Re: [Caml-list] Type generalization confusion

[Reed Wilson <cedilla@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1770 bytes --]

But this similar type has no problem:

type 'a field = Int of int
let a = Int 3
let b = Int (1 + 2)

Both a and b receive the same generalized type. Why does the GADT style
prevent this from occuring?

On Mon, May 8, 2017 at 8:49 PM, Gabriel Scherer <gabriel.scherer@gmail.com>
wrote:

> This is the value restriction at work: only values are generalized
> (often of the form "fun x -> ..."), and while (Int 3) is a value, (Int
> (1 + 2)) is not.
>
> On Mon, May 8, 2017 at 11:32 AM, Reed Wilson <cedilla@gmail.com> wrote:
> > I've been working with some of the new features introduced with GADTs,
> and
> > ran into a confusing instance of the "type variables that cannot be
> > generalized" error.
> >
> > The simplified program is as follows:
> > type _ field = Int : int -> 'a field
> > let a = Int 3
> > let b = Int (1 + 2)
> > let inside = 1 + 2
> > let c = Int inside
> >
> > ocamlc -i returns:
> > type _ field = Int : int -> 'a field
> > val a : 'a field
> > val b : '_a field
> > val inside : int
> > val c : 'a field
> >
> > with b remaining non-generalized. The problem I'm having with it is that
> the
> > type variable doesn't depend on the value at all, so I don't see how that
> > can prevent generalization.
> >
> > Also, the manual says the reason some types aren't generalized is due to
> > "polymorphic mutable data structures". Nothing I created is mutable, so
> why
> > was generalization turned off in the first place?
> >
> > Finally, I'm confused why separating the function from the definition is
> > enough to fix this; c is generalized simply by defining 1+2 in a separate
> > value (which must be global, apparently).
> >
> > Thanks,
> > Reed Wilson
> >
> > --
> > ç
>



-- 
ç

[-- Attachment #2: Type: text/html, Size: 2575 bytes --]