To Coroutine, or Not to Coroutine

Unity Engine
1

I have a design question.

I’ve found a few times I have come across a method which makes sense to have a coroutine and non-coroutine version.

For example, let’s say I have a State class, and that State class has two methods: Activate() and Deactivate(). State.Activate() is called when the State is first enabled, and handles a transition to that state. Similarly, State.Deactivate() is called before a new state is activated.

In some instances, I just want to call State.Deactivate() and leave it at that. But then in other instances, I want to script a sequence, so I’d do something like this instead:

yield return StartCoroutine(OldState.Deactivate());
NewState.Activate();

Without getting into a discussion about better ways to handle State management (LOL!), I’m wondering if anyone has come across this problem of wanting a coroutine and non-coroutine version of a method, and how you handle it when you do?

Do you simply have two version of the method?

void Activate();
void ActivateCoroutine();

Or do you do something much more clever?

2

I do stupid things like:

enum stateType{fart, sit, idle, punch, whatever}
var state:stateType = stateType.idle;

Then it’s just a switch statement for (state) in the code. That’ll be running regardless forever. I don’t really optimise or make things confusing. I probably should. In general, I just do less if I want it faster.

I think all this Activate and Deactivate is a recipe for sphagetti code.

3

Leaving aside the whole question of state implementation (I’ve spent many a sleepless night pondering an elegant solution to that issue), I do exactly what you suggest - simply write two versions of the method and call whichever is appropriate.

4

Unless you specifically need yield, you could start Activate via InvokeRepeating, then you don’t need two separate versions.

http://unity3d.com/support/documentation/ScriptReference/MonoBehaviour.InvokeRepeating.html