[antlr-interest] predicate question

[Loring Craymer]

I'll vote nay.  p2 is guaranteed to be tested out of context in a synpred (the sempred will execute but not the action).  PCCTS did not (IIRC) worry about actions when hoisting sempreds.

This is a playing with fire issue.  The rule needs to be "semantic context changes which will affect the outcome of a decision cannot legally be made within the scope of that decision".  You can never be sure when you will need to insert a synpred in a grammar, so sempred behavior within synpreds should be the same as when they are encountered outside a synpred.

I believe that changing the currently implemented behavior would be unwise--this is the sort of change that introduces new and more exciting behaviors (bug city).

--Loring


----- Original Message ----
> From: Terence Parr <parrt at cs.usfca.edu>
> To: ANTLR Interest <antlr-interest at antlr.org>
> Sent: Monday, May 5, 2008 6:20:28 PM
> Subject: [antlr-interest] predicate question
> 
> Consider:
> 
> parser grammar T;
> 
> a : {p1}? {action} {p2}? A
>    | A
>    ;
> 
> Someone pointed out that p2 should not be tested out of context (i.e.,  
> before action executes).  ANTLR v2 (PCCTS) correctly ignored preds  
> after actions.  v3 definitely walks right over actions looking for  
> preds.   I propose to prevent p2 from being hoisted into the decision  
> for rule 'a'.  Right now it generates enclosed DFA.  After fix, it  
> would only see p1.  Not sure how to fix just yet, but i'll figure it  
> out.
> 
> Contrary opinions?
> 
> Ter
> 
> 




      ____________________________________________________________________________________
Be a better friend, newshound, and 
know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ