[antlr-interest] Grammar Perplexity in v3.0 (More)

[Randall R Schulz]

Hello again,

I've found what appears to be a genuine anomaly. Using ANTLRworks, I've 
shown that the AtomicWord lexical rule works properly when explicitly 
applied to inputs that it describes.

When applied to the production rule "plainTerm" that directly addresses 
AtomicWord, ANTLRworks reports "MismatchedTokenException(4!=6)" as the 
direct descendant of plainTerm:

- [plainTerm]
  [MismatchedTokenException(4!=6)]


When the test input (the two characters "gt") is applied to AtomicWord, 
the expected derivation occurs. but when the same input is submitted to 
plainTerm:

plainTerm
    :    AtomicWord ( '(' arguments ')' ) ?


... the aforementioned error occurs.


Surely this is a bug? Or am I confused about something?


Here are the pertinent rules:


plainTerm
    :    AtomicWord ( '(' arguments ')' ) ?

AtomicWord
    :   LowerWord
    ;

fragment
LowerWord
    :   LowerAlpha ( Alphanumeric * )
    ;

fragment
Alphanumeric
    :   ( LowerAlpha | UpperAlpha | Numeric | '_' )
    ;

fragment
LowerAlpha
    :   ( 'a'..'z' )
    ;

fragment
UpperAlpha
    :   ( 'A'..'Z' )
    ;

fragment
Numeric
    :   ( '0'..'9' )
    ;



Am I missing something, or is this some kind of parser (or parser 
generator) error?


Randall Schulz