Re: New Yacc/Bison like parser generator for C#

"Eric" <englere.geo@yahoo.com>
22 Sep 2005 23:42:34 -0400

          From comp.compilers

Related articles
New Yacc/Bison like parser generator for C# w.kelly@qut.edu.au (Wayne Kelly) (2005-09-17)
Re: New Yacc/Bison like parser generator for C# englere.geo@yahoo.com (Eric) (2005-09-22)
| List of all articles for this month |
From: "Eric" <englere.geo@yahoo.com>
Newsgroups: comp.compilers
Date: 22 Sep 2005 23:42:34 -0400
Organization: http://groups.google.com
References: 05-09-068
Keywords: tools, C#
Posted-Date: 22 Sep 2005 23:42:33 EDT

Very interesting use of the new generics in C# 2.0. I'm also familar
with the excellent Gardens Point Component Pascal project. I was just
reading John Gough's book the other day.


John mentioned that he likes the Eli parser generator, but oddly he
said he used a hand-tweaked output from Coco/R to develop the parser
for GPCP. I am guessing that's because he wasn't using C++ (I guess Eli
is for C++ only).


I also have Pat Terry's book "Compiling in C# and Java", which is also
excellent.


I assume you're familiar with Coco/R, which has a C# version (among
others) and produces LL(1) parsers.


I'm guessing you probably prefer the LALR parsers in general. Do you
find it easier to work with LALR grammar specificiations, or was this
selected because you already have a LALR grammar for your desired
target language (Ruby)?


Did you also consider the use of the Gold parser generator (also LALR)?


I'm trying to formulate a strategy for choosing between LALR and LL(1),
or even LL(k) as implemented in ANTLR. It's a big field with a lot of
choices, but they all require a great deal of effort to get a deireable
outcome so it seems wise to do some general research up front.


Eric


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.