[antlr-interest] Evaluator bailing out early?

[Olivier Lefevre]

I was using a test rig, not ANTLRWorks, but yes I was calling prog()

> If you add an EOF to the prog rule you should get an error as the
> rule will be forced to handle ALL the tokens for a match to be made.

I see. That is sorta like bracketing regex strings with '^' and '$' to 
force the entire input string to be matched, correct?

> I often make a special rule, for example test_rule, where I add the EOF token.

Good idea. That works for test rigs, too. Why would you _not_ want to match 
the entire input, though?

Thanks again,

-- O.L.