(j3.2006) C-interoperable interface bodies

[Van Snyder]

On Fri, 2007-02-23 at 15:27 -0600, Bill Long wrote:

> > I wrote:
> > I don't have a compiler that does BIND yet, but in the ones I have, when
> > I put an interface body in a module, they don't produce a symbol
> > definition:
> > 
> > module A
> > 
> >   interface
> >     subroutine P_A ( X )
> >       real :: X
> >     end subroutine P_A
> >   end interface
> > 
> > end module A
> > 
> > #lf95 -c interfaceTest.f90
> > # nm interfaceTest.o
> > 00000000 T a_
> 
> I get:
> 
> % nm van.o
> 0000000000000000 t %%COMP_UNIT_BEGIN
> 0000000000000008 t %%FUNC_BEGIN_1
> 0000000000000080 t %%FUNC_END_1
> 0000000000000080 T __p_a_
> 0000000000000080 T a_
> %

Assuming the symbol __p_a arose from the interface body for P_A, I think
your compiler has a bug in it.  There's no definition of P_A here, so I
don't think there ought to be a definition of __p_a in the .o file.
There is a declaration that an external subroutine known by the Fortran
name P_A might exist somewhere in the program -- but it's not even
required to exist so long as it isn't invoked or used as a target in a
pointer assignment statement.

If you embellish the example a little bit:

module A

  interface
    subroutine P_A ( X )
      real :: X
    end subroutine P_A
  end interface

end module A

subroutine P_A ( X )
  real :: X
end subroutine P_A

program Test_P_A
  use A, only: P_A
  real :: X
  call P_A (x)
end program TEST_P_A

in separate files, do you get a duplicate-symbol-definition error from
your linker?

Here's what I get

from lf95:                     from NAG f95:

% nm *.o                       % nm *.o

interfaceTest.o:               interfaceTest.o:
00000000 T a_
                               
main.o:                        main.o:
00000000 R .jwe_ssn_MAIN__              U f90_finish
         U jwe_xcop                     U f90_init
         U jwe_xstp            00000000 T main
00000000 T MAIN__                       U p_a_
         U p_a_                
                               
p_a.o:                         p_a.o:
00000000 R .jwe_ssn_p_a_       00000000 T p_a_
00000000 T p_a_

Notice that the symbol p_a is defined in the file p_a.o in both cases,
referenced in the main program in both cases and not even mentioned in
interfaceTest.o (where module A is defined) in both cases.

There are no complaints from the linker, in either case.

What do you get if you put "BIND(C,name="foobar")" after
"subroutine P_A (X)" both in the module and in the external subroutine
definition?