Reducer

Reducer is a function that takes the current state and an action and returns a new state. This new state would then be saved as the current state to be used on the next call to the reducer. In Kaskade, there's a reducer created for every action.

on<TodoAction.Update> {
    todoRepository.updateItem(action.todoItem)
    TodoState.OnUpdated(action.position, action.todoItem)
}

In the DSL example the reducer is abstracted as the lambda in the on method. It has an ActionState, which is a pair of action and state, receiver. ActionState is for the DSL to easily access both action and state.

Extending the Reducer

The interface of a Reducer is:

interface Reducer<ACTION : Action, STATE : State> {
    operator fun invoke(action: ACTION, state: STATE, onStateChanged: (state: STATE) -> Unit)
}

The reducer is being invoked every time an action is processed.

The default reducer is the BlockingReducer which synchronously runs the lambda in the on method of the DSL.

class BlockingReducer<ACTION : Action, STATE : State>(
    private val transformerFunction: ActionState<ACTION, STATE>.() -> STATE
) : Reducer<ACTION, STATE> {
    override fun invoke(action: ACTION, state: STATE, onStateChanged: (state: STATE) -> Unit) {
        onStateChanged(getState(action, state))
    }

    fun getState(action: ACTION, state: STATE) = transformerFunction(ActionState(action, state))
}

It is not a good idea to run long running tasks on it. For asynchronous purposes there's a ScopedReducer and RxReducer in the Kaskade modules.