[antlr-interest] disambiguating sempred in a closure?

J Chapman Flack jflack at math.purdue.edu
Thu Dec 27 09:21:36 PST 2007


Terence Parr wrote:
> v3 is way better at this stuff.  You use a gated semantic pred, which 
> always get hoisted into decision:
> 
> prod : something
>       ( ( { muppet }?=> ( KERMIT | FOZZIE )
>         | { turtle }?=> ( LEONARDO | DONATELLO )
>         )
>      something_else
>       )*
>     ;

Ah ... ok (shrugs, downloads v2v3-0.2.tar.gz ... was originally
planning to finish a v2 version first, get that working, and
use that as a regression base for a v3 port, but maybe not if
it means bending over backwards in 2 to do what's easy in 3).

-Chap


More information about the antlr-interest mailing list