From: Eric Mahurin <eric.mahurin@...>
Date: 2008-09-07T11:20:59+09:00
Subject: [ANN] Grammar 0.8

------=_Part_60764_15395766.1220754420270
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Since I was giving a presentation about Grammar at the Lonestar Ruby
Conference (which ended today), I thought it was about time to release
another version of Grammar.  I posted it yesterday here:

http://rubyforge.org/projects/grammar/

and the presentation is here (before a few modifications):

http://grammar.rubyforge.org/0.8/grammar_lonestar2008.swf

Here are some release notes:

* second complete rewrite since Grammar 0.5
* BNF-like grammar written directly in Ruby
* LL(1) parser
* separation of Grammar and parser generator ("engine")
* engines could generate a parser in a particular target language or do
something else with the Grammar tree
* current engines: Ruby (generate stand-alone very flat Ruby parser), Ruby0
(directly parse while traversing Grammar), Ruby2CExt (Ruby + convert to C
using ruby2cext)
* lexer and parser grammars are specified in exactly the same way - one
parses characters and the other tokens
* can write a lexer-free parser
* may define tokens however is appropriate
* methods for a lexer (or just token) Grammar to a parser Grammar -
including multi-threading
* arbitrary lookahead can be achieved with the backtrack method
* tail-call optimization of recursion (right recursion)
* directly handles left recursion (historically only doable with (LA)LR
parsers only).
* very very fast - on par with the best hand crafted/tuned recursive descent
LL(1) or Regexp (LL(*)) parsers.  100X faster than many other parser
generators.  And that is before Ruby2CExt.

Comments and feedback are quite welcome.

Eric

------=_Part_60764_15395766.1220754420270--