How to generate and handle errors in Swift - ios

Question

I'm working on an app where users need to login. The webapplication uses a token, that users can get from calling a webservice using their username and password. My question now is what's the best way the handle errors that can occur. What I have now:
LoginViewController.swift
self.api.token(forUsername: "a", password: "b") { (result) in
switch (result) {
case .failure(let error):
print("something when wrong \(error)")
case .success(let token):
print("token \(token)")
}
LoginAPIClient (the self.api) using https://github.com/3lvis/Networking
class LoginAPIClient {
enum ResponseError: Error {
case unexpectedDataError
case unknownError
}
enum Result {
case success(String)
case failure(Error)
}
lazy var networking: Networking = {
let networking = Networking(baseURL: APIClient.serverURL())
return networking
}()
func token(forUsername username: String, password: String, completion: #escaping (Result) -> Void) {
let parameters = ["username" : username, "password" : password]
networking.post("/api/login", parameters: parameters) { (result) in
switch result {
case .failure(let response):
return completion(.failure(response.error))
case .success(let response):
if var token = response.headers["Authorization"] as? String {
token = token.replacingOccurrences(of: "Bearer ", with: "")
return completion(.success(token))
}
return completion(.failure(ResponseError.unknownError))
}
}
}
}
Here for example I'm creating my own error return completion(.failure(ResponseError.unknownError)) in case the server responds with a successful status code (200) but somehow the Authorization header is missing from the response.
This works, the only problem is that now when I handle the error in the ViewController I don't know the exact reason why it fails. For example, from the Networking library I get an error code (400 or 401 etc) but this is lost because first it was an NSError. I could use an NSError but somehow this doesn't feel right. Could anyone point me in the right direction?
One of the solutions I thought of was to add an extra enum and then do something like this:
enum Result {
case success(String)
case networkFailure(FailureJSONResponse)
case failure(Error)
}
self.api.token(forUsername: "a", password: "b") { (result) in
switch (result) {
case .failure(let error):
print("something when wrong \(error)")
case .networkFailure(let response):
print("something when wrong \(error)")
case .success(let token):
print("token \(token)")
}
But I rather have just 1 success and 1 failure in the switch.

Answer 1

Every application usually has its own "Error", so in your case you can define
public enum AppError: Swift.Error, CustomStringConvertible {
case networkError(code: Int)
case anotherError(message: String?)
case underlying(Swift.Error)
public var description: String {
switch self {
case .networkError(let code):
return "Network error with code \(code)"
default://I'm not gonna cover all cases, but you should :)
return "Error"
}
}
}
then you can use your approach
enum Result {
case success(String)
case failure(AppError)
}

Answer 2

You can add second parameter to completion block:
func token(forUsername username: String, password: String, completion: #escaping (Result, error: ResponseError?) -> Void)