> In both cases, the reason is that exhaustiveness checking is based
> only on the type of the expression, not on its value.

Is it correct to explain let-binding for this syntax

let <pattern> = <expression1> in <expression2>

as the type inference mechanism computes type T1 of <pattern> and type
T2 of <expression1> inferring the type of variables in <pattern> using
T2, and then T1 must be the same as T2?

•  4
  New Members

>
> In both cases, the reason is that exhaustiveness checking is based
> only on the type of the expression, not on its value. So the value
of
> g() is [], but its type is 'a list, which includes values other than
> []; the value of h() is 0, but its type is int, which includes
values
> other than 0.
>
> However, the type of f() is unit, which includes only the value ().
> Therefore the match is exhaustive.
>
> Does that help?

Yes it does help. Thank you Jon and Jake.

Warning P: this pattern-matching is not exhaustive.
Here is an example of a value that is not matched:
_::_

The warning says _::_ is not matched, i.e., the values represented by
_::_ (that might be the value of f() because of its type) are not
matched by [].

•  4
  New Members

> > The value of g() is [].
>
> The value of g() is not available to the static type checker at
compile time.
>
> > Why doesn't [] match g()?
>
> It does.
>
> You are not seeing a run-time match failure. You are seeing a
warning from the
> static type checker at compile time stating that it has not been
able to
> prove that it is safe to remove run-time checks from the generated
code.
>
> However, your use of the top-level makes the distinction difficult
to see. Put
> the program in a file:
>
> $ cat >a.ml
> let g() = []
> let [] = g()
>
> Compile it and note the compile-time warning:
>
> $ ocamlc a.ml -o a
> File "a.ml", line 2, characters 4-6:
> Warning P: this pattern-matching is not exhaustive.
> Here is an example of a value that is not matched:
> _::_
>
> Run it and note that the match does not actually fail:
>
> $ ./a
> $
>
> --
> Dr Jon Harrop, Flying Frog Consultancy Ltd.
> http://www.ffconsultancy.com/?e
>

Ok, I did that. Very helpful. Thank you.

•  4
  New Members