[plt-scheme] minor glitch in Lazy Scheme?

On 1/13/07, Eli Barzilay <eli at barzilay.org> wrote:
> On Jan 13, Robby Findler wrote:
> > Note that replacing cons in that manner is also less than optimal,
> > since it means that every cons has to have that contract.
>
> I'm not sure I'm following you.  Are you saying that every cons cell
> (whether generated by the `cons' function or not) must be following
> the same contract?  If this is correct, then this is what I meant when
> I said that it uses the fact that this
>
>   (define (list? x)
>     (or (null? x) (pair? x)))
>
> is a valid definition of `list?' (uses the same global cons
> restriction).
>
>
> > It is possible to do better (but you have to take over the
> > define-struct that generated cons).
>
> Are you suggesting a bit that tells you whether this cons is a proper
> list or not?  If so, wouldn't that bit be the same as adding a second
> kind of cons?

Not exactly. If you take over the define-struct you can make the
selectors recognize all the different cons's and still have different
contracts on them all.  That's the gist of the idea.

Check out define-contract-struct in the contract library. That's what
I'm talking about. It a way to get the contract on list-based cons in
some contexts but other contracts in other contexts.

Robby