[antlr-interest] resolving ambiguity

[Stephen Siegel]

No difference.

On May 22, 2012, at 11:42 AM, Terence Parr wrote:

> try (options {k=1;} : B)+
> Ter
> On May 21, 2012, at 4:12 PM, Stephen Siegel wrote:
> 
>> Thanks, Ter.  That's true, but the warnings are scary -- or at least suggest I'm not doing something quite right -- and I want to make them go away.  (This is an abstraction of a much larger grammar where I ran into this problem.  In the larger grammar, there are a bunch of such warnings.)   Is there a way to tell ANTLR that I want that behavior, so don't warn me?
>> -Steve
>> 
>> On May 21, 2012, at 6:14 PM, Terence Parr wrote:
>> 
>>> hi Steve, it will match greedily so that b_chain will match all B's in a row.
>>> Ter
>>> On May 21, 2012, at 8:50 AM, Stephen Siegel wrote:
>>> 
>>>> In the following grammar:
>>>> 
>>>> -----------------------------------------------
>>>> grammar g1;
>>>> options { output=AST; }
>>>> tokens { B_CHAIN; }
>>>> r	:	(A | b_chain)*;
>>>> b_chain	:	B+ -> ^(B_CHAIN B+);
>>>> A 	:	'A';
>>>> B	:	'B';
>>>> 
>>>> // BB should yield ^(BCHAIN B B)
>>>> -----------------------------------------------
>>>> 
>>>> there is an ambiguity since input "BB" could be parsed as (1) two "b_chain"s (each with one "B") or as (2) one b_chain (with two "B"s).  I want to tell ANTLR to choose option (2).
>>>> 
>>>> Right now, I get this warning:
>>>> 
>>>> -------------------------------------
>>>> [11:45:02] Checking Grammar g1.g...
>>>> [11:45:02] warning(200): g1.g:5:11: 
>>>> Decision can match input such as "B" using multiple alternatives: 1, 2
>>>> 
>>>> As a result, alternative(s) 2 were disabled for that input
>>>> -------------------------------------
>>>> 
>>>> Does anyone know how to resolve the ambiguity?