[MPI3 Fortran] J3 / etc.
nmm1 at cam.ac.uk
Wed Oct 13 05:13:08 CDT 2010
On Oct 12 2010, Rolf Rabenseifner wrote:
>has the C-binding of Fortran CHARACTER*(*) dummy arguments
>has also a chance to become part of Fortran 2013 by being
>included in a type 2 TR that passes already summer 2011?
Dunno, but it is something I have been badgering about for some time.
I hope so.
>I ask because this is another major open gap to have a full BIND(C)
>interface for all MPI routines.
That statement is a great help. My problem previously was that I
couldn't point to a reference saying that MPI badly needed it, so it
was ruled out of scope.
>(The last gap will be to look at our callback functions.)
And therein hangs a tale :-(
There are some evil issues there, to do with scoping etc. I am trying
to get the TR enhanced enough to be a starting point, but it doesn't
help that the current interoperability section and C standard are very
bad on the point. The main body of Fortran is merely weird and fiendishly
complicated (because of much history) ....
More information about the mpiwg-fortran