Best way to implement Finite State Machine for player character in Unity?

Question

I'd like to implement the player in my game as an FSM as described in this guide to programming patterns in games.

What is the best way to accomplish this? Should I implement each state as a monobehaviour with a state interface that can be enabled/disabled when the state transitions? Or would it better to the states as plain classes (with their own Update method) where the main player monobehaviour calls StateName.Update() from its own Update method (.e.g. void Update() { currentState.Update() })

Answer 1

I personally like to create a private enum with different states, with a private States variable that holds my actual state. Then in the Update function I create a switch-case to test in which state I am right now, and I call the relative function. Then inside the called function I assign the new value to the actual state. This is an easy to mantain solution, in my opinion.

{
    private enum States { state_0, state_1, state_2 };
    private States state_now;

    void Start () {
        state_now = States.state_0;
    }

    void Update () {
        switch (state_now) {
            case States.state_0 : state_0(); break;
            case States.state_1 : state_1(); break;
            case States.state_2 : state_2(); break;
            default : break;
        }

    void state_0() {

        // Business logic
        state_now = States.state_1;
    }

    // Others function, one for each state
}

I wrote this snippet without Unity on my PC to give you an idea, it's not intended to be a complete example.