Re: [Caml-list] phantom types and identity function

[Ivan Gotovchits <ivg@ieee.org>]

Gabriel Scherer <gabriel.scherer@gmail.com> writes:

> The reason for this behavior is that in the first case, t was inferred
> covariant, while the presence of a constraint disables variance
> inference (in the manual: "otherwise (ie. [...] for non-free
> parameters) the variance must be given explicitly)"). You can make the
> constrained version work with an explicit variance annotation:
>   type +'a t = {x : 'a} constraint 'a = [< `A | `B ];;
>   let f x = (x : [ `A ] t :> [ `A | `B ] t);;

Well, now I understand this. Thanks!

> While we're at it: are you sure you need to play with phantom
> polymorphic variant types now that we have GADTs? I have used phantom
> types in the past, but my personal use cases would be profitably
> rewritten using GADTs. They tend to produce hairy error messages, but
> phantom polymorphic variants are not better in this regard.

I wish I could, but I'm bond to Ocaml 3.11 =(

-- 
         (__) 
         (oo) 
   /------\/ 
  / |    ||   
 *  /\---/\ 
    ~~   ~~   
...."Have you mooed today?"...