scottmcm

(I'm personally happy with the usual "if it could be written on stable, then it's a libs-api question, not lang" here.)

`Ipv6Addr` is `Copy` -- how do you propose to store this additional string? I didn't see any obvious length restriction in the RFC, and for a type in `core` it...

> I feel it more surprising that valid IPv6 addresses are not properly parsed I think there's now multiple definitions of "address" and having those in different types would make...

I'm glad to see this leading away from indexing by mixed ranges. The thing I'm saddest about there are situations like "is `v[2..-2]` empty?" or "wait, what does `v[-2..2]` return?"...

I made a crate to try out the "slice adapter" idea above (https://github.com/rust-lang/rfcs/issues/2249#issuecomment-352287807) ```rust let r = [1, 2, 4, 9, 16, 25].rev(); assert_eq!(r[0], 25); assert_eq!(r[1..3].rev(), &[9, 16]); assert_eq!(r.split_first().unwrap().0, &25);...

In general, I'd expect custom map iterators to actually optimize worse, because they can't take advantage of all the specializations in the standard library. For example, `std::iter::Map` is if the...

RPITIT is a great call-out here. If these could be written as just `-> std::iter::Map` or whatever, things could well be different. (Though even then I'm not convinced that `.map_try_into()`...

> We'd like to get feedback and merge this now before we go too far in a direction that others have concerns with. I continue to be skeptical of the...

This reminds me of the old complaint about TAIT that it means that `type G = impl Iterator;` doesn't work in a copy-paste way for APIT. Should there be a...

@rfcbot concern require-repr-transparent The reference section spells out a bunch of requirements that I think should be deduped to referencing an existing feature. Rather than > 3. The struct must...