p0943 proposes a C++ header that behaves like the C one, but defines the C
facilities in terms of std::atomic. So we own our stdatomic.h and WG14 owns
theirs.

On Fri, Sep 25, 2020 at 1:09 PM Barry Revzin <barry.revzin_at_[hidden]> wrote:

>
>
> On Fri, Sep 25, 2020 at 2:32 PM Jonathan Wakely via SG10 <
> sg10_at_[hidden]> wrote:
>
>> Resending from the right email account....
>>
>>
>>> We looked at
>>> https://wiki.edg.com/pub/Wg21summer2020/LibraryWorkingGroup/d0943r6.html
>>> in the LWG telecon today. It proposes to add a __cpp_lib_atomic_compat
>>> macro.
>>>
>>> It was pointed out that "compat" is a bit generic and could mean
>>> different things in different contexts. Should we instead name the macro
>>> after the header? I think that's our policy when there's no reason to do
>>> otherwise. It's complicated here by the header being <stdatomic.h> and we
>>> can't have a dot in a macro name.
>>>
>>> Should it be __cpp_lib_stdatomic_h then?
>>>
>>
> Do we own <stdatomic.h> to put a macro in there? __cpp_lib_stdatomic_h
> seems fine. Maybe also __cpp_lib_c_atomic or something to that effect?
>
> Barry
>