Re: [PATCH 2/2] libstdc++: use new built-in trait __add_const

[Jonathan Wakely <jwakely@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1329 bytes --]

On Tue, 21 Mar 2023 at 11:21, Marc Glisse via Libstdc++ <
libstdc++@gcc.gnu.org> wrote:

> On Tue, 21 Mar 2023, Ken Matsui via Libstdc++ wrote:
>
> >   /// add_const
> > +#if __has_builtin(__add_const)
> > +  template<typename _Tp>
> > +    struct add_const
> > +    { using type = __add_const(_Tp); };
> > +#else
> >   template<typename _Tp>
> >     struct add_const
> >     { using type = _Tp const; };
> > +#endif
>
> Is that really better? You asked elsewhere if you should measure for each
> patch, and I think that at least for such a trivial case, you need to
> demonstrate that there is a point. The drawbacks are obvious: more code in
> libstdc++, non-standard, and more builtins in the compiler.
>

Right, this one isn't even getting rid of any partial specializations, but
it is giving the preprocessor more work to do.

Adding the extra built-ins to the compiler makes the compiler (very
slightly) bigger and slower, so a real benchmark would require comparing an
unpatched gcc (without the new built-in) to a patched gcc and patched
libstdc++ sources.



>
> Using builtins makes more sense for complicated traits where you can save
> several instantiations. Now that you have done a couple simple cases to
> see how it works, I think you should concentrate on the more complicated
> cases.
>
> --
> Marc Glisse
>
>