[RxSwift] Infallible

Infallinble is another flavor of `Observable` which is identical to it, but it's guaranteed to never fail and thus emit error.

 This means that when creating your own Infallible (using Infallible.create or one of the method mentioned in "Creating your first Observable"), you will not be allowed to emit errors.

 

 

 

Infallible은 Observable의 또 다른 유형 으로, Observable과 동일하지만 절대 실패하지 않으며, 에러를 방출할수 없습니다. 

즉,Infallible.create 또는 "Creating Your first Observable"에서 언급된 생성방법을 사용하여 Infallible을 만들때 , 에러를 방출하는것이 허용되지 않는다는 뜻입니다.

Infallible의 주요 특징 
1. 에러방출 불가능: onNext와 onCompleted만 방출
2. 에러 안전성 보장: 에러가 발생하지 않기에 단순한 처리가 가능
3. 주요 사용사례 : UI이벤트, 설정값 가져오기, 로컬캐시 데이터 등 

 

 

사용예제 ) UI event

let buttonTap = Infallible<String>.create {
	observer(.next("Button Tapped!"))
	observer(.completed)
	return Disposable.create
}

buttonTap
	.subscribe(.onNext: { print($0) })

 

 

 

 

Infallible is userful when you want to statically model and garantee a streem of values that  is known to never fail,

but don't want to commit to using `MainScheduler` and don't want to implictly user `share()` to share resources and side-effects, such as the case in Driver and Signal

 

Infallible 은 절대 실패하지 않는 값 스트림을 정적으로 모델링 하고 보장하고 싶을때 유용합니다.
이때 MainScheduler를 강제로 사용하지 않거나, Driver나 Signal처럼 share() 를 암묵적으로 사용해 리소의와 부작용을 공유하지 않으려는 경우에 특히 적합합니다.

Infallible은 에러를 방출하지 않는 데이터 스트림을 보장함 

Driver와 Signal은 기본적으로 MainScheduler를 사용하여 데이터를 방출하지만 Inallible은 Scheduler에 구애받지 않는 다는것이 큰 차이점 또한 Driver와 Signal은 기본적으로 share() 통해서 구독을 공유하여 리소스와 부작용을 방지 하지만 Infallible은 share()없이 각 구독이  독립적으로 동작할수 있음 

 

 

 

 

Life happens

so what if is's just too hard to solve come cases with custom operators?

You can exit the Rxmonad, perform actions in impreative world, and then tunnel results to Rx again using `Subject`s.

This isn't something that should be praticed often and is a bad smell, but you can do it

 

어떤 경우에는 Rx연산자 만으로 문제를 해결하기 어려울수 있습니다. 
이럴때는 RxMonad를 탈출하여 명령형 방식으로 작업을 수행한후,결과를 Rx로 다시 전달할수 있습니다. 이과정에서 Subjects를 사용할수 있습니다.

 

다만 이방법은 자주 사용해서는 안되며, 코드냄새가 나는 나쁜 방식으로 간주됩니다. 하지만 가능은 합니다.

 

let magicBeing: Observable<MagicBeing> = summonFromMiddleEarth()

magicBeings
	.subscribe(onNext: { being in  //Rx Monad 탈출
		self.doSomeStateMagic(being)
	})
	.disposed(by: DisposedBag)


let kitten = globalParty( // 명령형 방식으로 계산
	being.
	UIApplication.delegate.dataSomething.attendees
)
kitten.on(.next(kitten)) //결과를 다시 Rx로 전달

let kitten = BehaviorRelay(value: firstKitten) // 다시 Rx Monad로 돌아옴 

kittens.asObservable()
	.map {kitten in 
		return kitten.purr()
	}
// ...

 

이런 작업을 할때마다 다음과 같은 코드가 어딘가에 작성될 가능성이 큽니다.

 

kitten
	.subscribe( onNext: { kitten in 
	})
	.dispose(by: disposeBag)

 

 

그러니 이런방식은 피하는것이 좋습니다.


Rx Monad의 의미는 RxSwift의 선언형 프로그래밍 방식을 의미함 
Monad는 함수형 프로그래밍의 개념으로, 데이터를 체계적으로 연결하는 방식이라고 함
RxSwift에서는 Observable이 Monad로 동작하며, 데이터 스트림을 함수 체인으로 처리함 

>Monad에 대해서,,?  는 한번 공부해보면 좋을거 같다 ( 꽤 재미있어보이는 내용, scala/ Haskell)
>조금 검색해보다 보니 내용이 길어질것 같아서 나중에 다뤄보도록 하자 

Rx Monad는 데이터를 Observable 체인 안에서 처리하며, 명령형 코드(데이터에 직접 변경/조작 ) 를 지양함 
연산자 를 활용하여 데이터를 변환하고 처리하며 
부작용이 발생하지 않는 순수한 데이터 흐름을 지향함 

즉 Rx Monad를 탈출한다는 위와 같은 내용처럼 처리 하지 않고 체인의 밖에서 직접 명령형 방식으로 처리하거나 ,상태를 변경하는것으로 볼수있다.

 

 

일단 명령형 코드와 선언형 코드에 대한 기본지식이 어느정도 필요할것으로 생각되어서 정리를 해보자면 

 

명령형 코드는 
- 어떻게(**How**) 상태를 변경할지에 대한 초점.
- 상태변경과 작업과정을 명시적으로 작성

 

선언형 코드는 
- 무엇(**What**) 을 달성할지에 초점
- 데이터를 설명적으로 정의 하고 , 상태 변화는 내부적으로 처리.

 

 

명령형의 예시

class ViewController: UIViewController {
    @IBOutlet weak var label: UILabel!
    @IBOutlet weak var button: UIButton!
    
    override func viewDidLoad() {
        super.viewDidLoad()
        
        button.addTarget(self, action: #selector(buttonTapped), for: .touchUpInside)
    }

    @objc func buttonTapped() {
        label.text = "Button Tapped"
        label.textColor = .red
    }
}

 

선언형의 예시

class ViewController: UIViewController {
    @IBOutlet weak var label: UILabel!
    @IBOutlet weak var button: UIButton!
    
    let disposeBag = DisposeBag()

    override func viewDidLoad() {
        super.viewDidLoad()
        
        button.rx.tap
            .map { "Button Tapped" }
            .bind(to: label.rx.text)
            .disposed(by: disposeBag)
    }
}

 

한발 더나아가서 생각을 해본다면,,,?

UIKit은 기본적으로 명령형 프로그래밍 방식으로 설계되었다?

UIKit vs SwiftUI 에대한 부분이지 않을까 싶다. 

UIKit은 뷰와 상태를 직접 변경하고 text나 textColor 와 같은 속성들을 명시적으로 수정한다
또한 어떠한 이벤트를 처리하기위해 addTarget 메서드와 명령형 메서드를 실행하는 방식으로 돌아가는데 

상태기반으로 뷰를 업데이트 한다 
상태가 변경되면 뷰가 자동 업데이트 할수 있도록 동작한다.

(실제로 UIKit은 SwiftUI로 대체되어지는 중 이라고 생각한다. )