Am Mittwoch, den 10.02.2021, 10:25 -0800 schrieb Hans Boehm via Liaison:
> My suggestion on the WG14 reflector was to make _Atomic T ill-formed if
> _Atomic(T) does not have the same representation as T, and equivalent to
> _Atomic(T) otherwise. I think that would make it usable as a qualifier and
> atomic_ref<> replacement. That didn't seem to go over well there, but
> the reflector seemed to be a poor way to understand whether there was a
> WG14 consensus for any of the options.

The concern I had is that you can not use the _Atomic qualifier for
the atomic_ref<> use case in a portable way if is is then ill-formed
for some types T on some platforms. On the other hand, this would
be a major improvement because today you can not use atomics at all
if you have this requirement.

Best,
Martin

> I think it provides enough compatibility with the status quo that it might
> be an acceptable change by WG21 standards. That may not make it acceptable
> by WG14 standards. It breaks only code that uses the type qualifier
> notation for something that behaves very differently from a type qualifier,
> and it breaks it loudly.
>
> Making/keeping _Atomic T, _Atomic(T), and std::atomic(T) all synonymous
> also works for P0943. But it would eventually require an alternate
> atomic_ref<> solution for C, I think. And having _Atomic as an alternate
> spelling for _Atomic() (that buys nothing except breaking C++
> compatibility?) seems suboptimal to me.
>
> On Wed, Feb 10, 2021 at 9:43 AM Joseph Myers <joseph_at_codesourcery.com>
> wrote:
>
> > On Tue, 9 Feb 2021, Hans Boehm via Liaison wrote:
> >
> > > We've had similar discussions before on the WG14 reflector. The concern
> > > here is basically that if _Atomic were to be usable as a real type
> > > qualifier that can be added to parameters, then _Atomic T and T should
> >
> > have
> > > the same representation.
> >
> > Whereas in fact it's not a qualifier at all, other than purely
> > syntactically (_Atomic T is an alternative syntactic notation for _Atomic
> > (T)), and it's explicit in the C standard that references to qualified and
> > unqualified types do not count _Atomic as a qualifier by default.
> >
> > > I think C really wants the type qualifier to work, since
> > > it doesn't have atomic_ref<>. So something has to give, and we've had
> > > different proposals about what that should be.
> >
> > Whatever system you might add for atomic operations on non-atomic data, I
> > think making the syntactic-qualifier _Atomic into something semantically a
> > qualifier would be an unreasonably incompatible way of adding such a
> > feature.
> >
> > --
> > Joseph S. Myers
> > joseph_at_codesourcery.com
> >
>
> _______________________________________________
> Liaison mailing list
> Liaison_at_[hidden]socpp.org
> Subscription: https://lists.isocpp.org/mailman/listinfo.cgi/liaison
> Link to this post: http://lists.isocpp.org/liaison/2021/02/0317.php