[MPIWG Fortran] Provide Fortran datatypes if Fortran bindings not provided?

[Jeff Squyres (jsquyres)]

On Feb 18, 2016, at 4:43 PM, Rolf Rabenseifner <rabenseifner at hlrs.de> wrote:
> 
>> I think JeffH and Jed want (1).
>> I think I want (1), too.
> 
> I see the point.
> I have not a strong opinion and would like to know,
> what the other representatives in the Forum would like to see,
> because
> 
> (2) MPI_DATATYPE_NULL is what they mainly have today.

The standard currently requires that the Fortran datatypes *exist* in mpi.h, no matter what (which seems to be in contradiction to what we stated for the C++ datatypes).  That's why the implementations are currently this way.

> (1) "not existing" would be a real change.

Agreed.

I'll ask Martin if there's plenary time to get a straw vote.

-- 
Jeff Squyres
jsquyres at cisco.com
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/