An initial preview release of Rapid Scheme is available at:

  https://gitlab.com/nieper/rapid-scheme


Rapid Scheme is an implementation of the Scheme programming language
as described by the R7RS.

As Rapid Scheme is written in portable Scheme code itself, it depends
on the existence of a working implementation of the R7RS.

One may ask what the purpose of an implementation of R7RS is if one
already needs an implementation of R7RS in order to use it.

There are actually several reasons for the existence of Rapid Scheme:

  * Rapid Scheme outputs any error detected during expansion in a
    standard format (which is also used by gcc and consumable by
    Emacs) with exact line and column numbers of the erroneous
    construct including the context due to library imports, file
    inclusion and macro expansion.

  * Rapid Scheme tries to detect as many errors as possible. In
    particular, it helps writing portable code.

  * Rapid Scheme can and will be used to prototype syntactic
    extensions to the R7RS. This may come handy in sample
    implementations of SRFIs, in particular those envisaged for
    inclusion in R7RS-large.

  * Compiler writers can use Rapid Scheme as a front-end of a
    standalone implementation of R7RS. In particular, it should be
    relatively easy to combine an existing implementation of an
    earlier report such as R5RS with Rapid Scheme to produce a
    stand-alone implementation of R7RS.

  * Rapid Scheme itself may one day receive a back-end and become
    a stand-alone implementation.

  * Rapid Scheme is a proof that one can write non-trivial programs
    with the language as described by the R7RS. The development of
    Rapid Scheme helped to find a number of bugs in existing
    implementations of the R7RS, which are now fixed.

  * Each new implementation of R7RS is a proof that the R7RS has
    fulfilled its goals.


Currently, I am using Larceny during the development of Rapid Scheme.
I have also used Chibi Scheme, but for the Rapid Scheme code it is
unfortunately many orders of magnitude slower than Larceny. I am
thankful to the authors of the two implementations, especially to
Will Clinger and Alex Shinn, for making their implementations Free
Software, for their responsiveness to comments, questions, and bug
reports, and for their contributions to the prosperity of the Scheme
community.

(Due to a few bugs in Larceny's 0.99 release, one has to use a
development version to run Rapid Scheme. I am hoping that there will
soon be a release of Larceny that incorporates the latest bug fixes.)


Warning: Rapid Scheme has been announced as a *preview* release. As
such, it is buggy for sure. Any bug report or constructive criticism
is much appreciated.

Interesting! I'm trying to play around this and found a malformed and-let*
in the unpack-formals of primitive-environment.scm. The body of second
and-let* is inside of the variable declarations.

Cheers,

Thanks for catching this. The implementations of and-let* I have been using are so liberal that they didn't report the syntax violation (and did the right thing with the code).

I have pushed a corrected version to the GitLab repository.

--

Marc

Marc Nieper-Wißkirchen wrote:

> (Due to a few bugs in Larceny's 0.99 release, one has to use a
> development version to run Rapid Scheme. I am hoping that there will
> soon be a release of Larceny that incorporates the latest bug fixes.)

Prebuilt versions of a more-or-less current development version
of Larceny can be downloaded from

    http://www.cesura17.net/~larcenists/Nightly/

Will

> Prebuilt versions of a more-or-less current development version
> of Larceny can be downloaded from
>
>     http://www.cesura17.net/~larcenists/Nightly/
>
> Will

That's great! Thank you very much. I am going to update Rapid's documentation soon.

Marc

For those who want to follow development of the Rapid Scheme frontend, I have set up a simple web page with a blog, which is hosted at:

https://www.rapid-scheme.org

Stay tuned. :-)

Marc