This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug fortran/40949] FAIL: gfortran.dg/proc_ptr_7.f90
- From: "rguenther at suse dot de" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 4 Aug 2009 08:14:04 -0000
- Subject: [Bug fortran/40949] FAIL: gfortran.dg/proc_ptr_7.f90
- References: <bug-40949-10053@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #2 from rguenther at suse dot de 2009-08-04 08:14 -------
Subject: Re: FAIL: gfortran.dg/proc_ptr_7.f90
On Mon, 3 Aug 2009, burnus at gcc dot gnu dot org wrote:
> ------- Comment #1 from burnus at gcc dot gnu dot org 2009-08-03 23:17 -------
> > I see the Fortran FE accepts excess parameters to functions and only warns
> > for that with -fwhole-file ...
>
> Well, that's not surprising - if the function is not in CONTAINS
> (internal/module procedure) or in a module, the prototype is not available in
> Fortran thus it is just an external procedure without any known interface (even
> if it is just one line higher up).
Ah, ok.
> One purpose of -fwhole-file is to allow to diagnose such things as error.
>
> * * *
>
> I think this is fixed by the following patch. If I understand the code
> correctly, we currently only append the void_type_node if there is at least one
> argument, which allows for "(...)" for procedures without any argument.
Well, I think you are building a new fndecl at callsites with a
prototype that matches the actual arguments passed.
> Index: gcc/fortran/trans-types.c
> ===================================================================
> --- gcc/fortran/trans-types.c (Revision 150376)
> +++ gcc/fortran/trans-types.c (Arbeitskopie)
> @@ -2324,8 +2324,7 @@ gfc_get_function_type (gfc_symbol * sym)
> while (nstr--)
> typelist = gfc_chainon_list (typelist, gfc_charlen_type_node);
>
> - if (typelist)
> - typelist = gfc_chainon_list (typelist, void_type_node);
> + typelist = gfc_chainon_list (typelist, void_type_node);
>
> if (alternate_return)
> type = integer_type_node;
That looks correct to me.
Richard.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40949