[Bug fortran/47844] Array stride ignored for pointer-valued function results

["burnus at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47844

--- Comment #7 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-10-05 15:04:02 UTC ---
(In reply to comment #6)
> We could fix this in 4.7 by adding a sm field to array descriptors.
> If we added the sm field after the dimension array, we would not
> damage the exiting API.

I think that one can create cases, where it fails - for instance when combining
a GCC 4.7 compiled procedure with a GCC 4.6 compiled procedure. For instance
some combination of
  gcc_4_6_compiled()
    type(t), pointer :: A(:)
    call assign_in_gcc_4_7_compiled (A)
might fail, if one accesses "a.sm" in "assign_in_gcc_4_7_compiled". If we can
rule out that such issues occur for currently working code, we can do so.

You also have a small ABI issue with allocatable coarrays: In GCC 4.6 or 4.7,
using -fcoarray=single, one has for
  integer, allocatable :: A(:)[:]
  allocate (A(1)[3:*])
an extra dimension triplet for coarrays:
        a.dim[0].lbound = 1;
        a.dim[0].ubound = 1;
        a.dim[0].stride = 1;
        a.dim[1].lbound = 3;
Given that coarrays are not yet widely used and that coarrays in 4.6 had some
issues, it might be a smaller problem.