[Bug fortran/33997] New: Generic interface: Rejects non-ambiguous interface as ambiguous

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

Based on the example/report at
http://groups.google.com/group/comp.lang.fortran/browse_thread/thread/18873113b18cd5e9/
by Rich Townsend.

I think the following subroutines are not ambiguous. One is

  (COMPLEX, COMPLEX, procedure)
and the other is
  (COMPLEX, procedure [, COMPLEX])

Thus without foo= they are non-ambiguous and as the first "one" has a
non-optional z_2= (and "two" has no "z_2"), changing the order still should not
make it ambiguous.

For the original example, Michael Metcalf believes that it is valid (and he
stated that ifort compiles it).

My example below compiles with openf95, g95 and sunf95 and is rejected by ifort
(!), NAG f95 and gfortran.

If you believe that my example is invalid - or when you have a fix, please
check also the original example (see link above).


module test
 implicit none
 interface foo
   module procedure one, two
 end interface foo
contains
    subroutine one (z_1, z_2, f_func)
     complex, intent(in)           :: z_1
     complex, intent(in)           :: z_2
     interface
        function f_func () result (f)
          complex :: f
        end function f_func
     end interface
   end subroutine one

! works:
!  subroutine two (z_1, f_func)
! fails:
   subroutine two (z_1, f_func, f_1)
     complex, intent(in), optional :: f_1

     complex, intent(in)           :: z_1
     interface
        function f_func () result (f)
          complex :: f
        end function f_func
     end interface
   end subroutine two
end module
end


-- 
           Summary: Generic interface: Rejects non-ambiguous interface as
                    ambiguous
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Keywords: rejects-valid
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: burnus at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33997