[racket] Evaluating code written in non-SEXP language

From: Dmitry Pavlov (dpavlov at ipa.nw.ru)
Date: Wed Sep 11 14:33:26 EDT 2013

I think, you may simply add #lang into input-port when reading the file.
>

Good point, but how do I (dynamic-require) a port?
I could make a temporary file with the #lang line preceding
the contents of the original file, but then I lose all the references
to the original file in the parser error messages etc.

In case you referred to my original solution with
(make-evaluator), which I do not like: it works
without #lang all right.

Regards,

Dmitry



>
> Среда, 11 сентября 2013, 12:08 -05:00 от Robby Findler <
> robby at eecs.northwestern.edu>:
>
>   Generally speaking, Racket is set up to work well with the #lang lines,
> but not as well without. Lots of things get easier if they are there.
>
> Robby
>
>
> On Wed, Sep 11, 2013 at 11:57 AM, Dmitry Pavlov <dpavlov at ipa.nw.ru<https://e.mail.ru/sentmsg?mailto=mailto%3adpavlov@ipa.nw.ru>
> > wrote:
>
> Hi Joe,
>
> Thank you. Actually my goal is to run files that do
> not contain the #lang declaration--only pure non-Racket
> non-sexp code. Is there a way to (dynamic-require)
> a module with an externally preset language? I do not know.
>
> I tried your approach anyway. It works well itself,
> but I have failed (as with all the other approaches)
> to make a standalone executable from my "interpreter"
> program. I wonder if anybody tried that with programs
> of this kind.
>
> Best regards,
>
> Dmitry
>
>
>
>
>
> On 09/11/2013 05:46 PM, Joe Gibbs Politz wrote:
>
> Hi Dmitry,
>
> I've been using `dynamic-require` for Pyret, a non-SEXP based
> language.  For example, here's part of our command-line interface:
>
> https://github.com/brownplt/**pyret-lang/blob/master/src/**
> cmdline.rkt#L116<https://github.com/brownplt/pyret-lang/blob/master/src/cmdline.rkt#L116>
>
> You can ignore the parameterization under `check-mode`, which is Pyret
> specific.  The point is that `pyret-file` is a filename handed in as a
> command-line argument, and we load the module via `dynamic-require`.
> `dynamic-require` does the work of using the reader/environment
> specified for the #lang declared in the file, and gets things like
> relative module imports right, too (at least I haven't been able to do
> anything too unexpected).
>
> This should work for any language that you want to build a
> file-accepting interface for, not just non-sexp ones.  I'm not sure if
> this is the preferred style or not; if it's not, Pyret should probably
> use whatever better method there is as well.
>
> Cheers,
> Joe P
>
>
>
> On Tue, Sep 10, 2013 at 11:14 AM, Dmitry Pavlov <dpavlov at ipa.nw.ru<https://e.mail.ru/sentmsg?mailto=mailto%3adpavlov@ipa.nw.ru>>
> wrote:
>
> Hello,
>
> I have developed a non-SEXP DSL called SLON, and it works
> great in DrRacket via "#lang slon" declaration.
>
> I have properly set up a lexer, a parser which returns datum
> or syntax objects, slon/lang/reader.rkt,
> slon/lang/lang-into.rkt, slon/lang/confugure-runtime.**rkt
> and all the other stuff.
>
> Now I need it to work as a stand-alone interpreter
> which runs files specified by user (outside of DrRacket).
> How do I do that? The only thing I was able to come up
> with is using make-evaluator for racket/sandbox:
>
> (require slon/slon-parser
>           racket/sandbox)
>
> (call-with-input-file*
>     "myprogram.slon"
>     (lambda (port)
>       (parameterize ((sandbox-input (current-input-port))
>                      (sandbox-output (current-output-port)))
>         (call-with-trusted-sandbox-**configuration
>          (lambda ()
>            ((make-evaluator 'slon/slon-language
>                             #:allow-read '("/"))
>             (slon-read port))))))))
>
> Is it the correct way to create a stand-alone interpreter?
> Because I do not really need a "sandbox" of any kind,
> and also I was forced to specify #:allow-read '("/")
> because the SLON programs do read arbitrary files.
>
> Also I am having a lot of trouble in making that
> a stand-alone executable; but I will postpone
> discussing it because I am not sure whether
> I chose the right way to do the whole thing.
>
>
> Best regards,
>
> Dmitry
> ____________________
>   Racket Users list:
>   http://lists.racket-lang.org/**users<http://lists.racket-lang.org/users>
>
>
>
>
> ____________________
>  Racket Users list:
>  http://lists.racket-lang.org/**users <http://lists.racket-lang.org/users>
>
>
>  ____________________
>   Racket Users list:
>   http://lists.racket-lang.org/users
>
>
>
> --
> Roman Klochkov
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20130911/5798fa95/attachment-0001.html>

Posted on the users mailing list.