analyser.jj

/*******************************
***** SECTION 1 - OPTIONS *****
*******************************/
options { JAVA_UNICODE_ESCAPE = true; 
		   IGNORE_CASE = true;
		}
/*********************************
***** SECTION 2 - USER CODE *****
*********************************/
PARSER_BEGIN(SLPTokeniser)
public class SLPTokeniser{
	public static void main(String args[]){
	    String temp;
	    STC temp2;

	    if (args.length < 1)
	    {
	      System.out.println("Please pass in the filename.");
	      System.exit(1);
	    }

	    SLPTokeniser parser = new SLPTokeniser(anew FileInputStream(args[0]));

	    SimpleNode root = parser.program();

	    System.out.println("Abstract Syntax Tree:");

	    root.dump(" ");
}
PARSER_END(SLPTokeniser)
/*****************************************
***** SECTION 3 - TOKEN DEFINITIONS *****
*****************************************/
TOKEN_MGR_DECLS :
{
	static int commentNesting = 0;
}
SKIP : /*** Ignoring spaces/tabs/newlines ***/
{
	" "
	| "\t"
	| "\n"
	| "\r"
	| "\f"
}
SKIP : /* COMMENTS */
{
	<"/*"> { commentNesting++; } : IN_COMMENT 
	| < "//" (~["\n"])*"\n">
}
<IN_COMMENT> SKIP :
{
	"/*" { commentNesting++; }
	| "*/" { commentNesting--;
	if (commentNesting == 0)
		SwitchTo(DEFAULT);
	}
	| <~[]>
}
TOKEN : /* Keywords and punctuation */
{
	< SEMIC : ";" >
	| < COLON : ":" >
	| < ASSIGN : ":=" >
	| < LBR : "(" >
	| < RBR : ")" >
	| < COMMA : "," >
	| < PLUS_SIGN : "+" >
	| < MINUS_SIGN : "-" >
	| < MULT_SIGN : "*" >
	| < DIV_SIGN : "/" >
	| < RCBR : "}" > 
	| < LCBR : "{" >
	| < EXLM : "!" >
	| < QUESTION : "?" >
	| < PERCENT : "%" >
	| < EQUALS : "=" > 
	| < NOT_EQUAL : "!=" >
	| < LEFT_BRACE : "<" >
	| < RIGHT_BRACE : ">" >
	| < LEFT_BRACE_EQUAL : "<=" >
	| < RIGHT_BRACE_EQUAL : ">=" >
	| < DOT : "." >
	| < AND : "and" >
	| < BOOL : "boolean" >
	| < CONST : "constant" >
	| < DO : "do" >
	| < ELSE : "else" >
	| < FALSE : "false" >
	| < IF : "if" >
	| < INT : "integer" >
	| < MAIN : "main" >
	| < NOT : "not" >
	| < OR : "or" >
	| < REAL : "real" >
	| < RETURN : "return" >
	| < STR : "string" >
	| < THEN : "then" >
	| < TRUE : "true" >
	| < VAR : "var" >
	| < VOID : "void" >
	| < WHILE : "while" >
	
}
//Starts with a single-double quote, then anything on the keyboard
//Doesn't accept "" or \n, unless it finds a backslash, in which case they're allowed
TOKEN : /* Strings */
{
	< STRING : "\"" ((~[ "\"" , "\n", "\\"] )* | (("\\" ("\"" | "\n" | "\\" | "'\n\'"))))* "\"">
}
TOKEN : /* Numbers and identifiers */
{
	< NUM : (<DIGIT>)+ | (<DIGIT>) + "." (<DIGIT>) + >
	| < #DIGIT : ["0" - "9"] >
	| < ID : (<LETTER>)+ (<LETTER> | "_" | <DIGIT>) * >
	| < #LETTER : ["a" - "z", "A" - "Z"] >
}
TOKEN : /* Anything not recognised so far */
{
	< OTHER : ~[] >
}

//Grammar starts here

void program() : {}{
	(decl())*(function())*main_prog()
}

void decl() : {}{
	(var_decl() | const_decl())
}

void var_decl () : {}{
	<VAR>ident_list()<COLON>type()(<COMMA>ident_list()<COLON>type())*<SEMIC>
}

void const_decl() : {}{
	<CONST><ID><COLON>type()<EQUALS>expression()(<COMMA><ID><COLON>type()<EQUALS>expression())*<SEMIC>
}

void function () :{} {
	type()<ID><LBR>param_list()<RBR><LCBR>
		(decl())*(statement()<SEMIC>)*<RETURN><LBR>expression() | {}<RBR><SEMIC>
	<RCBR>
}
void param_list () :{} {
	(<ID><COLON>type()(<COMMA><ID><COLON>type())* | {})
}

void type () :{}{
	<INT> | <BOOL> | <REAL> | <STR> | <VOID>
}

void main_prog() :{} {
	<MAIN>
	<LCBR>
	(decl())*(statement()<SEMIC>)*
	<RCBR>
}

void statement () :{}{
	<ID><ASSIGN>expression()
	|<ID><ASSIGN><STR>
	|<EXLM>expression()
	|<QUESTION><ID>
	|<ID><LBR>arg_list()<RBR>
	|<LCBR>(statement()<SEMIC>)*<RCBR>
	|<IF>condition()<THEN>statement()
	|<IF>condition()<THEN>statement()<ELSE>statement()
	|<WHILE>condition()<DO>statement()
	|{}
}

void expression(): {}{
	fragment()((<PLUS_SIGN>|<MINUS_SIGN>|<MULT_SIGN>|<DIV_SIGN>)fragment())*
}

void fragment() :{}{
	<ID>
	|<NUM>
	|(<PLUS_SIGN>|<MINUS_SIGN>)fragment()
	|expression()
}

void condition() :{}{
	<NOT>expression()
	|expression()(<EQUALS>|<NOT_EQUAL>|<LEFT_BRACE>|<RIGHT_BRACE>|<LEFT_BRACE_EQUAL>|<RIGHT_BRACE_EQUAL>|<AND>|<OR>)
	|<ID>
}

void ident_list () :{}{
	<ID>(<COMMA><ID>)*
}

void arg_list():{}{
	(<ID>(<COMMA><ID>)* | {})
}