[J3] Ambiguous?

[Bader, Reinhold]

Hi Van, 

due to the auto-targetting feature, it actually is ambiguous. TC1 (f08/0001) added the words
"and not the INTENT(IN) attribute" at the end of the sentence that you are referring to to deal with the fallout of that feature.

Quite coincidentally, I've been looking at exactly that para earlier today, because I intend to hand in a F202X feature
request that actually also deals with this side effect ...

I've attached a draft of this and would be interested in feedback whether this is worth pursuing.

Cheers
Reinhold


> -----Ursprüngliche Nachricht-----
> Von: J3 <j3-bounces at mailman.j3-fortran.org> Im Auftrag von Van Snyder via
> J3
> Gesendet: Montag, 6. August 2018 22:21
> An: j3 <j3 at j3-fortran.org>
> Cc: Van Snyder <Van.Snyder at jpl.nasa.gov>
> Betreff: [J3] Ambiguous?
> 
> According to 12.4.3.4.5p3, third bullet in the list, in the 2008 standard, I
> believe the interface in the following module should not be ambiguous.  But
> none of my processors accept it.  Is it standard conforming, or have I not
> understood 12.4.3.4.5p3?
> 
> module Ambiguous_m
> 
>   implicit NONE
>   private
> 
>   public :: Ambiguous
> 
>   interface Ambiguous
>     module procedure Ambiguous_A, Ambiguous_P
>   end interface Ambiguous
> 
> contains
> 
>   subroutine Ambiguous_A ( This )
>     class(*), intent(in), allocatable, target :: This(:)
>   end subroutine Ambiguous_A
> 
>   subroutine Ambiguous_P ( This )
>     class(*), intent(in), pointer :: This(:)
>   end subroutine Ambiguous_P
> 
> end module Ambiguous_m
> 
> I also tried it with REAL instead of CLASS(*) arguments and got the same
> complaints.
> 

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: persistency.txt
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20180806/25b5bcb9/attachment-0001.txt>