Thanks in advance,
Carter.
I would also be interested in a response to this message :)
up to now, the only things that I saw is that as long as there several
parallel parsers any error simply gets rid of the parser, and there happens
to be a single parser remaining, a usual method is used (as with normal
(LA)LR parsers). Is there any better?
Armel