I always took error types with custom messages for granted in other languages, so I was a bit surprised at the amount of work I had to do to implement them in Swift. I had to spend a while digging through outdated Stack Overflow posts before coming upon this solution, so hopefully this will be useful to you!

The Swift documentation on the Error protocol spends a while discussing how to add associated values to error enums, and even lists

var localizedDescription: String

as part of the protocol. Yet, neither the associated values nor the localizedDescription are actually printed out when we throw an error! See the following example:

import XCTest

enum MyError: Error {
    case first(message: String)
    case second(message: String)

    var localizedDescription: String { return "Some description here!" }
}

final class ErrorTest: XCTestCase {
    static var allTests: [(String, (ErrorTest) -> () throws -> Void)] {
        return [
            ("testMyError", testMyError)
        ]
    }

    func testMyError() throws {
        throw MyError.first(message: "first error")
    }
}

This looks like it should print the associated messages, or maybe the localizedDescription, when we throw the error. And yet, my output, running the test from the command line:

failed: caught error: The operation couldn’t be completed.(MyTests.MyError error 0.)

While it does tell me that the first error case was thrown (error 0), I don’t get any of the useful information I worked so hard to store with my error!

The solution I found for this was that, rather than implementing Error, you can implement the LocalizedError protocol instead (which inherits from Error.) This protocol specifies a property:

var errorDescription: String? { get }

which is printed out when the error is thrown.

Modifying the previous example to implement LocalizedError, we get:

enum MyError: LocalizedError {
    case first(message: String)
    case second(message: String)

    var errorDescription: String? { return "Some description here!" }
}

When we run the same test as before, the output is now:

failed: caught error: Some description here!

Finally, what if we want to take advantage of our error’s associated values to give better error messages? We can get fancier with the description, and do something like:

enum MyError: LocalizedError {
    case first(message: String)
    case second(message: String)

    var errorDescription: String? {
        switch self {
        case let .first(message), let .second(message):
            return message
        }
    }
}

Now, when we run the test, we get:

failed: caught error: first error

Exactly the message we specified when we originally threw the error!

Thanks for reading, and let me know any questions or feedback you have via Twitter or email 🙂