caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Aleksey Nogin <nogin@cs.caltech.edu>
To: Nicolas Pouillard <nicolas.pouillard@gmail.com>
Cc: OCaml Mailing List <caml-list@yquem.inria.fr>,
	MetaPRL Developers <metaprl-devel@metaprl.org>
Subject: Re: [Caml-list] [Camlp4] Beta-release
Date: Tue, 16 May 2006 13:16:31 -0700	[thread overview]
Message-ID: <446A331F.4000205@cs.caltech.edu> (raw)
In-Reply-To: <cd67f63a0605150441l60ad0e10ra4502fa8c98082f1@mail.gmail.com>

I have a question/feature request.

For our camlp4 usage (in MetaPRL theorem prover - http://metaprl.org/ ), 
we need to:
- extend the OCaml grammar
- refer to a "state" when parsing the extensions.

Currently we have to put the state in a global ref (yuck!), which causes 
all kinds of difficulties.

Is this something that can be done better in your reimplementation? Can 
something like this be added to your reimplementation?

Basically, I would really like to see camlp4 help with the following:

1) Defining parsers that depend on a shared state that is passed around.

Currently one can either
a) use a global pointer ref, or
b) manually define the productions of "state -> ..." type and manually 
pass the state around.

(a) is obviously bad, but even (b) is not only a lot of extra typing, 
but also means that the execution of productions will be delayed until 
the whole thing is parsed, which would mean that when something goes 
wrong, you would often get a campl4 parsing error instead of a (usually 
more meaningful) error generated by the production code.

2) Allow passing shared state even for grammars that extend things like 
OCaml grammar.

3) (completely unrelated) I wish there was a way to tell camlp4 to 
detect and report conflicts!

P.S. The bulk of our camlp4 code is in 
http://svn.metaprl.org/viewcvs/mojave/metaprl/filter/filter/term_grammar.ml 
(the shared state is hidden in a way here - the grammar is created in a 
functor where the input module provides functions like mk_opname; when 
the functor will be instantiated, the functions would be made dependent 
on a number of global refs) and in 
http://svn.metaprl.org/viewcvs/mojave/metaprl/filter/filter/filter_parse.ml 
(here we actually extend the OCaml grammar, using some of the stuff from 
term_grammar.ml in that extension).

If you are interested in downloading MetaPRL sources to look at this 
stuff in more detail, please see the download instructions at 
http://metaprl.org/install.html

-- 
Aleksey Nogin

Home Page: http://nogin.org/
E-Mail: nogin@cs.caltech.edu (office), aleksey@nogin.org (personal)
Office: Moore 04, tel: (626) 395-2200


  parent reply	other threads:[~2006-05-16 20:16 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-05-15 11:41 Nicolas Pouillard
2006-05-16 14:55 ` [Caml-list] " Alain Frisch
2006-05-16 20:16 ` Aleksey Nogin [this message]
2006-05-17  7:52   ` Nicolas Pouillard
2006-05-22 22:39     ` Aleksey Nogin
2006-05-18 14:36 ` Stefano Zacchiroli
2006-05-18 14:41   ` Nicolas Pouillard

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=446A331F.4000205@cs.caltech.edu \
    --to=nogin@cs.caltech.edu \
    --cc=caml-list@yquem.inria.fr \
    --cc=metaprl-devel@metaprl.org \
    --cc=nicolas.pouillard@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).