Brainstorm for what goes in std inclusion

[bluss]

What parts of Itertools are solid enough for std?

.join(sep: &str) -> String

where Self::Item: Display

• Pro Very popular, easy to use
• Con Using Display is very "do what I mean" (which is mostly good) but not a zero cost abstraction; formatting uses dynamic dispatch

Plan: ?

.collect_vec() -> Vec<_>

• Pro Popular, often what you want
• Con Std prefers generic over specific. Name is bad

Plan: Propose under name .into_vec() for iterators.

.format(sep: &str)

• Pro Unique, does something smart neatly that otherwise requires some boring trait impl code
• Con Bends how format strings are used (once per element instead of just once)
• Con Is only once formattable (due to being once-iterable).

Plan: Propose as is?

.flatten()

• Pro simple, no closure in the type
• Pro skip double-endedness on introduction, simplifying all kinds of special implementations for the iterator

No-gos

[pickfire]

I think join is not a good idea, join should work with PathBuf as well and use Separator like the one in slice::join. https://rust-lang.zulipchat.com/#narrow/stream/219381-t-libs/topic/join.20on.20Iterator

flatten seemed to be there in std already.