By the way, upgrading JavaCC is taking a whole. First, JavaCC “improved” its
messages, which means that we have a lot of tests to change. I’m trying to
massage the new messages so they look more like the old ones.
Second, the parser generated for Babel is now very different from the other
pars
I agree that this is not something that should prevent us from
upgrading to a newer JavaCC. It's interesting that the generated
parser code is different but this is something that we could
investigate in a follow-up. Even if Babel becomes slower after the
upgrade we cannot stay with the 2006 releas
I am working on https://issues.apache.org/jira/browse/CALCITE-5541
(upgrading JavaCC) and the Babel parser is having problems deducing
whether a keyword is reserved. Investigating this, I took a look at
the generated code, and found something interesting.
Here are the NonReservedKeyWord and NonRes