Inout parameter in async callback does not work as expected

Question

I'm trying to insert functions with inout parameter to append data received from async callback to an outside array. However, it does not work. And I tried everything I know to find out why - with no luck.

As advised by @AirspeedVelocity, I rewrote the code as follows to remove unnecessary dependencies. I also use an Int as the inout parameter to keep it simple.
The output is always:
c before: 0
c after: 1

I'm not able to figure out what goes wrong here.

func getUsers() {
    let u = ["bane", "LiweiZ", "rdtsc", "ssivark", "sparkzilla", "Wogef"]
    var a = UserData()
    a.userIds = u
    a.dataProcessor()
}

struct UserData {
    var userIds = [String]()
    var counter = 0
    mutating func dataProcessor() -> () {
        println("counter: \(counter)")
        for uId in userIds {
            getOneUserApiData(uriBase + "user/" + uId + ".json", &counter)
        }
    }
}

func getOneUserApiData(path: String, inout c: Int) {
    var req = NSURLRequest(URL: NSURL(string: path)!)
    var config = NSURLSessionConfiguration.ephemeralSessionConfiguration()
    var session = NSURLSession(configuration: config)
    var task = session.dataTaskWithRequest(req) {
        (data: NSData!, res: NSURLResponse!, err: NSError!) in
        println("c before: \(c)")
        c++
        println("c after: \(c)")
        println("thread on: \(NSThread.currentThread())")
    }

    task.resume()
}

Thanks.

Answer 1

Sad to say, modifying inout parameter in async-callback is meaningless.

From the official document:

Parameters can provide default values to simplify function calls and can be passed as in-out parameters, which modify a passed variable once the function has completed its execution.

...

An in-out parameter has a value that is passed in to the function, is modified by the function, and is passed back out of the function to replace the original value.

Semantically, in-out parameter is not "call-by-reference", but "call-by-copy-restore".

In your case, counter is write-backed only when getOneUserApiData() returns, not in dataTaskWithRequest() callback.

Here is what happened in your code

1. at getOneUserApiData() call, the value of counter 0 copied to c₁
2. the closure captures c₁
3. call dataTaskWithRequest()
4. getOneUserApiData returns, and the value of - unmodified - c₁ is write-backed to counter
5. repeat 1-4 procedure for c₂, c₃, c₄ ...
6. ... fetching from the Internet ...
7. callback is called and c₁ is incremented.
8. callback is called and c₂ is incremented.
9. callback is called and c₃ is incremented.
10. callback is called and c₄ is incremented.
11. ...

As a result counter is unmodified :(

---

Detailed explaination

Normally, in-out parameter is passed by reference, but it's just a result of compiler optimization. When closure captures inout parameter, "pass-by-reference" is not safe, because the compiler cannot guarantee the lifetime of the original value. For example, consider the following code:

func foo() -> () -> Void {
    var i = 0
    return bar(&i)
}

func bar(inout x:Int) -> () -> Void {
    return {
        x++
        return
    }
}

let closure = foo()
closure()

In this code, var i is freed when foo() returns. If x is a reference to i, x++ causes access violation. To prevent such race condition, Swift adopts "call-by-copy-restore" strategy here.

Answer 2

Essentially it looks like you’re trying to capture the “inout-ness” of an input variable in a closure, and you can’t do that – consider the following simpler case:

// f takes an inout variable and returns a closure
func f(inout i: Int) -> ()->Int {
    // this is the closure, which captures the inout var
    return {
        // in the closure, increment that var
        return ++i
    }

}

var x = 0
let c = f(&x)

c() // these increment i
c()
x   // but it's not the same i

At some point, the variable passed in ceases to be x and becomes a copy. This is probably happening at the point of capture.

edit: @rintaro’s answer nails it – inout is not in fact semantically pass by reference

If you think about it this makes sense. What if you did this:

// declare the variable for the closure
var c: ()->Int = { 99 }

if 2+2==4 {
    // declare x inside this block
    var x = 0
    c = f(&x)
}

// now call c() - but x is out of scope, would this crash?
c()

When closures capture variables, they need to be created in memory in such a way that they can stay alive even after the scope they were declared ends. But in the case of f above, it can’t do this – it’s too late to declare x in this way, x already exists. So I’m guessing it gets copied as part of the closure creation. That’s why incrementing the closure-captured version doesn’t actually increment x.