[Bug fortran/46478] New: Missing diagnosis for combining SUBROUTINE and FUNCTION in a GENERIC interface

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46478

           Summary: Missing diagnosis for combining SUBROUTINE and
                    FUNCTION in a GENERIC interface
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Keywords: accepts-invalid
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: burnus@gcc.gnu.org


Found at:
http://groups.google.com/group/comp.lang.fortran/browse_thread/thread/2084596e400fdf2d

gfortran allows generic interfaces which combine FUNCTIONs with SUBROUTINEs.

That's invalid Fortran, but allowed as vendor extension as it is not a
constraint.

gfortran diagnoses the problem if one does not use "(MODULE) PROCEDURE" but
includes the declaration directly in the interface body - albeit the error
message is not very clear:
  Error: SUBROUTINE at (1) does not belong in a generic function interface


Other compiles:
- NAG rejects it with
  Error: SUBR_NAME and FUNC_NAME must both be SUBROUTINEs or both be FUNCTIONs
for generic NEW_NAME

- g95 with
  Error: Interfaces of generic 'new_name' at (1) must all be FUNCTIONs
  or SUBROUTINEs

- ifort has a warning -- and only with -stand f03:
  warning #7953: Standard Fortran requires that within a scoping unit, two
  procedures that have the same generic name shall both be subroutines or both
  be functions.   [SUBR_NAME]


Even with -std=f95, gfortran does not even issue a warning.

Expected:
- Reject it with -std=f95/f2008
- Consider print a warning by default for -std=gnu; printing an error
  would also be possible


MODULE something
  INTERFACE new_name
    PROCEDURE func_name, subr_name
  END INTERFACE
CONTAINS
   LOGICAL FUNCTION func_name(arg1,arg2)
!   :
   END FUNCTION

   SUBROUTINE subr_name(arg1,arg2,arg3)
!   :
   END SUBROUTINE
END MODULE