On Tue, Feb 24, 2015 at 11:35 PM, Matt Oliveri <[email protected]> wrote:
> On Tue, Feb 24, 2015 at 10:14 PM, Jonathan S. Shapiro <[email protected]> 
> wrote:
>> If I am not mistaken, Keeans' enumeration of possible arities is intended to
>> be a deep enumeration. I think he's mistaken, because the return type 'r can
>> unify with a function type and thereby extend the arities. But assuming that
>> does not occur, his enumeration seems reasonable.
>
> The enumeration of deep arities was mine. I only took the format from
> Keean. In my understanding, the reason why an abstract return type
> doesn't change things is because of the maximum arity, which it seems
> we also disagree about.

No, that's wrong, sorry. I got the arity enumeration right though
because I said r was a metavariable for a concrete non-function type.
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to