[antlr-interest] (newbie question) where can SLIST get its information from?

[Kay Roepke]

Hi!

On Jul 8, 2007, at 5:30 PM, ali azimi wrote:

>   If there is no left curly in some rule, where can an imaginary  
> node SLIST get its line and column information from? for example,  
> in the following rewrite rule, what can I put instead of question  
> mark in front of lc=?
>
>   textualsystemdefinition
>  :lc=?  packagereferenceclause*('system' systemname end  
> entityinsystem+ 'endsystem'(systemname)?end|typebasedsystemdefinition)
>  ->^(SLIST[$lc] packagereferenceclause* 'system'systemname end  
> entityinsystem+ 'endsystem' systemname? end typebasedsystemdefinition)
>  ;

As you don't have an explicit token reference you will need to use  
the first token that has been matched by textualsystemdefinition.
Consider this grammar:

grammar ASTConst;

options {
	output=AST;
	ASTLabelType=CommonTree;
}

tokens {
SLIST;
}

rule
	:	subrule? subrule1 -> ^(SLIST[$rule.start] subrule? subrule1)
	;
	
subrule
	:	'A';

subrule1
	:	'B';


$rule.start refers to the first token that has been matched by rule,  
making it the ideal candidate for your purposes :)
(For reference: Page 130 in the ANTLR book lists the predefined  
parser rule attributes. Among those is start. Haven't looked whether  
this
is in the wiki already.)

HTH,
-k

-- 
Kay Röpke
http://classdump.org/