Working and testing async Vapor instructions

The right way to run async instructions in Vapor?

The async / await function is comparatively new in Swift and a few framework authors have not transformed all the things to reap the benefits of these new key phrases. At the moment, that is the state of affairs with the Command API in Vapor 4. You’ll be able to already outline async instructions, however there is not any technique to register them utilizing the Vapor framework. Thankfully, there’s a comparatively simple workaround that you should use if you wish to execute instructions utilizing an asynchronous context. 🔀

First we’ll outline a helper protocol and create an asyncRun perform. We’re going to lengthen the unique Command protocol and supply a default implementation for the run technique.

import Vapor

public protocol AsyncCommand: Command {
    
    func asyncRun(
        utilizing context: CommandContext,
        signature: Signature
    ) async throws
}

public extension AsyncCommand {

    func run(
        utilizing context: CommandContext,
        signature: Signature
    ) throws {
        let promise = context
            .utility
            .eventLoopGroup
            .subsequent()
            .makePromise(of: Void.self)
        
        promise.completeWithTask {
            strive await asyncRun(
                utilizing: context,
                signature: signature
            )
        }
        strive promise.futureResult.wait()
    }
}

This fashion it’s best to be capable of create a brand new async command and it’s best to implement the asyncRun technique if you wish to name some asynchronous Swift code.

import Vapor

closing class MyAsyncCommand: AsyncCommand {
    
    static let identify = "async"
    
    let assist = "This command run asynchronously."

    struct Signature: CommandSignature {}

    func asyncRun(
        utilizing context: CommandContext,
        signature: Signature
    ) async throws {
        context.console.data("That is async.")
    }
}

It’s attainable to register the command utilizing the configure technique, you possibly can do that out by operating the swift run Run async snippet if you’re utilizing the usual Vapor template. 💧

import Vapor

public func configure(
    _ app: Utility
) throws {

    app.instructions.use(
        MyAsyncCommand(),
        as: MyAsyncCommand.identify
    )

    strive routes(app)
}

As you possibly can see it is a fairly neat trick, it is also talked about on GitHub, however hopefully we do not want this workaround for too lengthy and correct async command help will arrive in Vapor 4.x.

Unit testing Vapor instructions

This matter has actually zero documentation, so I believed it could be good to let you know a bit about find out how to unit check scripts created through ConsoleKit. To start with we’d like a TestConsole that we are able to use to gather the output of our instructions. This can be a shameless ripoff from ConsoleKit. 😅

import Vapor

closing class TestConsole: Console {

    var testInputQueue: [String]
    var testOutputQueue: [String]
    var userInfo: [AnyHashable : Any]

    init() {
        self.testInputQueue = []
        self.testOutputQueue = []
        self.userInfo = [:]
    }

    func enter(isSecure: Bool) -> String {
        testInputQueue.popLast() ?? ""
    }

    func output(_ textual content: ConsoleText, newLine: Bool) {
        let line = textual content.description + (newLine ? "n" : "")
        testOutputQueue.insert(line, at: 0)
    }

    func report(error: String, newLine: Bool) {
        
    }

    func clear(_ sort: ConsoleClear) {
        
    }

    var measurement: (width: Int, peak: Int) {
        (0, 0)
    }
}

Now contained in the check suite, it’s best to create a brand new utility occasion utilizing the check setting and configure it for testing functions. Then it’s best to provoke the command that you just’d like to check and run it utilizing the check console. You simply need to create a brand new context and a correct enter with the required arguments and the console.run perform will handle all the things else.

@testable import App
import XCTVapor

closing class AppTests: XCTestCase {
    
    func testCommand() throws {
        let app = Utility(.testing)
        defer { app.shutdown() }
        strive configure(app)
        
        let command = MyAsyncCommand()
        let arguments = ["async"]
        
        let console = TestConsole()
        let enter = CommandInput(arguments: arguments)
        var context = CommandContext(
            console: console,
            enter: enter
        )
        context.utility = app
        
        strive console.run(command, with: context)

        let output = console
            .testOutputQueue
            .map { $0.trimmingCharacters(in: .whitespacesAndNewlines) }
        
        let expectation = [
            "This is async."
        ]
        XCTAssertEqual(output, expectation)
    }
}

The great factor about this resolution is that the ConsoleKit framework will routinely parse the arguments, choices and the flags. You’ll be able to present these as standalone array parts utilizing the enter arguments array (e.g. ["arg1", "--option1", "value1", "--flag1"]).

It’s attainable to check command teams, you simply have so as to add the precise command identify as the primary argument that you just’d prefer to run from the group and you may merely examine the output via the check console if you’re searching for the precise command outcomes. 💪