I was able to decouple the visuals from the action system by having two events on the Unit class. Now anything can select or deselect units based on its own logic:
public event Action OnUnitSelected;
public event Action OnUnitDeSelected;
public void Select() {
OnUnitSelected?.Invoke();
}
public void DeSelect() {
OnUnitDeSelected?.Invoke();
}
Then in the UnitActionSystem it can select, or deselect units:
private void SetSelectedUnit(Unit unit) {
// Warning about lifetime check omitted, not important in this example
selectedUnit?.DeSelect();
unit.Select();
selectedUnit = unit;
}
Now, the UnitSelectedVisual class only cares about the state of the unit its attached to, and its events, and has no coupling to any other classes that can alter the selected state of the unit:
private void Start() {
unit.OnUnitSelected += HandleSelectedUnitSelected;
unit.OnUnitDeSelected += HandleSelectedUnitDeSelected;
// Show in edit mode, disable when play mode starts
_meshRenderer.enabled = false;
}
private void HandleSelectedUnitSelected() {
_meshRenderer.enabled = true;
}
private void HandleSelectedUnitDeSelected() {
_meshRenderer.enabled = false;
}
Of course, I haven’t gone through this whole course, so implementing it this way may give me a huge headache later trying to convert from the intended way to my way. So warning to anyone else who wants to try this.