Suppose I have a Swift class like this:
#objc final MyClass : NSObject
{
let classPropertyString = "A class property"
func doStuff()
{
let localString = "An object local to this function"
DispatchQueue.global(qos: .userInitiated).async { [classPropertyString] in
// Do things with 'classPropertyString' and 'localString'
}
}
}
My question is: when I write a capture list, am I responsible for EXHAUSTIVELY listing all the things to which I want the closure to hold a strong reference?
In other words, if I omit localString from my capture list (as I've done here), will the closure still automatically capture a strong reference to it or am I in for a bad time?
There are several minor quirks with your question that make it tricky to answer clearly, but I think I understand the underlying concern, and the short answer is "no." But your example is impossible, so the answer is "it's impossible." And if it were possible, there'd be no strong reference (nor would there be a need for one), so the question still would be kind of "it's impossible." Even so, let's walk through what's going on here.
First, closure can't reference localString unless it's reassigned somehow in the comment inside doStuff(). closure is assigned at a level where localString is not in scope. Closures can only capture variables that are in scope when they are assigned, not when they're called. But let's go back to the original version of this question, before it was edited. That version did have the case you're describing:
#objc final myClass : NSObject
{
let classPropertyString = "A class property"
func doStuff()
{
let localString = "An object local to this function"
DispatchQueue.global(qos: .userInitiated).async { [classPropertyString] in // (1)
// Do things with 'classPropertyString' and 'localString'
}
// (2)
}
}
There's no problems here. classPropertyString is copied into the closure, avoiding any retain loops. localString is referenced by the closure, and so it's preserved as long as the closure exists.
Because you listed classPropertyString in the capture list, it is evaluated at point (1) and copied into the closure. Because you implicitly captured localString, it is treated as a reference. See Capture Lists in the Swift Programming Language Reference for some excellent examples of exactly how this works in different cases.
In no case (*) will Swift allow the underlying storage for something you're using in a closure to disappear behind your back. That's why the typical concern is excessive retains (memory leaks) rather than dangling references (crashes).
(*) "In no case" here is a lie. There are several ways that Swift will allow it, but almost all of them involve "Unsafe" which is your warning about that. The major exception is unowned, and of course anything involving ! types. And Swift is not typically thread-safe, so you need to be careful about that...
The last comment about thread-safety is a place where the subtle distinctions between implicit and explicit captures can really matter. Consider this case where you modify an implicitly captured value on two queues:
func doStuff() -> String
{
var localString = "An object local to this function"
DispatchQueue.global(qos: .userInitiated).async {
localString = "something else"
callFunction(localString)
}
localString = "even more changes"
return localString
}
What happens in that case? Good grief, never do that. I believe it's undefined behavior and that localString could be anything including corrupted memory, at least in the most general case (it might be defined behavior for calling .async; I'm not sure). But don't do it.
But for your normal cases, there is no reason to explicitly capture local variables. (I sometimes wish Swift had gone the C++ way and said it was required, but it isn't.)
Ok, one more way implicit and explicit are different that might drive home how they work. Consider a stateful closure like this (I build these pretty often):
func incrementor() -> () -> Int {
var n = 0
return {
n += 1
return n
}
}
let inc = incrementor()
inc() // 1
inc() // 2
inc() // 3
let inc2 = incrementor()
inc2() // 1
See how the local variable n is captured by the closure, and can be modified after it goes out of scope. And see how inc2 has its own version of that local variable. Now try that with explicit capture.
func incrementor() -> () -> Int {
var n = 0
return { [n] in // <---- add [n]
n += 1 // Left side of mutating operator isn't mutable: 'n' is an immutable capture
return n
}
}
Explicit captures are copies and they're immutable. Implicit captures are references, and so have the same mutability as the thing they reference.
Related
It seems logical to me that escaping closures would capture structs by copying. But if that was the case, the following code makes no sense and should not compile:
struct Wtf {
var x = 1
}
func foo(){
var wtf = Wtf()
DispatchQueue.global().async {
wtf.x = 5
}
Thread.sleep(forTimeInterval: 2)
print("x = \(wtf.x)")
}
Yet it compiles successfully and even prints 5 when foo is called. How is this possible?
While it might make sense for a struct to be copied, as your code demonstrates, it is not. That's a powerful tool. For example:
func makeCounter() -> () -> Int {
var n = 0
return {
n += 1 // This `n` is the same `n` from the outer scope
return n
}
// At this point, the scope is gone, but the `n` lives on in the closure.
}
let counter1 = makeCounter()
let counter2 = makeCounter()
print("Counter1: ", counter1(), counter1()) // Counter1: 1 2
print("Counter2: ", counter2(), counter2()) // Counter2: 1 2
print("Counter1: ", counter1(), counter1()) // Counter1: 3 4
If n were copied into the closure, this couldn't work. The whole point is the closure captures and can modify state outside itself. This is what separates a closure (which "closes over" the scope where it was created) and an anonymous function (which does not).
(The history of the term "close over" is kind of obscure. It refers to the idea that the lambda expression's free variables have been "closed," but IMO "bound" would be a much more obvious term, and is how we describe this everywhere else. But the term "closure" has been used for decades, so here we are.)
Note that it is possible to get copy semantics. You just have to ask for it:
func foo(){
var wtf = Wtf()
DispatchQueue.global().async { [wtf] in // Make a local `let` copy
var wtf = wtf // To modify it, we need to make a `var` copy
wtf.x = 5
}
Thread.sleep(forTimeInterval: 2)
// Prints 1 as you expected
print("x = \(wtf.x)")
}
In C++, lambdas have to be explicit about how to capture values, by binding or by copying. But in Swift, they chose to make binding the default.
As to why you're allowed to access wtf after it's been captured by the closure, that's just a lack of move semantics in Swift. There's no way in Swift today to express "this variable has been passed to something else and may no longer be accessed in this scope." That's a known limitation of the language, and a lot of work is going into fix it. See The Ownership Manifesto for more.
I'm rewriting some code to append jobs to an array of closures rather than execute them directly:
var someObject: SomeType?
var jobsArray: [() -> ()] = []
// before rewriting
doExpensiveOperation(someObject!.property)
// 1st attempt at rewriting
jobsArray.append {
doExpensiveOperation(someObject!.property)
}
However, because the value of someObject might change before the closure is executed, I'm now adding a closure list as follows:
// 2nd attempt at rewriting
jobsArray.append { [someObject] in
doExpensiveOperation(someObject!.property)
}
Hopefully then if, say, someObject is subsequently set to nil before the closure executes, the closure will still access the intended instance.
But what's the neatest way to deal with the possibility that the value of .property might change before the closure is executed? Is there a better way than this?
// 3rd attempt at rewriting
let p = someObject!.property
jobsArray.append {
doExpensiveOperation(p)
}
I'm not very keen on this solution because it means changing the original line of code. I'd prefer this but it doesn't work:
// 4th attempt at rewriting
jobsArray.append { [someObject!.property] in
doExpensiveOperation(someObject!.property)
}
Pretty new to Swift, so all guidance gratefully received. Thanks!
A capture list such as [someObject] is actually syntactic sugar for [someObject = someObject], where the right hand side can be an arbitrary expression that gets bound to a new constant upon the closure being formed.
Therefore one option is to write your example as:
jobsArray.append { [property = someObject!.property] in
doExpensiveOperation(property)
}
Reference cycles in Swift occur when properties of reference types have strong ownership of each other (or with closures).
Is there, however, a possibility of having reference cycles with value types only?
I tried this in playground without succes (Error: Recursive value type 'A' is not allowed).
struct A {
var otherA: A? = nil
init() {
otherA = A()
}
}
A reference cycle (or retain cycle) is so named because it indicates a cycle in the object graph:
Each arrow indicates one object retaining another (a strong reference). Unless the cycle is broken, the memory for these objects will never be freed.
When capturing and storing value types (structs and enums), there is no such thing as a reference. Values are copied, rather than referenced, although values can hold references to objects.
In other words, values can have outgoing arrows in the object graph, but no incoming arrows. That means they can't participate in a cycle.
As the compiler told you, what you're trying to do is illegal. Exactly because this is a value type, there's no coherent, efficient way to implement what you're describing. If a type needs to refer to itself (e.g., it has a property that is of the same type as itself), use a class, not a struct.
Alternatively, you can use an enum, but only in a special, limited way: an enum case's associated value can be an instance of that enum, provided the case (or the entire enum) is marked indirect:
enum Node {
case None(Int)
indirect case left(Int, Node)
indirect case right(Int, Node)
indirect case both(Int, Node, Node)
}
Disclaimer: I'm making an (hopefully educated) guess about the inner workings of the Swift compiler here, so apply grains of salt.
Aside from value semantics, ask yourself: Why do we have structs? What is the advantage?
One advantage is that we can (read: want to) store them on the stack (resp. in an object frame), i.e. just like primitive values in other languages. In particular, we don't want to allocate dedicated space on the heap to point to. That makes accessing struct values more efficient: we (read: the compiler) always knows where exactly in memory it finds the value, relative to the current frame or object pointer.
In order for that to work out for the compiler, it needs to know how much space to reserve for a given struct value when determining the structure of the stack or object frame. As long as struct values are trees of fixed size (disregarding outgoing references to objects; they point to the heap are not of interest for us), that is fine: the compiler can just add up all the sizes it finds.
If you had a recursive struct, this fails: you can implement lists or binary trees in this way. The compiler can not figure out statically how to store such values in memory, so we have to forbid them.
Nota bene: The same reasoning explains why structs are pass-by-value: we need them to physically be at their new context.
Quick and easy hack workaround: just embed it in an array.
struct A {
var otherA: [A]? = nil
init() {
otherA = [A()]
}
}
You normally cannot have a reference cycle with value types simply because Swift normally doesn't allow references to value types. Everything is copied.
However, if you're curious, you actually can induce a value-type reference cycle by capturing self in a closure.
The following is an example. Note that the MyObject class is present merely to illustrate the leak.
class MyObject {
static var objCount = 0
init() {
MyObject.objCount += 1
print("Alloc \(MyObject.objCount)")
}
deinit {
print("Dealloc \(MyObject.objCount)")
MyObject.objCount -= 1
}
}
struct MyValueType {
var closure: (() -> ())?
var obj = MyObject()
init(leakMe: Bool) {
if leakMe {
closure = { print("\(self)") }
}
}
}
func test(leakMe leakMe: Bool) {
print("Creating value type. Leak:\(leakMe)")
let _ = MyValueType(leakMe: leakMe)
}
test(leakMe: true)
test(leakMe: false)
Output:
Creating value type. Leak:true
Alloc 1
Creating value type. Leak:false
Alloc 2
Dealloc 2
Is there, however, a possibility of having reference cycles with value types only?
Depends on what you mean with "value types only".
If you mean completely no reference including hidden ones inside, then the answer is NO. To make a reference cycle, you need at least one reference.
But in Swift, Array, String or some other types are value types, which may contain references inside their instances. If your "value types" includes such types, the answer is YES.
Reference cycles in Swift occur when properties of reference types have strong ownership of each other (or with closures).
Is there, however, a possibility of having reference cycles with value types only?
I tried this in playground without succes (Error: Recursive value type 'A' is not allowed).
struct A {
var otherA: A? = nil
init() {
otherA = A()
}
}
A reference cycle (or retain cycle) is so named because it indicates a cycle in the object graph:
Each arrow indicates one object retaining another (a strong reference). Unless the cycle is broken, the memory for these objects will never be freed.
When capturing and storing value types (structs and enums), there is no such thing as a reference. Values are copied, rather than referenced, although values can hold references to objects.
In other words, values can have outgoing arrows in the object graph, but no incoming arrows. That means they can't participate in a cycle.
As the compiler told you, what you're trying to do is illegal. Exactly because this is a value type, there's no coherent, efficient way to implement what you're describing. If a type needs to refer to itself (e.g., it has a property that is of the same type as itself), use a class, not a struct.
Alternatively, you can use an enum, but only in a special, limited way: an enum case's associated value can be an instance of that enum, provided the case (or the entire enum) is marked indirect:
enum Node {
case None(Int)
indirect case left(Int, Node)
indirect case right(Int, Node)
indirect case both(Int, Node, Node)
}
Disclaimer: I'm making an (hopefully educated) guess about the inner workings of the Swift compiler here, so apply grains of salt.
Aside from value semantics, ask yourself: Why do we have structs? What is the advantage?
One advantage is that we can (read: want to) store them on the stack (resp. in an object frame), i.e. just like primitive values in other languages. In particular, we don't want to allocate dedicated space on the heap to point to. That makes accessing struct values more efficient: we (read: the compiler) always knows where exactly in memory it finds the value, relative to the current frame or object pointer.
In order for that to work out for the compiler, it needs to know how much space to reserve for a given struct value when determining the structure of the stack or object frame. As long as struct values are trees of fixed size (disregarding outgoing references to objects; they point to the heap are not of interest for us), that is fine: the compiler can just add up all the sizes it finds.
If you had a recursive struct, this fails: you can implement lists or binary trees in this way. The compiler can not figure out statically how to store such values in memory, so we have to forbid them.
Nota bene: The same reasoning explains why structs are pass-by-value: we need them to physically be at their new context.
Quick and easy hack workaround: just embed it in an array.
struct A {
var otherA: [A]? = nil
init() {
otherA = [A()]
}
}
You normally cannot have a reference cycle with value types simply because Swift normally doesn't allow references to value types. Everything is copied.
However, if you're curious, you actually can induce a value-type reference cycle by capturing self in a closure.
The following is an example. Note that the MyObject class is present merely to illustrate the leak.
class MyObject {
static var objCount = 0
init() {
MyObject.objCount += 1
print("Alloc \(MyObject.objCount)")
}
deinit {
print("Dealloc \(MyObject.objCount)")
MyObject.objCount -= 1
}
}
struct MyValueType {
var closure: (() -> ())?
var obj = MyObject()
init(leakMe: Bool) {
if leakMe {
closure = { print("\(self)") }
}
}
}
func test(leakMe leakMe: Bool) {
print("Creating value type. Leak:\(leakMe)")
let _ = MyValueType(leakMe: leakMe)
}
test(leakMe: true)
test(leakMe: false)
Output:
Creating value type. Leak:true
Alloc 1
Creating value type. Leak:false
Alloc 2
Dealloc 2
Is there, however, a possibility of having reference cycles with value types only?
Depends on what you mean with "value types only".
If you mean completely no reference including hidden ones inside, then the answer is NO. To make a reference cycle, you need at least one reference.
But in Swift, Array, String or some other types are value types, which may contain references inside their instances. If your "value types" includes such types, the answer is YES.
C.f. Apple's website page on Swift: https://developer.apple.com/swift/
Are there blocks in Swift like in objective-c? How are they created and called?
How would do an asynchronous request in Swift?
Is it easy to create block related memory leaks in swift? If yes, how would you avoid them?
The Swift equivalent of an (Objective-)C block is called a closure. There's a whole chapter about them in The Swift Programming Language book.
Depending on the context where you use a closure, you can declare/use it with very concise syntax. For example, a method that takes a completion handler whose signature is (success: Bool, error: NSError) - > Void can be called like this:
someMethod(otherParameters: otherValues, completionHandler:{ success, error in
if !success { NSLog("I am a leaf on the wind: %#", error) }
})
There's also a trailing closure syntax that reads nicely in cases where a closure essentially provides flow control. And you can drop the parameter names when you want to be really brief (at some cost to readability, but that's okay in some obvious cases like the below). Often a return statement is implicit, too.
myArray.sort { $0 < $1 }
let squares = myArray.map { value in
value * 2
}
Swift itself doesn't have anything for asynchronous requests, so you use existing API for that. You can use the trailing closure syntax, though:
dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0)) {
// do some async stuff
NSOperationQueue.mainQueue().addOperationWithBlock {
// do some main thread stuff stuff
}
}
In most cases, you don't need to worry about creating reference cycles with Swift closures the way you do with ObjC blocks. To put it simply, the capture semantics are similar enough to "just work" the way you want it to for most stuff, but different enough that the common patterns for block/closure uses (e.g. dispatch to background/main thread and referencing self's properties) don't cause cycles.
Cycles are still possible, though, and there is a solution for them. This answer's a bit long already, so check out Strong Reference Cycles for Closures in the docs for the complete explanation.
Blocks in Swift are called closures. They operate much the same as blocks (though are more flexible, and operate in more places). Reference cycles are possible with closures in Swift, and can be avoided with closure capture lists.
“Swift provides an elegant solution to this problem, known as a closure capture list. However, before you learn how to break a strong reference cycle with a closure capture list, it is useful to understand how such a cycle can be caused”
Excerpt From: Apple Inc. “The Swift Programming Language.” iBooks. https://itun.es/us/jEUH0.l
As I said in another question, you have many ways offered to pass a block equivalent to function in Swift.
I found three.
To understand this I suggest you to test in playground this little piece of code.
func test(function:String -> String) -> String
{
return function("test")
}
func funcStyle(s:String) -> String
{
return "FUNC__" + s + "__FUNC"
}
let resultFunc = test(funcStyle)
let blockStyle:(String) -> String = {s in return "BLOCK__" + s + "__BLOCK"}
let resultBlock = test(blockStyle)
let resultAnon = test({(s:String) -> String in return "ANON_" + s + "__ANON" })
println(resultFunc)
println(resultBlock)
println(resultAnon)
Update: There are 2 special cases to the Anonymous function.
The first is that function signature can be inferred so you don't have to rewrite it.
let resultShortAnon = test({return "ANON_" + $0 + "__ANON" })
The second special case works only if the block is the last argument, it's called trailing closure
Here is an example (merged with inferred signature to show Swift power)
let resultTrailingClosure = test { return "TRAILCLOS_" + $0 + "__TRAILCLOS" }
Finally, as an example:
Using all this power what I'd do is mixing trailing closure and type inference (with naming for readability)
PFFacebookUtils.logInWithPermissions(permissions) {
user, error in
if (!user) {
println("Uh oh. The user cancelled the Facebook login.")
} else if (user.isNew) {
println("User signed up and logged in through Facebook!")
} else {
println("User logged in through Facebook!")
}
}