[antlr-interest] executing actions during backtrack

[Terence Parr]

Howdy,

We need to discuss v3's behavior for executing actions during  
backtracking (guessing).  ANTLR must parse ahead to see if something  
matches.  If it fails, then ANTLR tries the next viable alt.  Upon  
failure, it's pretty hard to undo actions in general so ANTLR gates  
actions out with something like this:

if ( backtracking==0 ) {
      SymbolTable.getPredefinedType("void"), symtab.getDefaultPkg());
}

Labels are still defined, because semantic preds might need them:

                             packageDefinition1=packageDefinition();

AST actions are off during backtracking:

                             if ( backtracking==0 )  
list_packageDefinition.add(packageDefinition1.tree);

Upon success, antlr still rewinds the input and then does the same  
parse again "with feeling". ;)

Clearly you don't want actions executed when that alt ultimately  
doesn't succeed.  Similarly, when the alt does succeed, you do not  
want actions executed twice.  This all makes sense.

The problem arises when you want actions to execute *during*  
backtracking so that semantic predicates make sense during  
backtracking.  If you are combining backtracking (syntactic  
predicates) and semantic predicates for tough languages like C++ then  
you must execute actions during the backtrack but then avoid them  
during the parse.

In v2, a quirk of implementation became a feature: init actions for  
rules, subrules were not gated out:

( {always do this} foo | bar )

This was "wrong" in the sense that it executed the action twice.

Anybody have a suggestion?  We have named actions now, perhaps  
instead of @init{...} or plain {...} we do @guessing{...} or something?

Ter