xqparser.java

A framework written in Java for implementing high-level and dynamic languages, compiling them into J

  static final int ELEMENT_TOKEN = 251; // 'element' followed by '{' or alpha
  static final int ATTRIBUTE_TOKEN = 252;// 'attribute' followed by '{' or alpha
  static final int TEXT_TOKEN = 253; // 'text' followed by '{'
  static final int COMMENT_TOKEN = 254; // 'text' followed by '{'
  static final int PI_TOKEN = 255; // 'processing-instruction' followed by '{' or alpha
  static final int DOCUMENT_TOKEN = 256; // ;document' followed by '{'
  
  private int saveToken;
  private Object saveValue;

  public void mark ()
    throws java.io.IOException
  {
    super.mark();
    saveToken = curToken;
    saveValue = curValue;
  }

  public void reset()
    throws java.io.IOException
  {
    curToken = saveToken;
    curValue = saveValue;
    super.reset();
  }

  private int setToken (int token, int width)
  {
    curToken = token;
    curLine = port.getLineNumber() + 1;
    curColumn = port.getColumnNumber() + 1 - width;
    return token;
  }

  void checkSeparator (char ch)
  {
    if (XName.isNameStart(ch))
      error('e', "missing separator", "XPST0003");
  }

  int getRawToken()
      throws java.io.IOException, SyntaxException
  {
    int next;
    for (;;)
      {
	next = read();
	if (next < 0)
	  return setToken(EOF_TOKEN, 0);
	if (next == '\n' || next == '\r')
	  {
	    if (nesting <= 0)
	      return setToken(EOL_TOKEN, 0);
	  }
	else if (next == '(')
	  {
	    if (checkNext(':'))
	      skipComment();
            else if (checkNext('#'))
              return setToken(PRAGMA_START_TOKEN, 2);
	    else
	      return setToken('(', 1);
	  }
	else if (next == '{')
	  {
	    if (! checkNext('-'))
	      return setToken('{', 1);
	    next = read();
	    if (next != '-')
	      {
		// FIXME backup 2 chars. Can fix using special token for '{-'.
		unread();
		unread();
		return setToken('{', 1);
	      }
	    skipOldComment();
	  }
	else if (next != ' ' && next != '\t')
	  break;
      }
    tokenBufferLength = 0;
    curLine = port.getLineNumber() + 1;
    curColumn = port.getColumnNumber();
    char ch = (char) next;
    switch (ch)
      {
      case ')':  case '[':  case ']':  case '}':
      case '$':  case '@':  case ',':  case '?':  case ';':
	break;
      case ':':
	if (checkNext('='))
	  ch = COLON_EQUAL_TOKEN;
	else if (checkNext(':'))
	  ch = COLON_COLON_TOKEN;
	break;
      case '|':
	ch = OP_UNION;
	break;
      case '*':
	ch = OP_MUL;
	break;
      case '+':
	ch = OP_ADD;
	break;
      case '-':
	ch = OP_SUB;
	break;
      case '!':
	if (checkNext('='))
	  ch = OP_NEQ;
	break;
      case '/':
	if (checkNext('/'))
	  ch = SLASHSLASH_TOKEN;
	break;
      case '=':
	if (checkNext('>'))
	  ch = ARROW_TOKEN;
	ch = OP_EQU;
	break;
      case '>':
	ch = checkNext('=') ? (char) OP_GEQ
	  : checkNext('>') ? (char) OP_GRTGRT : (char) OP_GRT;
	break;
      case '<':
	ch = checkNext('=') ? (char) OP_LEQ
	  : checkNext('<') ? (char) OP_LSSLSS : (char) OP_LSS;
	break;
      case '\'':  case '\"':
	char saveReadState = pushNesting ((char) next);
	for (;;)
	  {
	    next = read();
	    if (next < 0)
	      eofError("unexpected end-of-file in string starting here");
	    if (next == '&')
	      {
		parseEntityOrCharRef();
		continue;
	      }
	    else if (ch == next)
	      {
		next = read ();
		if (ch != next)
		  {
		    unread(next);
		    break;
		  }
	      }
	    tokenBufferAppend((char) next);
	  }
	popNesting(saveReadState);
	ch = STRING_TOKEN;
	break;
      default:
	if (Character.isDigit(ch)
            || (ch == '.' && Character.isDigit((char) peek())))
	  {
	    boolean seenDot = ch == '.';
	    for (;; )
	      {
		tokenBufferAppend(ch);
		next = read();
		if (next < 0)
		  break;
		ch = (char) next;
		if (ch == '.')
		  {
		    if (seenDot)  break;
		    seenDot = true;
		  }
		else if (! Character.isDigit(ch))
		  break;
	      }
	    if (next == 'e' || next == 'E')
	      {
		tokenBufferAppend((char) next);
		next = read();
		if (next == '+' || next == '-')
		  {
		    tokenBufferAppend((char) next);
		    next = read();
		  }
		int expDigits = 0;
		for (;;)
		  {
		    if (next < 0)
		      break;
		    ch = (char) next;
		    if (! Character.isDigit(ch))
		      {
                        checkSeparator(ch);
			unread();
			break;
		      }
		    tokenBufferAppend(ch);
		    next = read();
		    expDigits++;
		  }
		if (expDigits == 0)
		  error('e', "no digits following exponent", "XPST0003");
		ch = DOUBLE_TOKEN;
	      }
	    else
	      {
		ch = seenDot ? DECIMAL_TOKEN : INTEGER_TOKEN;
		if (next >= 0)
                  {
                    checkSeparator((char) next);
                    unread(next);
                  }
	      }
	  }
        else if (ch == '.')
          {
            if (checkNext('.'))
              ch = DOTDOT_TOKEN;
	    break;
          }
	else if (XName.isNameStart(ch))
	  {
	    for (;;)
	      {
		tokenBufferAppend(ch);
		next = read();
		ch = (char) next;
		if (! XName.isNamePart(ch))
		  break;
	      }
	    if (next < 0)
	      ch = NCNAME_TOKEN;
	    else
	      {
		if (next != ':')
		    ch = NCNAME_TOKEN;
		else
		  {
		    next = read();
		    if (next < 0)
		      eofError("unexpected end-of-file after NAME ':'");
		    ch = (char) next;
		    if (XName.isNameStart(ch))
		      {
			tokenBufferAppend(':');
			for (;;)
			  {
			    tokenBufferAppend(ch);
			    next = read();
			    ch = (char) next;
			    if (! XName.isNamePart(ch))
			      break;
			  }
			ch = QNAME_TOKEN;
		      }
		    else if (ch == '=')
		      {
			unread(ch);
			ch = NCNAME_TOKEN;
		      }
		    else
		      ch = NCNAME_COLON_TOKEN;
		  }
		unread(next);
	      }
	  }
	else if (ch >= ' ' && ch < 127)
	  syntaxError("invalid character '"+ch+'\'');
	else
	  syntaxError("invalid character '\\u"+Integer.toHexString(ch)+'\'');
      }
    curToken = ch;
    return ch;
  }

  /** Scan until a given delimiter.
   * On success, text upto the delimiter is in then tokenBuffer (with
   * tokenBufferLength marking its length); the delimiter is not included.
   */
  public void getDelimited(String delimiter)
      throws java.io.IOException, SyntaxException
  {
    tokenBufferLength = 0;
    int dlen = delimiter.length();
    char last = delimiter.charAt(dlen-1);
    for (;;)
      {
	int ch = read();
	if (ch < 0)
	  eofError("unexpected end-of-file looking for '"+delimiter+'\'');
	int dstart, j;
	// Look for a match for the last delimiter character.
	if (ch == last
	    && (dstart = tokenBufferLength - (j = dlen - 1)) >= 0)
	  {
	    // Check that the initial part of the delimiter has also been seen.
	    do
	      {
		if (j == 0)
		  {
		    tokenBufferLength = dstart;
		    return;
		  }
		j--;
	      }
	    while (tokenBuffer[dstart+j] == delimiter.charAt(j));
	  }
	tokenBufferAppend((char) ch);
      }
  }

  public void appendNamedEntity(String name)
  {
    name = name.intern();
    char ch = '?';
    if (name == "lt")
      ch = '<';
    else if (name == "gt")
      ch = '>';
    else if (name == "amp")
      ch = '&';
    else if (name == "quot")
      ch = '"';
    else if (name == "apos")
      ch = '\'';
    else
      error("unknown enity reference: '"+name+"'");
    tokenBufferAppend(ch);
  }

  boolean match (String word1, String word2, boolean force)
      throws java.io.IOException, SyntaxException
  {
    if (match(word1))
      {
        mark();
        getRawToken();
        if (match(word2))
          {
            reset();
            getRawToken();
            return true;
          }
        reset();
        if (force)
          {
            error('e', "'"+word1+"' must be followed by '"+word2+"'",
                  "XPST0003");
            return true;
          }
      }
    return false;
  }

  /** Return the current token, assuming it is in operator context.
   * Resolve NCNAME_TOKEN (identifier) to 'and', 'or', 'div', etc.
   */
  int peekOperator()
      throws java.io.IOException, SyntaxException
  {
    while (curToken == EOL_TOKEN)
      {
	if (nesting == 0)
	  return EOL_TOKEN;
	getRawToken();
      }
    if (curToken == NCNAME_TOKEN)
      {
	int len = tokenBufferLength;
        char c1, c2, c3;
        switch (len)
          {
          case 2:
            c1 = tokenBuffer[0];
            c2 = tokenBuffer[1];
            if (c1 == 'o' && c2 == 'r')
              curToken = OP_OR;
            else if (c1 == 't' && c2 == 'o')
              curToken = OP_RANGE_TO;
            else if (c1 == 'i' && c2 == 's')
              curToken = OP_IS;
            else if (c1 == 'e' && c2 == 'q')
              curToken = OP_EQ;
            else if (c1 == 'n' && c2 == 'e')
              curToken = OP_NE;
            else if (c1 == 'g')
              {
                if (c2 == 'e')  curToken = OP_GE;
                else if (c2 == 't')  curToken = OP_GT;
              }
            else if (c1 == 'l')
              {
                if (c2 == 'e')  curToken = OP_LE;
                else if (c2 == 't')  curToken = OP_LT;
              }
            break;

          case 3:
            c1 = tokenBuffer[0];
            c2 = tokenBuffer[1];
            c3 = tokenBuffer[2];
            if (c1 == 'a')
              {
                if (c2 == 'n' && c3 == 'd')
                  curToken = OP_AND;
              }
            else if (c1 == 'm') {
              if (c2 == 'u' && c3 == 'l')
                curToken = OP_MUL;
              if (c2 == 'o' && c3 == 'd')
                curToken = OP_MOD;
            }
            else if (c1 == 'd') {
              if (c2 == 'i' && c3 == 'v')
                curToken = OP_DIV;
            }
            break;
          case 4:
            if (match("idiv"))
              curToken = OP_IDIV;
            else if (match("cast", "as", true))
              curToken = OP_CAST_AS;
            break;
          case 5:
            if (match("where"))
              curToken = OP_WHERE;
            else if (match("isnot"))
              curToken = OP_ISNOT;
            else if (match("union"))
              curToken = OP_UNION;
            else if (match("treat", "as", true))
              curToken = OP_TREAT_AS;
            break;
          case 6:
            if (match("except"))
              curToken = OP_EXCEPT;
            break;
          case 8:
            if (match("instance", "of", true))
              curToken = OP_INSTANCEOF;
            else if (match("castable", "as", true))
              curToken = OP_CASTABLE_AS;
            break;
          case 9:
            if (match("intersect"))
              curToken = OP_INTERSECT;
            break;
          case 10:
            if (match("instanceof")) // obsolete
              {
                warnOldVersion("use 'instanceof of' (two words) instead of 'instanceof'");
                curToken = OP_INSTANCEOF;
              }
            break;
          default:
            break;
          }
      }
    return curToken;