[antlr-interest] Please help with arbitrary arrangements and oldBNF syntax

[Geir Ove Skjaervik]

Hello,


I haven't tried this, but I am pretty sure the following would do the
trick:
I use a semtic predicates to rule out choices already chosen. Maybe
there are other ways to do this.
I put the code together real fast, so exuse me if there are logical
errors:


written 
	: w1 w2 w3
	;

w1
	:  a:author | p:program | d:dataOrigin w2[a, p, d]
	;

w2[AST a, AST p, AST d]	
	: {a != null}? p1:program | d1:dataOrigin w3[p1, d1]
	| {p != null}? a2:author | d2:dataOrigin w4[a2, d2]
	| {d != null}? a3:author | p3:program w5[a3, p2]
	;


w3[AST p, AST d]	
	: {p != null}? dataOrigin 
	| {d != null}? program 
	;

w4[AST a, AST d]	
	: {a != null}? dataOrigin 
	| {d != null}? program 
	;


w5[AST a, AST p]	
	: {a != null}? program 
	| {p != null}? author 
	;


Geir Ove

-----Original Message-----
From: antlr-interest-bounces at antlr.org
[mailto:antlr-interest-bounces at antlr.org] On Behalf Of Brannon King
Sent: 16. desember 2005 20:46
To: antlr-interest at antlr.org
Subject: [antlr-interest] Please help with arbitrary arrangements and
oldBNF syntax


Suppose I have the following parser rule snipped from edif.org:

written :
	LP! "written"
	     timeStamp
	     ( < author > | < program > | < dataOrigin > |
	     property | comment | userData )*
	RP!;

Where the angle brackets mean that those can only appear once (in an
arbitrary order). How do I enforce this in an ANTLR grammar?

In a similar vein, is this the right way to limit an identifier length?

IDENTIFIER
	options {testLiterals=true;} //make sure it's not a built in
function
	:	(ALPHA|'&')	a:('_'|ALPHA|DIGIT)*
{a.getText().length() < 256}? ; //less than 256 chars total

Thanks for your time