[MPI3 Fortran] Summary of items on the table

Bill Long longb at cray.com
Mon Apr 28 16:09:22 CDT 2008

Aleksandar Donev wrote:
> On Monday 28 April 2008 13:29, Bill Long wrote:
>> The problem is that the Fortran standard supports those semantics only
>> for a Fortran pointer, not for a C pointer.
> I will take the liberty to interpret the standard words "any pointer 
> associated with the dummy shall become associated with the actual" to apply 
> to C pointers as well. 
I don't think anyone (well, almost anyone) would interpret the Fortran 
standard's use of the well defined term "pointer" to include C 
pointers.  However, I do agree that want you are saying makes 
operational sense.  This seems like a candidate for an interp, since I 
don't see any words in the current standard to support this idea.   Even 
with this fix, there is still the problem of the target attribute and 
all the other baggage that comes with pointers in Fortran.


> Otherwise TYPE(C_PTR) is useless. Of course one can 
> modify the object it points to through the pointer, in C. What else is there?
> Maybe the standard is not super clear about it, but that's a different issue. 
> I believe the intent has been that the C standard describes that, no need to 
> repeat in the Fortran one.
> Best,
> Aleks

Bill Long                                   longb at cray.com
Fortran Technical Support    &              voice: 651-605-9024
Bioinformatics Software Development         fax:   651-605-9142
Cray Inc., 1340 Mendota Heights Rd., Mendota Heights, MN, 55120


More information about the mpiwg-fortran mailing list