (!advanced)
          {
            Token wrong = getNextToken();
            error("unexpected '" + wrong.getImage() + "'", wrong);
            buffer.setLength(0);
            buffer.append(wrong.getImage());
            _handleText();
          }
      }
  }

  /**
   * Read the element attributes, adding them into attribute set.
   * @param element The element name (needed to access attribute
   * information in dtd).
   */
  protected void readAttributes(String element)
  {
    Token name;
    Token value;
    Token next;
    String attrValue;

    attributes = new htmlAttributeSet();

    optional(WS);

    attributeReading: 
    while (getTokenAhead().kind == NUMTOKEN)
      {
        name = getNextToken();
        optional(WS);

        next = getTokenAhead();
        if (next.kind == EQ)
          {
            mustBe(EQ);
            optional(WS);

            next = getNextToken();

            switch (next.kind)
              {
                case QUOT :

                  // read "quoted" attribute.
                  buffer.setLength(0);
                  readTillTokenE(QUOT);
                  attrValue = buffer.toString();
                  break;

                case AP :

                  // read 'quoted' attribute.
                  buffer.setLength(0);
                  readTillTokenE(AP);