grammar cfg;

PIPE	: '|';
PLUS	: '+';
STAR	: '*';

LPAREN	: '(';
RPAREN	: ')';
LBRACE	: '{';
RBRACE	: '}';

COLON	: ':';
SEMICOLON : ';';


NUMBER
	: '0'..'9'+
	;

fragment LETTER 
	: 'a'..'z' | 'A'..'Z' | '0'..'9' | '_' | '-'
	;

// Encompasses both terminals and nonterminals. Distinguishing them is done later.
// This is way more broad than your typical programming language identifier: leading
// digits are OK and hyphens are OK.
SYMBOL
	: LETTER+
	;
	
// For separating productions and rules
ARROW	
	: '->'
	| COLON
	;

WHITESPACE
	: '\r' | '\n' | ' ' { $channel = HIDDEN; }
	;
	


// Either a symbol or id:symbol, or a parenthisized expression.
atom
	: SYMBOL
	| NUMBER COLON SYMBOL
	| LPAREN disjunction RPAREN
	;

term
	: atom (STAR|PLUS)?
	;

// Represents concatenation	
termString
	: term+
	;

disjunction
	: termString (PIPE termString)*
	; 	


production
	: SYMBOL ARROW disjunction SEMICOLON
	;

rules
	: production+
	;