Cat in Scheme - Cat Language

Message from discussion Cat in Scheme

Ben Chambers

More options Sep 25 2007, 2:27 pm

From: Ben Chambers <bjchamb...@gmail.com>

Date: Tue, 25 Sep 2007 18:27:13 -0000

Local: Tues, Sep 25 2007 2:27 pm

Subject: Re: Cat in Scheme

  Reply to author
  |
  Forward
  | Print
  | View thread
  | Show original
  | Report this message
  | Find messages by this author
  

I have most everything working from cat the specification. I have all
of the level 0 and level 1 primitives working, except for some of the
ones that use while and empty. As far as I can tell there is a
mistake in the specification of either some of the unit tests or some
of the functions. Specifically, if the predicate for a while changes
the stack, these changes should actually affect the stack, right? If
yes, than the empty primitive shouldn't pop the list off the stack (at
least, according to the test cases for a bunch of the things that use
whilene -- or whilene should duplicate the list before calling empty).

Can anyone confirm that this is a bug?

On Sep 25, 8:58 am, "Tom Schouten" <tom.goto10....@gmail.com> wrote:

> hi ben,

> about 'environment-ref':

> in PLT scheme, using 'indentifier-binding', you can make the
> constant/function distinction at compile time. i think this is part of R6RS
> now too. i use this to distinguish between lexical/module/toplevel
> identifiers.

> for my global identifiers i use a separately implemented namespace so it
> won't clash with the scheme toplevel namespace. all names are resolved once
> from a hash table when they are executed using delay/force. all symbols are
> assumed to be functions.

> (it should be possible to use PLT module namespaces only, basicly you get a
> CAT module system for free then.)

> as you allude this makes it possible to expand a code list

> (A 123 C)

> to

> (lambda s (apply C (cons 123 (apply A s))))

> where 'apply' is replaced by 'cons' directly if you know that the atom is a
> constant.

> my code that does this is here:

> http://darcs.goto10.org/brood/brood/rpn-tx.ss

> (http://darcs.goto10.org/brood/is the root of a darcs archive)

> it abstracts the way the global namespace is handled, and defers the real
> parsing using a 'compiler' object: i'm using it for a couple of different
> syntaxes (plain CAT, "writer monad lifted", and compositional forth macros /
> postponed forth)

> cheers
> tom

Reply to author Forward