Re: PSA: Default C++ dialect is now C++17

[Marek Polacek <polacek@redhat.com>]

On Mon, Sep 14, 2020 at 11:13:18AM -0400, Jason Merrill via Gcc-patches wrote:
> On Mon, Jun 29, 2020 at 1:25 PM Martin Liška <mliska@suse.cz> wrote:
> >
> > On 6/29/20 4:57 PM, Marek Polacek wrote:
> > > On Mon, Jun 29, 2020 at 09:51:57AM +0200, Martin Liška wrote:
> > >> On 6/26/20 9:34 PM, Marek Polacek via Gcc-patches wrote:
> > >>> As discussed last month:
> > >>> <https://gcc.gnu.org/pipermail/gcc-patches/2020-May/545879.html>
> > >>> it's time to change the C++ default to gnu++17.  I've committed the patch after
> > >>> testing x86_64-pc-linux-gnu and powerpc64le-unknown-linux-gnu.  Brace yourselves!
> > >>>
> > >>> Marek
> > >>>
> > >>
> > >> Just a small note that 510.parest_r SPEC 2017 benchmark can't be built now
> > >> with default changed to -std=c++17. The spec config needs to be adjusted.
> > >
> > > Interesting, do you know why?  Does it use the register keyword?
> >
> > Apparently it needs -fno-new-ttp-matching for successful compilation.
> > There's a reduced test-case I made:
> >
> > cat fe.ii
> > template <int, int> class FiniteElement;
> > template <int, int> class DoFHandler;
> > class FETools {
> >    template <int dim, class InVector, class OutVector, int spacedim>
> >    void back_interpolate(const DoFHandler<dim, spacedim> &, const InVector &,
> >                          const FiniteElement<dim, spacedim> &, OutVector &);
> >    template <int dim, template <int> class DH, class InVector, class OutVector,
> >              int spacedim>
> >    void back_interpolate(const DH<dim> &, InVector,
> >                          const FiniteElement<dim, spacedim> &, OutVector);
> > };
> > template <int dim, int = dim> class DoFHandler;
> > template <int dim, int = dim> class FiniteElement;
> > template <int dim, class InVector, class OutVector, int spacedim>
> > void FETools::back_interpolate(const DoFHandler<dim, spacedim> &,
> >                                 const InVector &,
> >                                 const FiniteElement<dim, spacedim> &,
> >                                 OutVector &) {}
> > template void FETools::back_interpolate(const DoFHandler<3> &, const float &,
> >                                          const FiniteElement<3> &, float &);
> 
> Hmm, looks like I never sent this.
> 
> Further reduced:
> 
> template <int N, int M = N> class A;
> template <int N, int M> void fn(A<N, M> &) {}
> template <int N, template <int> class TT>  void fn(TT<N> &);
> template void fn(A<3> &);
> 
> This breaks due to the C++17 changes to template template parameters
> causing A to now be considered a valid argument for TT; with that
> change both function templates are valid candidates, and neither is
> more specialized than the other, so it's ambiguous.
> 
> There are still some open core issues around these changes.

Thanks.  I just pushed a patch to introduce GCC 11 porting_to:
<https://gcc.gnu.org/gcc-11/porting_to.html> and documented this change.

Let me know if you have any comments.

Marek