[Caml-list] ignore semantics

[Caml-list] ignore semantics

[Philippe Veber]

[-- Attachment #1: Type: text/plain, Size: 595 bytes --]

Dear all,

This is a rather minor issue, but I was surprised by the following answers
of the interpreter:

        OCaml version 4.00.1

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

I naively used to think the two expressions were equivalent. Reading
pervasives.mli, I found that ignore is a primitive:

external ignore : 'a -> unit = "%ignore"

which means it is treated as a special case. Just for curiosity, what is
the rationale for the warning in the first case?

Cheers,
  Philippe.

[-- Attachment #2: Type: text/html, Size: 1429 bytes --]

Re: [Caml-list] ignore semantics

[John Whitington]

Hi,

Philippe Veber wrote:
> This is a rather minor issue, but I was surprised by the following
> answers of the interpreter:
>
>          OCaml version 4.00.1
>
> # ignore succ;;
> Warning 5: this function application is partial,
> maybe some arguments are missing.
> - : unit = ()
> # (fun _ -> ()) succ;;
> - : unit = ()
>
> I naively used to think the two expressions were equivalent. Reading
> pervasives.mli, I found that ignore is a primitive:
>
> external ignore : 'a -> unit = "%ignore"
>
> which means it is treated as a special case. Just for curiosity, what is
> the rationale for the warning in the first case?

Imagine you wrote:

ignore (output_something_and_return_something)

When what you wanted was

ignore (ouput_something_and_return_something ())

You'd want the partial application error there, even though you are 
(intentionally) ignoring the return value of the function.

Cheers,

-- 
John Whitington
Director, Coherent Graphics Ltd
http://www.coherentpdf.com/

Re: [Caml-list] ignore semantics

[Philippe Veber]

[-- Attachment #1: Type: text/plain, Size: 1224 bytes --]

Hi John,

That's a good reason indeed, good enough to justify having this special
case semantics. Thanks!

ph.


2012/12/17 John Whitington <john@coherentgraphics.co.uk>

> Hi,
>
>
> Philippe Veber wrote:
>
>> This is a rather minor issue, but I was surprised by the following
>> answers of the interpreter:
>>
>>          OCaml version 4.00.1
>>
>> # ignore succ;;
>> Warning 5: this function application is partial,
>> maybe some arguments are missing.
>> - : unit = ()
>> # (fun _ -> ()) succ;;
>> - : unit = ()
>>
>> I naively used to think the two expressions were equivalent. Reading
>> pervasives.mli, I found that ignore is a primitive:
>>
>> external ignore : 'a -> unit = "%ignore"
>>
>> which means it is treated as a special case. Just for curiosity, what is
>> the rationale for the warning in the first case?
>>
>
> Imagine you wrote:
>
> ignore (output_something_and_return_**something)
>
> When what you wanted was
>
> ignore (ouput_something_and_return_**something ())
>
> You'd want the partial application error there, even though you are
> (intentionally) ignoring the return value of the function.
>
> Cheers,
>
> --
> John Whitington
> Director, Coherent Graphics Ltd
> http://www.coherentpdf.com/
>
>

[-- Attachment #2: Type: text/html, Size: 1907 bytes --]

Re: [Caml-list] ignore semantics

[Francois Berenger]

On 12/17/2012 09:11 PM, Philippe Veber wrote:
> Dear all,
>
> This is a rather minor issue, but I was surprised by the following
> answers of the interpreter:
>
>          OCaml version 4.00.1
>
> # ignore succ;;
> Warning 5: this function application is partial,
> maybe some arguments are missing.
> - : unit = ()
> # (fun _ -> ()) succ;;
> - : unit = ()

Didn't you intended to write

ignore (succ 0);;

(or with any other integer instead of 0) ?

In that case it compiles but I don't get what the code
is intended to do.

Regards,
F.

> I naively used to think the two expressions were equivalent. Reading
> pervasives.mli, I found that ignore is a primitive:
>
> external ignore : 'a -> unit = "%ignore"
>
> which means it is treated as a special case. Just for curiosity, what is
> the rationale for the warning in the first case?
>
> Cheers,
>    Philippe.
>
>