REPL: Weak pointers don't point to None when pointee hits end of scope

[ADKaster]

Running samples/weak/basic.jakt produces different results for comptime vs build time execution, suggesting that either foo is not destroyed at the end of the scope, or the weak pointer is not updated:

class Foo {
    public function hello(this) => "friends"
}

function main() {
    mut weak_foo: weak Foo? = None

    println("weak_foo has_value? {}", weak_foo.has_value())

    {
        let foo = Foo()
        weak_foo = foo

        println("weak_foo has_value? {}", weak_foo.has_value())
        println("weak_foo hello: {}", weak_foo!.hello())
    }

    println("weak_foo has_value? {}", weak_foo.has_value())
}

~/jakt$ ./build/bin/jakt -r samples/weak/basic.jakt
weak_foo has_value? false
weak_foo has_value? true
weak_foo hello: friends
weak_foo has_value? true <------ !!!!
~/jakt$ ./build/bin/jakt -cr samples/weak/basic.jakt
weak_foo has_value? false
weak_foo has_value? true
weak_foo hello: friends
weak_foo has_value? false

[ADKaster]

This is expected because WeakPtr::has_value() is not implemented for comptime.

At the same time, this is unexpected because WeakPtr::has_value() is not implemented for comptime. It should fail to parse/compile the code with -r

[alimpfard]

I think this is just because the interpreter doesn't actually perform the implicit casts required for our weakptr mess to work, so it's seen as an Optional instead.