Ruby

• Interesting. Can you share a bit more detail about your "if two arrays intersect but not actually caring about the intersection" use case? For instance if you have open sourced such code, a URL for it helps us a lot.

• If you have zero interest to the intersection itself I think you can avoid creating the temporary array.

The most recent example I encountered was authorizing a user in a Rails endpoint. Each user has a list of abilities, and each endpoint has a list of abilities that is authorized to perform the action. We just need to check if there is an intersection between the two lists.

I updated the PR to avoid creating the intermediate array.

Demonstrated use cases helps the core team & matz assess on the usefulness of a proposed
change, which can help in accepting issue requests ultimately in the long run (if the
use case is considered sufficiently useful). :)

I threw together an example in a gist which is very similar to the use case I described. I have also run into numerous other situations where this functionality would have been useful.

I added some comments in the gist but I'll cross-post them here for reference:

The current behavior creates an intermediate array for both intersection tests when the resulting array is clearly not needed. With the desired behavior, we could avoid creating the intermediate array and produce a faster best case runtime.

It also seems like it comes up fairly commonly on stackoverflow/blogs:
https://stackoverflow.com/questions/2603895/how-can-i-check-if-a-ruby-array-includes-one-of-several-values
https://www.ruby-forum.com/t/find-if-an-array-has-any-element-present-in-another-array/166877
https://irb.rocks/comparison-arrays-ruby/
https://tosbourn.com/set-intersection-in-ruby/

It might make sense to use ary1.to_set.intersect?(ary2). That way it makes explicit the fact that ary1 must be converted to a set. But Set#intersect? would have to support any Enumerable.

I tried setting that up here: https://github.com/ruby/ruby/pull/2598

I ignored infinite ranges for now

Matz/Marc-Andre: anything I can do to help here as far as documentation or providing more use cases? I/my colleagues would love to see this make it into Ruby.

Accepted.

Matz.

We discussed this in today's developer meeting. Here are some remarks:

• It takes one argument for now.
• It means !(a & b).empty? instead of (a & b).any? so [nil].intersect?([nil]) evaluates to true.
• The implementation will internally use a hash to match the behavior of Array#&.