It's faster as it checks for ICollection, IList, also checks if it can use spans in case of primitive types etc.

Also improved CollectionComparer.GetHashCode.

The old implementation used Distinct + OrderBy + a combiner to normalize hash codes:

• Sorting was used to make the combination order-independent.
• This worked, but sorting is unnecessary, allocates memory, and costs O(n log n) time.
• Hash codes are called frequently by dictionaries and hash sets, so they should be cheap.

The new implementation avoids sorting:

• Uses a HashSet to process distinct elements only.
• Combines hash codes with XOR in a single pass.
• No extra allocations for sorting, much faster, simpler.

It directly matches the documented semantics:

Order of the sequence is not relevant, neither are multiple entries of the same value.

So the new GetHashCode is simpler, faster, and fully correct for the intended behavior.

Checklist:

• I've read the contribution guidelines
• My code follows the style of this project
• I've added or updated relevant unit tests