I'm having trouble getting antlr to skip whitespace where I want it to.

I've defined the following rules in my Lexer:

WHITESPACE

: ( ' '

| ' '

| 'f'

| NEWLINE

)+

{ $setType(Token.SKIP); }

;



protected

NEWLINE

: ( options { generateAmbigWarnings=false; }

: "
" // Macintosh style

| "
" // MS-DOS style

| "
" // UNIX style

) { newline(); }

;



protected

LEGAL_CHAR

: (~( '{' | '}' | ',' | '[' | ']' | ' ' | ' ' | 'f' | '
' | '
'));



LABEL_STRING

: '[' (LEGAL_CHAR)+ ( ',' (LEGAL_CHAR)+)* ']';



When I define LABEL_STRING in this fashion, I have trouble with Strings that include whitespace between the first '[' and the first block of LEGAL_CHARs, such as "[ a]".

If I redefine the rule as such:

LABEL_STRING

: '[' (WHITESPACE)? (LEGAL_CHAR)+ ( ',' (LEGAL_CHAR)+)* (WHITESPACE)? ']';



This circumvents the problem, but why do I need to include the additional (WHITESPACE)? ?

Thanks,

-Nathan