I've uploaded Marpa-R2-2.084000 to CPAN. Value added this time includes
some speedups to the LATM processing, and portability to platforms where
the C compiler does not support "long long". (Sone MS C compilers
apparently do not have a "long long" type).
The most visible change is to the documentation and the examples in the
synopses. The examples have all been converted to use longest
acceptable token matching (LATM), and the documentation revised so that
LATM is treated as normal, and longest token matching (LTM) is now
treated as an advanced topic. (LTM, for backward compatibility reasons,
remain the default.)
LATM is "smart lexing" -- it means the SLIF uses Marpa knowledge about
the state of the parse to guide the lexer, so that only tokens
acceptable to the parse are looked for. LTM is the tradition because
pre-Marpa parsers weren't smart enough to know which tokens were
acceptable, so they could not guide the lexer. LATM is far easier to
use and more beginner-friendly than LTM. For that matter, it's the
choice experts will almost always make.
This release is, by intention, exactly the same as its release
candidate, except for the change log and the version number. As usual
with these stable, indexed releases, testing is appreciated.
-- jeffrey
--
You received this message because you are subscribed to the Google Groups "marpa
parser" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
