Re: [Chicken-users] can't get 'the' to work

[Jörg F . Wittenberger]

On Mar 25 2013, Moritz Heidkamp wrote:

> >  (let ((foobar (the (or null (list-of number)) '())))
> >
> > Warning: in toplevel procedure `foo#bar':
> >  expression returns a result of type `null', but is declared to return
> > (list-of number)', which is not a subtype
>
> It seems you can trick the compiler by using something like this:
>
>  (the (list-of number) (apply list '()))
>
> I'm not sure whether that's how THE BOSS intended the -strict-types
> option to be (ab)used though :-)

Thanks for this idea!

> Better follow Peter's advice of getting
> rid of the mutation.

Hm, the example is stripped down from some argument parsing.

Given N many bindings, wouldn't it be it a bit expensive to
unroll the simple case-dispatch into N eq?-Tests?  Let alone
that the code would become quite unreadable.  (Though this
could eventually be handled by macro expansion.)

So I'd better resort to some compiler-trickery to get the
explicit "the" declaration to work.

Thanks again

/Jörg



........