On Friday, 9 January 2026 14:14:29 Brasilia Standard Time Frederick Virchanza
Gotham via Std-Proposals wrote:
> Wait.... what? You mean some stuff _can_ be optional depending on the
> target? Really?

std::uintptr_t and the bit-exact integer types are indeed optional, in the C
standard. Though as David said, the machines where they aren't supported are
rare and esoteric. In fact, I don't think any of those currently target C++17
or later.

That's not what I understood from you. I didn't get the impression you wanted
an optional one. You wanted one to always be there, which would either mean
you didn't know there were modern machines with proper C++ support where it
can't be supported, or you wanted them to be left in the dustbin of history.

I skipped the discussion about faking behaviour. That's either disingenuous,
for debugging only, or matches the behaviour we already have in the
std::atomic types. Why should we add another type that is indistinguishable
from the one we have?

-- 
Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
  Principal Engineer - Intel Data Center - Platform & Sys. Eng.