[antlr-interest] Evaluation boolean expressions

[Xavier Benveniste]

Hi,

First, thanks both of you for your follow-up.

In fact, the atom rule suggested by John B. Brodie was one of some I had
already tried.
And, indeed, this new grammar doesn't suit my expectations.
But I'm not going to give up for all that.
Currently, I'm (re)reading some tutorials from the site and I hope I'll be
able to build it some day.
But I guess this is one huge concern about ANTLR :
- the grammar examples are either very straightforward or quite
incomprehensible for newbies (as I am), and the learning curve seems
endless.

Another point is that the mailing list isn't very helpful : too few people
try to help new ANTLR users (so, I'm very grateful to both of you and some
others).

Actually, I'm pretty sure that what I'm trying is do is a five minutes
working time for an experienced ANTLR and java developer.
Anyway, thanks again.
----- Original Message ----- 
From: "Martin Probst" <mail at martin-probst.com>
To: "Xavier Benveniste" <xavier.benveniste at free.fr>
Cc: "John B. Brodie" <jbb at acm.org>; <antlr-interest at antlr.org>
Sent: Wednesday, July 27, 2005 3:52 PM
Subject: Re: [antlr-interest] Evaluation boolean expressions


> Hi,
>
> > But I'm a little bit confused : does it worth using ANTL if such a
simple
> > algorithm is so hard to develop?
> > Or can't ANTLR evaluates such simple expression as (4 + (5*2) > 15) OR
(4>1)
>
> ANTLR is being used to implement turing complete languages by many
> different projects and many people. It's a great help for parser
> generation, but not necessarily for evaluation. That's of course because
> all languages are quite different in their execution model.
>
> The actual execution is usually what you implement yourself. Treeparsers
> are used by most people to translate ASTs into their own execution model
> (e.g. some kind of expression objects), or for translation (e.g. into
> C).
>
> Martin
>
>
>