(j3.2006) IMPLICIT TYPE(*)

[Tobias Burnus]

Bill Long wrote:
> Question arose:
> Do we (did we intent to) allow this in the interop TS:
>
> IMPLICIT TYPE(*)  (A)
>
> as long as the only variables starting with the letter A are dummy arguments?

I don't know whether someone intended it, but I do not see a reason not 
to allow "type(*)" in IMPLICIT while allowing other TYPE/CLASS. (That 
implicit typing shouldn't be used is another matter.)


Bader, Reinhold wrote:
> Given that IMPLICIT is followed by a <declaration-type-spec>, the following should be valid in F2008:
[...]
>    implicit class(foo) (a-b)
>    implicit class(*) (c)
[...]
> although only one of the compilers at my disposal accepts this. If this code is valid, I see no reason why IMPLICIT TYPE(*) (...) shouldn't be.

For what it is worth, I just tried "implicit type(*) (a)" successfully 
with one compiler.

Tobias