On Mon, 25 Sept 2023 at 01:03, Thiago Macieira via Std-Proposals
<std-proposals_at_[hidden]> wrote:
> I'm missing something.
>
> How is this any different to std::strong_ordering?

It isn't. It implements the same thing, without using std::strong_ordering.

> The implementations do use values -1, 0, 1 and something else (2, -2, -127,
> etc.). They just happen to wrap them in a class that can relate to
> partial_ordering, weak_ordering.
>
> So what's the problem with the current implementation?

What Chris seems to be suggesting is that what's in the standard
should've been less coupled and lower-level,
so that you can polyfill your own stuff in the case of a buggy stdlib.
That has its own downsides, so this
is really a "water under the bridge, fix those libraries, or have them fixed".