Eric Mahurin wrote:
> I tried this parser, but couldn't get it to work with the String interface
> of the quize testbench.  Using an IO/stream is how a parser should work,

Not a packrat parser. If you have memory to memoize each grammar rule
attempted at each byte-offset of the input, you have memory to have all
the input in memory at once, and it's more efficient to do so.

Obviously a simple adapter can fix that, but steve's doesn't, and
neither does mine (emailed to JEG before the 24 hours - I'll post
it soon). The adapter is needed because the parse() method expected
is different from the one that Treetop's parsers provide.

Clifford Heath.