Specialize Zip::nth for TrustedRandomAccess

[scottmcm]

Fixes the performance difference observed in https://users.rust-lang.org/t/performance-difference-between-iterator-zip-and-skip-order/15743

Before:

test iter::bench_skip_then_zip ... bench: 154 ns/iter (+/- 4)
test iter::bench_zip_then_skip ... bench: 4,725 ns/iter (+/- 225)

After:

test iter::bench_skip_then_zip ... bench: 154 ns/iter (+/- 2)
test iter::bench_zip_then_skip ... bench: 81 ns/iter (+/- 19)

[rust-highfive]

r? @kennytm

(rust_highfive has picked a reviewer for you, use r? to override)

[kennytm]

Could you add a test to ensure side effects are executed?

fnmain(){letmut a = Vec::new();letmut b = Vec::new();let value = (1..7).map(|n| {
a.push(n);
n *10}).zip((2..9).map(|n| {
b.push(n *100);
n *1000})).skip(1).nth(3);assert_eq!(value,Some((50,6000)));assert_eq!(a, vec![1,2,3,4,5]);assert_eq!(b, vec![200,300,400,500,600]);}

[scottmcm]

@kennytm Good call.

It looks like Range: !TrustedRandomAccess, though, so I swapped it out for cloning slice iterators.

[kennytm]

Minor nit: why this { is on a new line 😄

[scottmcm]

Now that is an excellent question 😅

[kennytm]

(This also)

[kennytm]

Travis passed, so

@bors r+

[bors]

📌 Commit 5105fc1 has been approved by kennytm