[libcxx-commits] [libcxx] [libcxx] Fix `long double` literals in <numbers> (PR #222830)

[A. Jiang via libcxx-commits]

frederick-vs-ja wrote:

Mostly LGTM. But I wonder why we haven't brought affected platforms to CI.

> Future work:
> Add support for `__float128` and other types in `<numbers>`.

This seems to be a part of [P1467R9](https://wg21.link/p1467r9) (tracked in #105196).

https://github.com/llvm/llvm-project/pull/222830