https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113997
Thomas Koenig <tkoenig at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords|needs-stdcheck |
--- Comment #8 from Thomas Koenig <tkoenig at gcc dot gnu.org> ---
(In reply to Thomas Koenig from comment #7)
> Hm, I am not actually sure that the code is correct.
> I think I will ask on the J3 mailing list.
Forget the last comment, I was getting confused about the
two 'x' in the test case; this test case fails as well, and should
not:
subroutine acc_attach(x)
use iso_c_binding, only : c_loc
implicit none (external, type)
type(*), dimension(..), target :: x
interface
subroutine acc_attach_c(a) bind(C, name="acc_attach")
use iso_c_binding
type(c_ptr) :: a
end subroutine
end interface
call acc_attach_c(c_loc(x))
end
which gives
asdf.f90:8:5:
8 | subroutine acc_attach_c(a) bind(C, name="acc_attach")
| 1~~~~~~~~~~~~~~~~~~~~~~
Warning: Interface mismatch in global procedure 'acc_attach_c' at (1): Type
mismatch in argument 'a' (TYPE(c_ptr)/TYPE(*))
It seems that the code checking for the global procedure looks through
the c_loc() call, and it should not.
