How to perform implementation of shared viewModel using precompose-koin?

[farhazulmullick-pw]

Is there any way so that we can use same viewModel reference in all the screen of a group using precompose-koin instead of implementeing like.

val vm by inject<MYViewModel>() in every screen

[Tlaster]

Sorry but I don't get it, if you want to get something that is injected from koin, at least you need to call something likeval sth = koinInject<Something>(), so in your case, you still need to call val vm = koinViewModel<ViewModel>() in every screen even you do can get the same instance of view model by passing the root LocalStateHolder.current to every screen like this val vm = koinViewModel<ViewModel>(stateHolder = stateHolder)

[farhazulMullick]

@Tlaster
Is there any way to get the lifecycle of a graph/nested-graphs in a NavHost. Like when we are releasing the resources?
Is it kept in memory till NavHost composable's disposableEffect is called?

[Tlaster]

There's a much more simple way to do it, just use DisposableEffect at the same level of your NavHost like this:

DisposableEffect(Unit) {
    //...
    onDispose {
        // release your resource
    }
}
NavHost(
    //...
)

If you want a per back stack's lifecycle, there's also a LocalLifecycleOwner.current that you can use.

[farhazulMullick]

Okay , thanks.
I wanted to scope my viewModel to a particular group in a NavHost? How can i do that

[Tlaster]

As I said before, you can pass the root LocalStateHolder.current to every screen, here is a complete example:

val stateHolder = LocalStateHolder.current
NavHost(
    navigator = navigator,
    initialRoute = "/home",
) {
    scene("/home") {
        val viewModel = koinViewModel(NoteListViewModel::class, stateHolder = stateHolder)
    }
}