[Patch, fortran] PR30887 and PR30888 - overdone constraints on %VAL

[Paul Thomas]

Sorry all:

Both the testcases need a

C { dg-options "-O0" }

to cancel out the implicit -pedantic

Paul
> Jerry and Tobias,
>
> The attached are completely self explanatory fixes for the above to
> %VAL problems.
>
> Bootstrapped and regtested on x86_ia64/FC5.
>
> I cannot do anything with them until at least the 6th, so be my guest.
>
> Best regards
>
> Paul
>
> ------------------------------------------------------------------------
>
> Index: gcc/fortran/resolve.c
> ===================================================================
> *** gcc/fortran/resolve.c	(revision 122229)
> --- gcc/fortran/resolve.c	(working copy)
> *************** resolve_actual_arglist (gfc_actual_argli
> *** 1016,1037 ****
>   		 since same file external procedures are not resolvable
>   		 in gfortran, it is a good deal easier to leave them to
>   		 intrinsic.c.  */
> ! 	      if (ptype != PROC_UNKNOWN && ptype != PROC_EXTERNAL)
>   		{
>   		  gfc_error ("By-value argument at %L is not allowed "
>   			     "in this context", &e->where);
>   		  return FAILURE;
>   		}
> - 
> - 	      if (((e->ts.type == BT_REAL || e->ts.type == BT_COMPLEX)
> - 		   && e->ts.kind > gfc_default_real_kind)
> - 		  || (e->ts.kind > gfc_default_integer_kind))
> - 		{
> - 		  gfc_error ("Kind of by-value argument at %L is larger "
> - 			     "than default kind", &e->where);
> - 		  return FAILURE;
> - 		}
> - 
>   	    }
>   
>   	  /* Statement functions have already been excluded above.  */
> --- 1016,1029 ----
>   		 since same file external procedures are not resolvable
>   		 in gfortran, it is a good deal easier to leave them to
>   		 intrinsic.c.  */
> ! 	      if (ptype != PROC_UNKNOWN
> ! 		    && ptype != PROC_DUMMY
> ! 		    && ptype != PROC_EXTERNAL)
>   		{
>   		  gfc_error ("By-value argument at %L is not allowed "
>   			     "in this context", &e->where);
>   		  return FAILURE;
>   		}
>   	    }
>   
>   	  /* Statement functions have already been excluded above.  */
> Index: gcc/testsuite/gfortran.dg/c_by_val_5.f
> ===================================================================
> *** gcc/testsuite/gfortran.dg/c_by_val_5.f	(revision 0)
> --- gcc/testsuite/gfortran.dg/c_by_val_5.f	(revision 0)
> ***************
> *** 0 ****
> --- 1,17 ----
> + C { dg-do compile }
> + C Tests the fix for PR30888, in which the dummy procedure would
> + C generate an error with the %VAL argument, even though it is
> + C declared EXTERNAL.
> + C
> + C Contributed by Peter W. Draper <p.w.draper@durham.ac.uk>
> + C
> +       SUBROUTINE VALTEST( DOIT )
> +       EXTERNAL DOIT
> +       INTEGER P
> +       INTEGER I
> +       I = 0
> +       P = 0
> +       CALL DOIT( %VAL( P ) )
> +       CALL DOIT( I )
> +       CALL DOIT( %VAL( P ) )
> +       END
> Index: gcc/testsuite/gfortran.dg/c_by_val_4.f
> ===================================================================
> *** gcc/testsuite/gfortran.dg/c_by_val_4.f	(revision 0)
> --- gcc/testsuite/gfortran.dg/c_by_val_4.f	(revision 0)
> ***************
> *** 0 ****
> --- 1,11 ----
> + C { dg-do compile }
> + C Tests the fix for PR30887, in which P would be rejected because it
> + C is not of default kind.
> + C
> + C Contributed by Peter W. Draper <p.w.draper@durham.ac.uk>
> + C
> +       SUBROUTINE VALTEST2()
> +       INTEGER*8 P
> +       P = 0
> +       CALL DOIT( %VAL( P ) )
> +       END
>