[J3] F2023:C7108 questions

[Steven G. Kargl]

Malcolm,

Thanks for the clarification.  As it's a numbered
constraint, is a Fortran processor required to 
detect and report this disambiguating constraint?

-- 
steve

On Sun, Mar 24, 2024 at 09:27:09AM +0900, Malcolm Cohen wrote:
> Hi Steve,
> 
> This is what we call a disambiguating constraint. It makes no requirement on
> the user program. Its only effect is that when something could be
> interpreted as a structure constructor or a generic procedure reference, the
> generic procedure reference takes precedence.
> 
> And that resolution is determined by the type, kind, rank, and procedureness
> of the arguments.
> 
> Cheers,
> -- 
> ..............Malcolm Cohen, NAG Oxford/Tokyo.
> 
> -----Original Message-----
> From: J3 <j3-bounces at mailman.j3-fortran.org> On Behalf Of Steven G. Kargl
> via J3
> Sent: Sunday, March 24, 2024 6:11 AM
> To: J3 Fortran <j3 at mailman.j3-fortran.org>
> Cc: Steven G. Kargl <kargl at uw.edu>
> Subject: [J3] F2023:C7108 questions
> 
> All,
> 
> I am seeking clarification on the intent of F2023:C7108.
> This constraint can be traced to F2003:C489.  Consider the
> following code:
> 
>    !
>    ! Fortran 2023:C7108 (R756) If derived-type-spec is a type name
>    ! that is the same as a generic name, the component-spec-list
>    ! shall not be a valid actual-arg-spec-list for a function
>    ! reference that is resolvable as a generic reference to that
>    ! name (15.5.5.2).
>    !
>    module foo
>       implicit none
>       type :: params
>          real :: x
>          real :: y
>       end type params
>       interface params
>          module procedure default_params
>       end interface
>       contains
>          function default_params(x, y) result(this)
>             real, intent(in) :: x, y
>             type(params) :: this
>             this%x = x * 2 
>             this%y = y * 2
>          end function
>    end module
> 
>    program chk_type_define
>       use foo
>       implicit none
>       type(params) :: p
>       !
>       ! Is this a structure constructor or generic function reference?
>       !
>       p = params(3.0, 2.0)
>       if (p%x /= 3 .or. p%y /= 2) stop 1
>    end program chk_type_define
> 
> It seems to me that the constraint is to ensure that no ambiguity
> occurs with the assignment in 'p = params(3.0, 2.0)'.  Is this 
> correct?  If so, if the names of the components or dummy arguments
> are changed, do the use of keywords remove the ambiguity?  That is,
> 'p = params(a = 3.0, 2.)' would be a reference to the generic function
> if the first dummy argument of 'default_params' is 'a'.
>   
> -- 
> Steve

-- 
Steve
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3950 bytes
Desc: not available
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20240323/68b863d0/attachment.bin>