Re: [boost] Re: Formal Review: The Spirit Parser Framework.

From: Joel de Guzman (djowel_at_[hidden])
Date: 2002-10-19 19:44:05

----- Original Message ----- 
From: "Carl Daniel" <cpdaniel_at_[hidden]>

> Summary:  I think that Spirit should be accepted into Boost.

Thanks!

> Comments:
> 
> I've already posted a variety of documentation and code nits - I won't
> repeat those.
> 
> I will echo the concern of others that the documentation, while excellent,
> is also insufficient.  A complete reference section is absolutely necessary,

Agreed.

> along with some significant discussion of theory of operation, and tradeoffs
> between all the different mechanisms. e.g. When would I use a
> rule/subrule/grammar?  Why is one better than the other?  Why do they all
> exist?

Understood.

> It's been many years (15?) since any of my projects have needed a parser,
> but I was able to build a parser for a moderately complex Pascal-like "4GL"
> in about 3 hours, using one of the samples as a guide.  I expect a second
> go-around would cut that development time in half (or better).

I'd be really interested with your Pascal-like "4GL" language.
I'm also interested with Doug's review experiments. I'm
sure these will be nice additions to the samples if either or both
of you make them available.

> One concern I have, which I belive is based mostly in ignorance, is the
> ability of a Spirit-generated parser to perform error recovery and issue
> good error messages.  I didn't have time to experiment with the exceptions
> and guards mechanism, so perhaps my concern is unfounded.

This is a strong point of LL (including RD). I'll try to make a strong
case in this regard coupled with examples and tests.

> Great work Joel, Dan, Harmut, Martin, JCAB, and everyone else!

And thanks to you too.

Regards,
--Joel