Karoy Lorentey

This is exciting work. I think these would be a great fit for this package, and I'm personally very much interested in getting these landed (& eventually added to the...

I'm landing this on a feature branch to start integration & performance work. 💯

Can you provide your reasoning behind these suggestions? > 1. Could the [collection APIs][] be moved to a separate type? > > ```swift > extension BitSet { > public func...

>> We can certainly precompute startIndex -- the big question there is if that will actually be helpful overall. > I don't understand the [benchmark](https://forums.swift.org/t/bit-array-and-bit-set-api-review-the-end-of-a-gsoc-project/51396#computed-startindex-property-26) in the original API review....

What do you think about `BitSet.insert` and similar operations being generic over `FixedWidthInteger`, rather than simply taking `Int` values? (Consider that `BitSet` is a collection of `Int`s -- it does...

What's your opinion on `BitArray`'s `bitPattern` initializer? Should `BitArray` provide a way to go the other way, i.e., initialize a binary integer based on the contents of a bitarray?

We currently have a `BitArray` initializer that takes a `BitSet`, but we don't have an initializer to go the other way. We probably should! These two types provide two different...

What about naming? I'm pretty sure bitset & bitarray work best as single words (e.g., C++ has `std::bitset`, not `std::bit_set`), but we're currently camelcasing them in their type names.

(On providing a labeled subscript in `SetAlgebra`) > Unfortunately, it's not obvious whether to use a conditional `insert(_:)` or an unconditional `update(with:)` operation. The `someSet[member: foo] = true` syntax reads...

>> What about naming? I'm pretty sure bitset & bitarray work best as single words (e.g., C++ has `std::bitset`, not `std::bit_set`), but we're currently camelcasing them in their type names....