Retrying-an-operation-if-it-fails.md

Retrying an operation if it fails

An AOperation can be retryable if you adopt RetryableOperation protocol to it. This protocol has one method that is required to conform. func new() -> Self which you should retrun a new object of conformed operation in its implementation.

Consider this operation:

class FetchUserInfoOperation: ResultableOperation<UserInfo> {

	override public func execute() {
	// Som task is done here
	}

}

By adopting RetryableOperation to this operation we make it retryable.

class FetchUserInfoOperation: ResultableOperation<UserInfo>, RetryableOperation {

	public func new() -> Self {
		FetchUserInfoOperation() as! Self
	}


	override public func execute() {
	// Som task is done here
	}

}

Use this feature in operation execution like below.

FetchUserInfoOperation()
.retryOnFailure({(numberOfRetries, error, retry) in
	retry(true)
}
.didFinish { result
//Update UI
}
.add(to: queue)

By calling retryOnFailure(_:) method on operation you can manage when it should be retry. This method has a closure with three parameter. A numberOfRetries which shows how many time you retried since now. At the first time operation fails the value of this parameter is 0, which means the operation doesn't retry yet. An error which is error of operation failure. A retry method that should be called at the end of closure and has an boolean input which tells to the operation wether it should execute again (true) or completely finished(false).

In another example we could manage retryOnFailure closure like below.

FetchUserInfoOperation()
.retryOnFailure({(numberOfRetries, error, retry) in
	if (error.publishedError as? URLError).errorCode == URLError.Code.timedOut {
		retry(true)
	}
	else {
		retry(false)
	}
}
.didFinish { result
//Update UI
}
.add(to: queue)

Note that didFinish(_:) doesn't call if you set retry(_:) input as true.