Added CanEqual support

[felix-hedenstrom]

At $work we are trying out multiversal equality. One of its bigger drawbacks is the need to specify a lot of CanEqual givens. Right now we are defining our own CanEqual for neotypes but it requires ugly given imports wherever we need to compare neotypes.

This PR adds a handy CanEqual given that makes it so that Neotypes will have an available CanEqual if the wrapped type has one.

[felix-hedenstrom]

Link to relevant scala3 documentation https://docs.scala-lang.org/scala3/reference/contextual/multiversal-equality.html

[felix-hedenstrom]

Looks like this revealed some test issues but also made me realize that this isn't a backwards compatible change as of now. Let me see if I can make it backwardscompatible.

[felix-hedenstrom]

I tried adding a regression test to make sure that old equality behavior remains where the flag is not enabled, but the chimney test is still failing.

I don't have any ideas what might be causing this. Not sure how to proceed...

[felix-hedenstrom]

I think all the errors that this PR detected are actual code errors that we should fix but that went undetected since there was no CanEqual defined.

I realized that the documentation mentions why this happens

Even though canEqualAny is not declared as given, the compiler will still construct a canEqualAny instance as answer to an implicit search for the type CanEqual[L, R], unless L or R have CanEqual instances defined on them, or the language feature strictEquality is enabled.

So me defining a CanEqual removes the use of the default defined one. The PR in this form would be a breaking change (even though I think it is a very useful breaking change, since it detected multiple issues within this repo alone). I haven't worked with hybrid codebases were CanEqual is defined for some cases but the flag is not enabled so I was not aware of this outcome.

@kitlangton should I move this into a separate module, something like neotype-canequal?

[felix-hedenstrom]

Hi @kitlangton. Have you had time to think about this? I would be fine making this into neotype-canequal if you think that it is appropriate.