Is it possible to handle calculations after didApplyConstraints inside of a Delegating class? - swift

Question

I made a subclass of SKSpriteNode called JoyStick that follows the delegate pattern, where its delegate is GameScene. In an effort to further clean up the GameScene, the JoyStick class also implements its own touchesBegan, touchesMoved, and touchesEnded methods.
The problem is that I have some position calculations in the JoyStick's touchesMoved method that need to be made after certain SKConstraints have been applied to the joystick (such as bounding the joystick handle to the base). I know that there is a didApplyConstraints instance method of SKScene, and that the documentation indicates that it should be overridden and not called as it is already called once per frame.
However, I'm not sure how to integrate didApplyConstraints into my delegating joystick class's touchesMoved method without resorting to moving logic back into my GameScene's didApplyConstraints.
Here is a simplified example of what I currently have implemented with the delegate pattern. Note that the below code allows the variable joyVector to take on values outside of the acceptable range because the constraint hasn't yet been applied.
GameScene.swift
import SpriteKit
import GameplayKit
class GameScene: SKScene {
var shipNode: SKSpriteNode?
var joyStick: JoyStick?
override func didMove(to view: SKView) {
self.view?.isMultipleTouchEnabled = true
// Init the ship
guard let shipNode = self.childNode(withName: "ShipNode") as? SKSpriteNode else {
fatalError("Player node not loaded!")
}
self.shipNode = shipNode
// Init the joystick
joyStick = JoyStick()
joyStick?.isUserInteractionEnabled = true
joyStick?.position = CGPoint(x: -500, y: -200)
joyStick?.delegate = self
self.addChild(joyStick!)
}
}
extension GameScene: JoyStickDelegate{
var objPhysicsBody: SKPhysicsBody? {
return self.shipNode?.physicsBody
}
}
JoyStick.swift
import Foundation
import SpriteKit
protocol JoyStickDelegate: class {
var objPhysicsBody: SKPhysicsBody? {get}
}
class JoyStick: SKSpriteNode {
weak var delegate: JoyStickDelegate!
var joyVector: CGVector = CGVector()
var handle: SKSpriteNode?
init () {
let baseTexture = SKTexture(imageNamed: "joyStickBase")
super.init(texture: baseTexture, color: UIColor.clear, size: baseTexture.size())
// Add the joystick handle onto the base
handle = SKSpriteNode(imageNamed: "joyStickHandle")
handle?.position = CGPoint(x: 0 , y: 0)
// Add constraints
let range = CGFloat(self.size.width/2 - (handle?.size.width)! / 2)
let moveRange = SKRange(lowerLimit: 0, upperLimit: range)
let rangeConstraint = SKConstraint.distance(moveRange, to: CGPoint(x: 0,y: 0))
handle?.constraints = [rangeConstraint]
self.addChild(handle!)
}
required init?(coder aDecoder: NSCoder) {
fatalError("init(coder:) has not been implemented")
}
func moveJoyStick(_ touches: Set<UITouch>) {
guard let fingerPosition = touches.first?.location(in: self) else { return }
self.handle!.position = fingerPosition
}
func getJoyVector(handlePosition: CGPoint) -> CGVector {
return CGVector(dx: handlePosition.x, dy: handlePosition.y)
}
func resetJoyStick() {
delegate?.objPhysicsBody?.velocity = CGVector(dx: 0, dy: 0)
//animation to return the joystick to the center...
}
override func touchesBegan(_ touches: Set<UITouch>, with event: UIEvent?) {
self.moveJoyStick(touches)
joyVector = getJoyVector(handlePosition: self.handle!.position)
delegate?.objPhysicsBody?.velocity = joyVector
}
override func touchesMoved(_ touches: Set<UITouch>, with event: UIEvent?) {
self.moveJoyStick(touches)
joyVector = getJoyVector(handlePosition: self.handle!.position)
delegate?.objPhysicsBody?.velocity = joyVector
}
override func touchesEnded(_ touches: Set<UITouch>, with event: UIEvent?) {
resetJoyStick()
}
override func touchesEstimatedPropertiesUpdated(_ touches: Set<UITouch>) {}
override func touchesCancelled(_ touches: Set<UITouch>, with event: UIEvent?) {}
}
An alternative implementation that did work correctly was to keep the touch event methods inside the JoyStick class but also drop the delegation pattern all together. Instead, I accessed the JoyStick handlePosition (as a vector) inside GameScene via a getter method and applied it as a velocity inside GameScene's didApplyConstraints method. Here are the two snippets of added code to make this work:
Add to JoyStick.swift:
func getPositionVelocity() -> CGVector {
guard let handlePosition = self.handle?.position else { return CGVector() }
return CGVector(dx: handlePosition.x, dy: handlePosition.y)
}
Add to GameScene.swift:
...
//joyStick?.delegate = self
...
override func didApplyConstraints() {
self.shipNode?.physicsBody?.velocity = self.joyStick?.getPositionVelocity() ?? CGVector()
}
While this achieves the desired behavior of only applying velocities bounded by the radius of the joystick, it seems far less elegant and greatly reduces the amount of encapsulation/generality that I had with the delegate pattern - which becomes noticeable as the game grows.
Ultimately, is it possible to avoid piling in post-constraint logic into a single didApplyConstraints definition in GameScene to achieve this functionality? And why is it that UIResponder touch events seem to always be handled before :didApplyConstraints - even though they seem to be outside of the SKScene Frame-Cycle?

Answer 1

You have two objects with different methods inside. You can not move both methods into one class directly.
But it is possible to do in this way:
class GameScene: SKScene {
var onJoysticNeedVerticalPosition: (()-> CGVector)?
....
override func didApplyConstraints() {
self.shipNode?.physicsBody?.velocity = onJoysticNeedVerticalPosition?() ?? CGVector()
}
}
and
protocol JoyStickDelegate: class {
var objPhysicsBody: SKPhysicsBody? {get}
var onJoysticNeedVerticalPosition: (()-> CGVector)? {get set}
}
then
class JoyStick: SKSpriteNode {
...
init () {
self.delegate?.onJoysticNeedVerticalPosition = {[weak self] in
self?.getPositionVelocity()
}
}
...
func getPositionVelocity() -> CGVector {
guard let handlePosition = self.handle?.position else { return CGVector() }
return CGVector(dx: handlePosition.x, dy: handlePosition.y)
}
}
So in fact it is still two different objects, but the first one (GameScene) hold the closure with calculation logic, that in fact calculated in second one (JoyStick), and second one give it in detail to the first one by using delegate method.