C++2a synchronisation inefficient in GCC 11

[Ville Voutilainen]

On Sat, 27 Feb 2021 at 02:22, Marc Glisse <marc.glisse@inria.fr> wrote:
>
> On Fri, 26 Feb 2021, Thiago Macieira via Libstdc++ wrote:
>
> > How about the other way around: only enable the new, experimental code if the
> > user defines _GLIBCXX_I_WANT_EXPERIMENTAL before the #includes?
>
> That's exactly what the user did, they passed -std=c++20...
>
> Ok, it might help if the option was called -std=c++2a until it is stable,
> or if we defined a macro that advertises if we are in an experimental mode
> or not.
>
> But ultimately, it is the user's responsibility to use working (not
> experimental) tools. And you can document that your library is only
> compatible with gcc-12+.

To be fair, I partially understand where Thiago may be coming from.
Doing that is *horrible* for things like.. say.. Qt.
While it's not nice to opt in to new features for seventeen different
implementations, it's not much fun to disable
features due to bugs in seventeen different implementations either.
(Yeah, seventeen is a bit of an exaggeration.)