Today, I saw something that is new for me in kotlin's syntax. In <https://github.com/zach-klippenste...

# getting-started

[MR3Y]

Today, I saw something that is new for me in kotlin's syntax. In https://github.com/zach-klippenstein/compose-fractal-nav, there is an interface FractalParent that is declared as follows:

internal interface FractalParent : FractalNavScope {
    // ...
    val zoomAnimationSpecFactory: () -> AnimationSpec<Float>
  // ...
}

and the class that implements the interface is:

internal class FractalNavStateImpl : FractalParent {
    // ...
    override lateinit var zoomAnimationSpecFactory: () -> AnimationSpec<Float>

my question is how he was able to override zoomAnimationSpecFactory and change its modifier from val to lateinit var without getting a compilation error?

[Casey Brooks]

Think of it in terms of what a property in an interface actually means. Interfaces cannot actually have backing fields which hold data, so all the compiler knows is that val zoomAnimationSpecFactory is effectively no different from fun getZoomAnimationSpecFactory().

So when FractalNavStateImpl implements the interface, override lateinit var does still include that “get function” from the property, which means it is implementing everything required from the interface. It just also adds fun setZoomAnimationSpecFactory(value), effectively

[ephemient]

this is explicitly mentioned in the docs, https://kotlinlang.org/docs/inheritance.html#overriding-properties

[ephemient]

also

interface Base {
    val foo
}
class Impl {
    override var foo
        private set
}

is a common pattern; this allows Impl and nothing else to update the value

[MR3Y]

thanks for the heads-up @ephemient 👏