On 2/14/07, Arto Bendiken <[EMAIL PROTECTED]> wrote:
On 2/9/07, Daishi Kato <[EMAIL PROTECTED]> wrote:
> On 2/9/07, Arto Bendiken <[EMAIL PROTECTED]> wrote:
> >
> > I simply meant that you can use (define-record) to wrap your own
> > structure around the environment structure, and include a parent
> > pointer (and any other needed goodies) in that.
>
> That doesn't sound so nice.
> I don't want to care about the parent pointer when I call (eval foo env).
> The layered lookup of bindings should be done internally.
Yes, but Chicken's eval won't do that for you (though you can
partially simulate it [1]) - thus, you need your own evaluator,
anyway. I intend to provide a basic one (derived from felix's sandbox
egg) in the contexts egg, eventually.
I see.
Maybe I can create my own evaluator.
> Maybe I could just wait for your egg, and discuss later.
I've just uploaded an initial version to the SVN repository (it's in
contexts/trunk/). I'll be adding unit tests and some documentation [2]
in the next few days - you may want to hold off till those are in
place (though the egg is not that difficult to understand from the
source, either, at a couple hundred LOC).
I think I understand what is going on.
My question would be how much the overhead is.
It should at least be slower than the normal eval.
Ah, I see. That is why you want to redesign it just like the sandbox.
Thanks for your egg anyway, many things for me to help learning.
Daishi
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
