Let's add javakotlindoc with example?

A README update with example would be great too

nit: it's a bit inconsistent with "None.component1"

Does it need to be open?

Otherwise it won't compile because Some also declares a component1

What about a test with unwrapping in lambda? Something like this:

listOf(Some(42)).forEach { (value) ->
    assertThat(value).isEqualTo(42)
}

Do we need that though? It's up to compiler right

And Kotlin compiler doesn’t have issues, yep.

It can serve as a checking sample as well, since it is a more desirable usage (at least for me).

Let's also specify return type explicitly here, that makes it easier for users when they look at sources

Let's add few tests to check type system behavior?

I was afraid that this PR would break previous construction:

val (value: String) = Some("str")

Because now Optional declares component1() as T?.

But gladly compiler actually generates proper function override in Some with @NotNull:

class Some extends Optional {
   @NotNull
   public final Object component1() {
      return this.value;
   }

So we get both T? if you destructure Optional and T if you destructure Some<T>, but I'm afraid that this behavior can be easily broken in compiler and that there are not so many similar use cases to report that to them quickly

Something like:

context("destructure Some") {

    val some = Some("string")

    it("destructures to non-null type") {
        val (value: String) = some
    }
}

But for all types: Optional, Some, None. Right now you test behavior but not the type information

Actually, playing more with it I think we should define it as an abstract function:

sealed class Optional {
    abstract operator fun component1(): T?
}

// `Some` overrides it automatically, 
// also looks like we can't define it manually in either case which worries me a bit

object None : Optional<Nothing> {
    override fun component1(): Nothing? = null
}

That will work faster for None because it will just return null without doing instanceof that toNullable() does

And actual behavior for Some will be more clear because right now it's not obvious that toNullable() is not actually executed due to override generated by compiler

nice i like that

This probably should be more generic and showcase destructuring of both types

and moved above Java Interop section, it's a bit confusing atm haha