Feature request: PartialEq<[T]> for set::Slice, PartialEq<[(K, V)]> for map::Slice

[clarfonthey]

This would be particularly nice for tests, so that I don't need to construct a new set/map just to assert that the values are in a particular order.

[cuviper]

You could assert!(slice.iter().eq([...])), or use itertools::assert_equal. For tests, the latter is nice because you'll get a useful debug message, including the position.

[clarfonthey]

I mean, both options are fair; the latter is best for understanding the failures, but I still was thinking this would be a relatively easy solution that doesn't require any extra dependencies.

It's just hard to beat assert_eq!(set.as_slice(), [1, 2, 3]) and this notation makes extra sense considering how the set is explicitly ordered, unlike regular HashSets.

[cuviper]

I just realized the iterator ways are more annoying for maps, since the items are different: &(K, V) vs. (&K, &V).

[cuviper]

It's just hard to beat assert_eq!(set.as_slice(), [1, 2, 3])

Playing with #376, it turns out that's not enough because the latter is an array, and you don't get automatic slice coercion here. I guess we could add impls for arrays too, but the scope is creeping...

I also wonder if we should make this more flexible, e.g. PartialEq<[U]> for Slice<T> like Vec does. That could be added later, but I think it's a minor breaking change because it affects type inference, especially for integer cases like your example.