Re: [PATCH V4 2/6] Fortran: Resolve dynamic properties of pointer types.

[Bernhard Heckel <bernhard dot heckel at intel dot com>]

On 05/10/2016 11:35, Yao Qi wrote:
> On Wed, Oct 5, 2016 at 9:44 AM, Yao Qi <qiyaoltc@gmail.com> wrote:
> > Bernhard Heckel <bernhard.heckel@intel.com> writes:
> >
> > > 2016-07-08  Bernhard Heckel  <bernhard.heckel@intel.com>
> > >
> > > gdb/Changelog:
> > >        * gdbtypes.c (resolve_dynamic_types_internal): Resolve pointer types.
> > This line doesn't match the change, which is in is_dynamic_type_internal.
> >
> > >          (resolve_dynamic_pointer): New.
> > Otherwise, patch is good to me.
> Do we already have a test case which has a Fortran pointer which has
> a dynamic associated property?  If no, we need to add one.
Yes, we have a testcase already.
vla-ptype.exp + vla.f90
This testcase was already passing as gfortran describes a pointer to an 
array as an array +  associated property.
For Ifort, a pointer in the program is still a pointer type in the debug 
info.

> Patch 3/6 resolves dynamic target type of a pointer, a pointer points
> to a dynamic type, but what does this patch do?

The patch only resolves the "associated property" of a pointer.
According to Joel review (some time ago), I only resolve target types of 
pointer when we actually have to (see following patches in this series).
I wanted to keep this patch separate from resolving dynamic target 
types, as this patch is independent from those but not vice verse.

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928