[Caml-list] ignore in postfix position

[Caml-list] ignore in postfix position

[Gerd Stolpmann]

[-- Attachment #1.1: Type: text/plain, Size: 971 bytes --]

Hi list,

the "ignore" function behaves specially when applied to a partial function:

# ignore (fun _ -> ());;
Warning 5: this function application is partial,
maybe some arguments are missing.

This very useful warning is not emitted, however, when I switch to a
slight variation, like:

# (fun _ -> ()) |> ignore

I like this post-fix syntax more than the original because "ignore"
isn't then as distracting, but it's a pity that the warning can no
longer be generated. I am wondering whether this behavior is deeply
anchored in the inference engine or whether this is improvable.

Gerd

-- 
------------------------------------------------------------
Gerd Stolpmann, Darmstadt, Germany    gerd@gerd-stolpmann.de
My OCaml site:          http://www.camlcity.org
Contact details:        http://www.camlcity.org/contact.html
Company homepage:       http://www.gerd-stolpmann.de
------------------------------------------------------------



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [Caml-list] ignore in postfix position

[Josh Berdine]

Hi Gerd,

> On Jan 9, 2018, at 11:49 AM, Gerd Stolpmann <info@gerd-stolpmann.de> wrote:
> 
> the "ignore" function behaves specially when applied to a partial function:
> 
> # ignore (fun _ -> ());;
> Warning 5: this function application is partial,
> maybe some arguments are missing.
> 
> This very useful warning is not emitted, however, when I switch to a
> slight variation, like:
> 
> # (fun _ -> ()) |> ignore
> 
> I like this post-fix syntax more than the original because "ignore"
> isn't then as distracting, but it's a pity that the warning can no
> longer be generated. I am wondering whether this behavior is deeply
> anchored in the inference engine or whether this is improvable.

I don't know if treating reverse application similarly to the current treatment of application would be difficult, but as an alternative I tend to put a type ascription on ignore:

# (fun _ -> ()) |> (ignore : unit -> _);;
Error: This expression has type unit -> unit
but an expression was expected of type ('a -> unit) -> 'b
Type unit is not compatible with type 'a -> unit

Cheers, Josh