Very nice, adding
(: get-user-var (case-> ...))
indeed does the trick. So I can use `case->` to do overloading (if that's
what it is called), by defining different functions for different type
signatures, and then put them together into a single function via a cond.
E.g.:
(: int-add (-> Integer Integer Integer))
(define (int-add x y)
(+ x y))
(: string-add (-> String String String))
(define (string-add x y)
(string-append x y))
(: int-string-add (case->
(-> Integer Integer Integer)
(-> String String String)))
(define (int-string-add x y)
(cond [(integer? x) (int-add x y)]
[(string? x) (string-add x y)]
[else
(error "Not a string or int" x)]))
I am wondering if there is a way to avoid the final definition of
`int-string-add`, by simply defining the same function twice (or more) with
different type signatures, and then when I call the function, which code
gets called depends on the type signature? I would be surprised if this
existed in Typed Racket right now, but it would be neat and good to know if
there is a more idiomatic/built-in way than what I do above, ie just write:
(: special-add (-> Integer Integer Integer))
(define-special (special-add x y)
(+ x y))
(: special-add (-> String String String))
(define-special (special-add x y)
(string-append x y))
I guess it would be pretty hard, since the type signature has to be coupled
more tightly to a specific function than TR requires right now, i.e. I'd
have to write (define-special (special-add [x : String] [y : String]) to
make the link explicit. Anyway, if it's not possible right now, totally
fine.
Regarding refinement types for symbols, I have no idea how useful it would
be, as the only example I came up with was the above one, which I think is
better dealt with in a different way.
Cheers,
Marc
On Sunday, February 23, 2020 at 6:33:33 PM UTC+1, Ben Greenman wrote:
>
> Try this case-> type instead:
>
> (case->
> (-> Integer 'name Symbol)
> (-> Integer 'age Integer)
> (-> Integer Symbol (U Symbol Integer)))
>
>
> I don't know what it would take to add refinements for symbols, but
> that might be useful. Andrew Kent's dissertation may have some
> pointers:
>
> https://pnwamk.github.io/docs/dissertation.pdf
>
--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/racket-users/98151240-f930-4dee-a8de-45f8ce6797ac%40googlegroups.com.
