Hi Jens,

The extra definition of integer below is a mistake. I must have copied
it and not marked it <ins> and then just not looked at it again. Weird
stuff.

> It would probably also be nice to clean up the core wording:
> "integral" is everything, "integer" just the signed / unsigned
> integer types. How much would change to make that happen?

I REALLY like this idea. We're already fairly close to that anyway.
- std::signed_integral is broader than signed integer
- std::unsigned_integral is broader than unsigned integer

Making "integer" a collective term for "signed and unsigned integers"
instead of a synonym for "integral type" would be a way more
consistent. It would mean that in every case, "integer" is the narrow
term that you likely want, and "integral" is the broader term.

I haven't yet looked at all the wording, but I don't think it would
require large changes based on what I remember from writing my
int128_t paper. The wording uses "signed integer" or "integral type"
etc. anyway, so the term "integer" can be more or less defined as we
please.

I'll get back to you later once I've gathered up all the affected wording.