(j3.2006) Binding label consistency in module procedures

[Van Snyder]

I believe it would be permitted for the procedure to have both binding
names.  Didn't we have an interp that allows an external procedure that
is declared with interface bodies in different scoping units to have
different binding labels?

On Mon, 2016-05-23 at 14:32 +0000, Lionel, Steve wrote:
> Consider:
> 
>  
> 
> module U610575A
> 
> interface
> 
> module subroutine sub1 () bind(C,name='SUB1')
> 
> end subroutine sub1
> 
> end interface
> 
> end module U610575A
> 
>  
> 
> submodule(U610575A) U610575AS
> 
> contains
> 
> module subroutine sub1 () bind(C,name='NOTSUB1')
> 
> end subroutine sub1
> 
> end submodule U610575AS
> 
>  
> 
> The standard says (F2008 12.4.3.2p7) for a specific interface that
> ?the characteristics shall be consistent with those specified in the
> procedure definition??. However, the list of procedure characteristics
> (12.3.1) includes only ?whether it has the BIND attribute? and not the
> value of NAME= in a language-binding-spec.
> 
>  
> 
> I don?t want to see the standard trying to enforce binding label
> consistency in general, but I wondered if submodules should be a
> special case. If a program tried to call sub1, it would probably fail
> to link because of the name mismatch. Thoughts?
> 
>  
> 
> Steve Lionel
> 
> Intel Developer Support
> 
> Merrimack, NH
> 
>  
> 
> 
> _______________________________________________
> J3 mailing list
> J3 at mailman.j3-fortran.org
> http://mailman.j3-fortran.org/mailman/listinfo/j3