CIT 594 Errata #5 for Bugs Grammar (and Code)
Spring 2008, David Matuszek

It seems like there's no end to the corrections....

Last year in CIT594 the main project used ints. The StreamTokenizer really wants to return doubles, so I had to cast them to int. This year we are using doubles, so the cast needs to be removed. I didn't think to update the Recognizer class (or the RecognizerTest class) to take this into account, which is why everything worked.

Was: 
Token nextToken() {
    int code;
    try { code = tokenizer.nextToken(); }
    catch (IOException e) { throw new Error(e); } // Should never happen
    switch (code) {
        case StreamTokenizer.TT_WORD:
            if (Token.KEYWORDS.contains(tokenizer.sval)) {
                return new Token(Token.Type.KEYWORD, tokenizer.sval);
            }
            else {
                return new Token(Token.Type.NAME, tokenizer.sval);
            }
        case StreamTokenizer.TT_NUMBER:
            return new Token(Token.Type.NUMBER, ((int) tokenizer.nval) + "");
        case StreamTokenizer.TT_EOL:
            return new Token(Token.Type.EOL, "\n");
        case StreamTokenizer.TT_EOF:
            return new Token(Token.Type.EOF, "EOF");
        default:
            return new Token(Token.Type.SYMBOL, ((char) code) + "");
    }
}
Now: 
Token nextToken() {
    int code;
    try { code = tokenizer.nextToken(); }
    catch (IOException e) { throw new Error(e); } // Should never happen
    switch (code) {
        case StreamTokenizer.TT_WORD:
            if (Token.KEYWORDS.contains(tokenizer.sval)) {
                return new Token(Token.Type.KEYWORD, tokenizer.sval);
            }
            else {
                return new Token(Token.Type.NAME, tokenizer.sval);
            }
        case StreamTokenizer.TT_NUMBER:
            return new Token(Token.Type.NUMBER, tokenizer.nval + "");
        case StreamTokenizer.TT_EOL:
            return new Token(Token.Type.EOL, "\n");
        case StreamTokenizer.TT_EOF:
            return new Token(Token.Type.EOF, "EOF");
        default:
            return new Token(Token.Type.SYMBOL, ((char) code) + "");
    }
}

You should be aware that the String value returned for a number written without a decimal point has now changed. For example, the number 5 will now be returned by the tokenizer as 5.0.

Corrections have been made to both Recognizer.java and RecognizerTest.java.