I have this code and I'm questioning if I need to use a capture list to make my reference to self weak.
Right now I'm thinking that getTextFileData and .main.async are static methods, therefore, this will not cause a retain cycle. However, I do access the games property, kinda not sure.
NPWebService.getTextFileData { (games, success) in
if success {
self.games = games
DispatchQueue.main.async {
self.updateUI()
}
}
}
This wouldn’t cause a retain cycle because it looks like a static method on a different type. So the getTextFileData method will temporarily retain the closure you are passing in until any asynchronous work is complete, and the closure will temporarily retain self. But when the work is complete and the closure is done executing, those temporary retains will expire and automatic memory management can clean up as appropriate.
The danger of a retain cycle is when you have a closure the that references / captures self, and self also retains the closure, for example. Like this:
class GameController {
var games: [Game]?
// self retains the closure as a property
let updateClosure:([Game], Bool)->() = {
games, success in
if success {
self.games = games // and the closure captures self. Each retains each other indefinitely, this is a retain cycle and neither this closure nor self will ever be deallocated
}
}
func load() {
NPWebService.getTextFileData(updateClosure)
}
}
So normally weak self or unowned self in closure capture lists are only needed:
if self or something retained by self will be retaining that closure (this is rarely the case when a closure is created locally and transiently at the call site to pass into a method as an argument)
if for non-retain cycle reasons, the desired behavior is to allow self to be deallocated before some asynchronous work is completed, and before the closure is invoked. In this case the closure shouldn’t strongly capture self and instead should weakly capture it and check if it still exists before making use of self when the closure is invoked
Related
Coming from a different language I got surprised by a silent bug where an object which is passed to a method as a callback suddenly is never called. The reference to a callback is somehow lost.
Minimal (not runnable) example of a problem:
class Foo: NSObject, AVCaptureFileOutputRecordingDelegate {
func bar() {
let out = AVCaptureMovieFileOutput()
let delegate = Foo() //nonsensical in this case, in normal case diff. object will be used
out.startRecording(to: /*...*/, recordingDelegate: delegate)
//Result: delegate methods are never called
}
}
Minimal (not runnable) example of a "solution":
class Foo: NSObject, AVCaptureFileOutputRecordingDelegate {
func bar() {
let out = AVCaptureMovieFileOutput()
out.startRecording(to: /*...*/, recordingDelegate: self)
//Result: delegate methods are called when approperiate
}
}
I'm puzzled...
Why does that happen?
How to prevent such situation?
Is such silent failure by design?
This question stems from the AVCaptureMovieFileOutput never calls delegate on screen recording
Most delgates are weak so that they do not create a retain cycle, see Automatic Reference Counting (ARC). Most of this is assuming that the delegate storage for what you are using is weak.
In your first example the only strong reference to the object is held by the function bar, as the delegate is a weak reference. Once the function ends, the only remaining strong reference is gone and the object is free to be deleted.
class Foo: NSObject, AVCaptureFileOutputRecordingDelegate {
func bar() {
let out = AVCaptureMovieFileOutput()
let delegate = Foo() //object created, and strong reference stored in variable called delegate
out.startRecording(to: /*...*/, recordingDelegate: delegate) // object passed in and likely stored in a weak variable inside of the `out` object. This means it will not keep a strong reference to your Foo object.
//Result: delegate methods are never called
}// local variable called delegate goes out of scope, strong reference to your Foo object goes away, there are no more strong references, can be deleted.
}
In the second example when using self as the delegate, self is likely sticking around after the end of the bar function, so the delegate remains.
class Foo: NSObject, AVCaptureFileOutputRecordingDelegate {
func bar() {
let out = AVCaptureMovieFileOutput()
out.startRecording(to: /*...*/, recordingDelegate: self) // pass `self`, which presumably has something else referencing it with a strong reference, so it stays alive
//Result: delegate methods are called when approperiate
} // `self` still has strong references to it (somewhere else) keeping it alive after the function call, so the weak reference that is being used to call the delegate methods can still operate! Yay!
}
Hopefully that answers the "why."
As for prevention, you need to make sure you keep a strong reference to any delegates (or weak variables) that you want to stay alive.
This behavior is by design as it is used to prevent retain cycles and memory leaks. When designing your own classes with delegates it is up to you to use weak appropriately to prevent retain cycles when necessary.
As for the silence of the failure, there are many cases where the delegates are optional, and it is not considered a failure for the delegate to be nil and for the delegate functions to not get called. Many times the functions are called like delegate?.delegateMethod() intentionally so that the function will be called if you want to have a delegate and it won't cause a problem if you don't want to have a delegate.
There are a lot of tutorials on ARC. But I am not understanding the clear working of unowned or weak as how the reference captured variables becomes null.
Apple Document :
Define a capture in a closure as an unowned reference when the closure
and the instance it captures will always refer to each other, and will
always be deallocated at the same time.
class RetainCycle {
var closure: (() -> Void)!
var string = "Hello"
init() {
closure = { [unowned self] in
self.string = "Hello, World!"
}
}
}
the closure refers to self within its body (as a way to reference self.string), the closure captures self, which means that it holds a strong reference back to the RetainCycle instance. A strong reference cycle is created between the two. By unowned its breaking reference cycle.
But I want to understand which scenario both will not be mutually deallocated at the same time and Unowned self becomes null just want to crash it.?
As I get, You ask How self can be null while closue is running. If I get this right I can give you a quite similar example this that I have seen before.
I wrote an extension to UIImageView that download image from given link and set itself like this.
public extension UIImageView{
func downloadImage(link: String){
let url = URL(string:link)
URLSession.shared.dataTask(with: url){ [unowned self]
if let image = UIImage(data: data){
DispatchQueue.main.async{
self.image = image
}
}
}
task.start()
}
}
But there was a problem. Downloading an image is a background task. I set completion method to UrlSession and increased its reference count. So, my closure remains even if imageView is deaollecated.
So What happens if I close my viewController that holds my UIImageView, before download completed. It crashes because of imageView is deallocated but closure still remains and tries to reach its image property. As I get, you want to learn this.
I changed unowned reference to weak to solve this problem.
which means that it holds a strong reference back to the RetainCycle instance
This isn't true. It has an unowned reference back to the RetainCycle instance. That's not the same thing as a strong reference.
But I want to understand which scenario both will not be mutually deallocated at the same time and Unowned self becomes nilI just want to crash it.?
Any time closure is captured by something outside of RetainCycle, and so outlives its owner:
var rc: RetainCycle? = RetainCycle() // create an RC
let cl = rc?.closure // Hold onto its closure
rc = nil // Deallocate the RC
cl?() // Access the closure safely, but its reference to `self` is invalid. Crash.
As a rule, closures that involve unowned self should be impossible to reference outside of self. It's sometimes difficult to know this is true. For example, here is a case that recently crashed an app I work on:
var completion: (() -> Void)?
...
DispatchQueue.main.async { [unowned self] in
self.completion()
self.completion = nil
}
This feels fine, but if self is deallocated between the time it enqueues the main-queue block and the time that block runs, boom.
BTW, in this case the right answer would be a regular, strong self. We want the retain loop to keep this object around until its completion handler runs, at which point the block goes away, the reference to self goes away, and self is properly deallocated. So [weak self] is not always the answer, either.
I just want to know if I am understanding this correctly or not. So according to the apple docs when you create a closure as a property of a class instance and that closure references self(The class that created the closure property) this would cause a strong retain cycle an ultimately the class nor the closure will ever get released. So in laymen terms that means that if I have a class that has a property and that property is a closure, and once I assign the functionality of that closure within the class that declares the closure property that will cause a strong retain cycle. Heres a quick example of what I mean
class SomeViewController{
let myClosure:()->Void
public func someFunction(){
....bunch of code
myClosure = {
self.dismiss(blahBlahBlah)
}
}
}
This ultimately causes a retain cycle since the closure keeps a strong reference to self, which is the class that creates the closure property. Now to fix this according to apple I would define a capture list like so
class SomeViewController{
let myClosure:()->Void
public func someFunction(){
....bunch of code
myClosure = { [weak self] in
self?.dismiss(blahBlahBlah)
}
}
}
Notice how I put the [weak self] before the in statement. This lets the closure know to only hold a weak reference to self and not a strong reference. IM supposed to use weak when self can out live the closure or unowned when the closure and self live for the same duration.
I got this information from here Automatic Reference Counting and in the Strong Reference Cycles for Closures section of that link, theres this sentence "A strong reference cycle can also occur if you assign a closure to a property of a class instance, and the body of that closure captures the instance" Im about 90% sure Im understanding this correctly but theres just that 10% of doubt. So do I have this correct?
The reason why im asking this is because I use callbacks for some of my buttons in my views. And those callbacks call to self but the self in that scenario is the view controller that is responding to the callback not the actual view itself. This is where im doubting myself because I from that sentence I highlighted I don't think I need to put [weak self] on all those button callbacks, but im just making sure. Heres an example of this
class SomeViewController {
let someSubview:UIView
override viewDidLoad() {
//Some Subview has a button and in that view I just have some action that gets fired off calling the callback here in the view controller I don't need to use the [weak self] in this scenario because closure property is not in this class correct?
someSubview.someButtonsCallback = {
....run code then
self?.dismiss(blahBlahBlah)
}
}
Yes, that can still cause a retain cycle.
The simplest retain cycle is 2 objects that each have strong references to each other, but 3-way and larger retain cycles are also possible.
In your case, you have view controller who's view contains a button (a strong reference.) The button has a strong reference to a closure. The closure strongly references the view controller using self. So the view owns the button. The button owns the closure. The closure owns the view controller. If you dismiss the view controller (say it was a modal) then it SHOULD be deallocated. However, since you have this 3-way retain cycle, it won't be deallocated. You should add a deinit method to your view controller with a print statement and try it.
The solution is to add a capture list (The [weak self] bit) just like you did in your first example.
Note that a common pattern is to add a capture list, and then map the weak variable to a strong variable inside the closure:
let myClosure = { [weak self] in
guard let strongSelf = self else { return }
//...
strongSelf.doSomething()
}
That way, if the closure is still active but the object that owns it was released, the guard statement at the beginning detects that self is nil and exits at the beginning of the closure. Otherwise you have to unwrap the optional every time you refer to it.
In certain cases it's also possible that the object in the capture list (self in these examples) could be deallocated in the middle of the closure being executed, which can cause unpredictable behavior. (Gory details: This is only when the closure is run on a different thread than the owner of the object in the capture list, but completion handlers are pretty commonly run on a background thread, so it does happen)
Imagine this:
let myClosure = { [weak self] in
self?.step1() //1
//time-consuming code
self?.property = newValue //2
//more time-consuming code
self?.doSomething() //3
//even more time-consuming code
self?.doSomethingElse() //4
}
With the code above, if the closure is run on a background thread, its possible that self would still be valid at step 1, but by the time you execute step 2, self has been deallocated. The same goes for steps 3 and 4. By adding the guard strongSelf = self else { return } at the beginning of the closure you test at the entry of the closure to make sure self is still valid, and if it is, you make the closure create a strong reference that only lives as long as the closure takes to run, and that prevents self from being deallocated while the closure code is executing.)
If I weakly reference self in a closure:
{[weak self] in self!.doSomethinmg()}
And self is deallocated, yet the closure continues to exist. Is it possible that self in the closure can become non-nil at some point in the future -- pointing to some random new object as determined by the runtime environment?
A pointer is a variable whose value is a memory address. The same is
true for a weak pointer, but the value of a weak pointer is
(due to some runtime magic) set to nil if the pointed-to object is
deallocated. It is irrelevant if any new object is created at the
same memory location later, the weak pointer variable stays nil.
So
{ [weak self] in /* ... */ }
creates a closure with a weak pointer variable. That variable
is set to nil when the pointed-to object is deallocated.
That may happen before the closure is called or during the execution
of the closure, so even
{ [weak self] in
if (self != nil) self!.doSomethinmg()
}
is unsafe and can crash if the object is deallocated between the test
and the method call.
But at any point, the weak pointer either points to
the original object instance or is nil. It will
never point to some other object instance. (That may happen with
unowned references.)
As Lorenzo says, don't
force unwrap the weak pointer. Either use optional chaining or
create a strong reference inside the closure.
It depends on what you want to achieve.
Martin is suggesting the right thing. In order to guarantee you don't have any crash and continue to deal with that instance you can use weak/strong dance. Otherwise, use self?.
So for example, with the weak/strong dance you can do the following:
{ [weak self] in
if let strongSelf = self {
strongSelf.doSomething()
}
}
Otherwise
{ [weak self] in
self?.doSomething()
}
While in the first case the instance that belongs to self will continue to exist until the closure will be discarded, in the second case, instead, since the instance will be put to nil, it will behave as a sort of non-op.
If all the strong references to an instance are removed, then the instance will eventually be deallocated. But that doesn't necessarily mean the references that were referring to it will be reused. All the strong references are obviously gone because they had to be in order for it to be deallocated in the first place. So whatever that means for the memory that those references required for their referring duties is irrelevant to the application, because the references don't exist anywhere in it anymore.
Which leaves unowned and weak references left. Apple makes it pretty clear that attempting to access the value of an unowned reference after the instance has been deallocated is a no, no:
If you try to access the value of an unowned reference after that instance has been deallocated, you’ll get a runtime error.
So that is also irrelevant to the application, simply because it's a hard-rule not to be broken.
Which leaves, last but not least, weak references. Apple says this about weak references:
Because a weak reference does not keep a strong hold on the instance
it refers to, it’s possible for that instance to be deallocated while
the weak reference is still referring to it. Therefore, ARC
automatically sets a weak reference to nil when the instance that it
refers to is deallocated.
...
You can check for the existence of a value in the weak reference, just
like any other optional value, and you will never end up with a
reference to an invalid instance that no longer exists.
The key take away here is that the reference is set to nil. And even though nil represents nothing, it's still a valid value. And I feel it may be safe to assume that the runtime environment won't steal a reference and use it for another purpose when it's still pointing to a valid value. Imagine what it'd be like otherwise.
So the relevance to the application here for weak references is just that they may become nil at some point; and as #MartinR showed in his answer, this may even happen mid-execution of the closure. So the final solution would seem to be creating a strong reference from the weak reference before using it, aka the "weak/strong dance":
Either:
{ [weak self] in
if let strongSelf = self {
strongSelf.doSomething()
}
}
or
{ [weak self] in
guard let strongSelf = self else { return }
strongSelf.doSomething()
}
If the closure consists of a single optional chain:
{ [weak self] in
self?.doSomething()
}
will simply result in a non-op if self is nil, but no idea if this bears any guarantee that self won't be deallocated mid-execution of the closure if it's arbitrarily longer:
{ [weak self] in
self?.doSomething()
...
self?.doSomethingDownTheLine()
}
So in this case, doing the weak/strong dance will guarantee that your closure will be all-or-nothing:
{ [weak self] in
guard let strongSelf = self else { return }
strongSelf.doSomething()
...
strongSelf.doSomethingDownTheLine()
}
Having a runtime error when program is compiled, purposefully the instantiation of SomeObject is short lived but the block getting passed captures the reference, doesn't work in playground but shows the error when the program is compiled and run.
Objective is to hold the reference temporarily of short lived objects SomeObject until callback gets completed.
Edit - if I comment [unowned self] inside go it works, as I believe it creates a strong reference, but hopefully no memory leak there??? (the caller object went out of scope anyway). please confirm that I shouldn't use [unowned self] in here.
import Foundation
class SomeObject {
func go() { //i guess problem is here, it can't find self,
anotherObject.asyncCall({ [unowned self] in //works if i comment this
self.complete()
})
}
func complete() { //can't move this routine inside block, part of parent class api
println("received callback after 5 sec")
}
}
class AnotherObject {
var callback: (() -> ())?
init() {}
func asyncCall(callback: () -> ()) {
self.callback = callback
let delay = 5 * Double(NSEC_PER_SEC)
let time = dispatch_time(DISPATCH_TIME_NOW, Int64(delay))
dispatch_after(time, dispatch_get_main_queue(), { [unowned self] in
self.callback!()
})
}
}
var anotherObject = AnotherObject() //not global object but permanent lived
for i in 1...3 { //can't change here, it's part of API that instantiates my objects
var instance = SomeObject() //short lived objects
instance.go()
}
please confirm that I shouldn't use [unowned self] in here
There's no need to concern yourself with memory management merely because an anonymous function mentions self.
If an anonymous function that mentions self is going to be a property of self, then you have a retain cycle and potential memory leak, and you should concern yourself with memory management. You can easily see whether you have a memory leak by implementing your deinit to log; if it doesn't log when you expect this object to be destroyed, it's leaking. If it is leaking, you might try using [weak self], not [unowned self]. unowned is more convenient but it is usable only in very limited circumstances.
However, I see no evidence that the object that's going to retain the callback is the same object that is referred to in it as self. It looks to me more like the opposite: you seem to be using [unowned self] on every anonymous function, apparently without the slightest knowledge of what you are doing. That is extremely dangerous. You should not interfere with memory management unless you have to and unless you know how to. My advice is that you start by deleting every single unowned in your code. Then implement deinit and see if you have any actual leaking objects. I'm betting that you don't.