Force exhaustion in iter::ArrayChunks::into_remainder

[krtab]

Closes: #123333

[rustbot]

r? @scottmcm

rustbot has assigned @scottmcm.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[zachs18]

Suggested change

	while let Some(_) = self.next() {}
	if self.remainder.is_none() {
	while let Some(_) = self.next() {}
	}

If I: DoubleEndedIterator + ExactSizeIterator (thus Self: DoubleEndedIterator), and self.next_back() has been called, and there is a remainder, then self.remainder is already set to Some, so we can skip exhausting the iterator. (Optimization , not required for correctness).

Hypothetically, specialization could be used to do the "fast" thing if I: DEI + ESI regardless of if the remainder has already been found, but in general that would be unsound (I think?), since user implementations of DEI/ESI can have lifetime requirements.

[krtab]

Nice catch regarding next_back. I'll implement it.
I didn't get your point about specialization.

[scottmcm]

Hmm, normally we say to use for_each(drop) for this

rust/library/core/src/iter/traits/iterator.rs

Line 1996 in 0e15f5e

#[must_use = "if you really need to exhaust the iterator, consider `.for_each(drop)` instead"]

but I guess this can't use that because it consumes it :/

Maybe it should be

Suggested change

	while let Some(_) = self.next() {}
	self.try_for_each(NeverShortCircuit::wrap_mut_1(drop));

?

Meh, we can always change it later if it ends up mattering. This is probably fine.

[scottmcm]

Thanks for the fix! I wondered for a bit about what would happen with this after exhaustion or if it's called twice, but then I remembered it's self so can't be called twice and ArrayChunks is always fused anyway.

@bors r+

[bors]

📌 Commit d9a8903 has been approved by scottmcm

It is now in the queue for this repository.