On Tue, 10 Feb 2026 at 13:10, Marcin Jaczewski
<marcinjaczewski86_at_[hidden]> wrote:
>
> wt., 10 lut 2026 o 11:57 Ville Voutilainen
> <ville.voutilainen_at_[hidden]> napisał(a):
> >
> > On Tue, 10 Feb 2026 at 12:40, Marcin Jaczewski
> > <marcinjaczewski86_at_[hidden]> wrote:
> > > > Or perhaps use the prefix '='. Like we.. ..already do.
> > >
> > > I recall that something like this was mentioned, but is it already
> > > available, what paper add this?
> > > And if yes then this whole discussion (about attributes) is a bit moot
> > > as it solves all problems.
> >
> > That would be https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2025/p3394r4.html
> >
> > We don't seem to have (p)reserved any annotation names for use by
> > standard facilities.
>
> Aaa right, this is for reflection, this is why I do not connect dots,
> but this could conflict with user defined values. And best would be
> that `[[foo]]` and `[[=foo]]` means the same but one is not ignorable.
> But technically we could define constexpr symbols that could be
> recognized by a compiler like `[[foo]]` equal to `[[=std::foo]]`. This
> could be used too to query normal std attributes if the compiler
> recognized them.

Precisely.