UIWindow in Xcode 6 templates - ios

I have a simple question, why is the main (and only UIWindow) on all of the Apple templates in Xcode declared as a var, rather than a let.
var window: UIWindow?
As I understand it we should use let wherever possible, and especially where the instance won't change.
Also using let with objects still allows you to modify their properties (frame etc).
I'm having a little trouble understanding when to use let with object (class types).
I assumed something like CLLocationManager and UIWindow would be perfect examples of when to use let with objects, but Apple don't seem to use let in their code.

Properties defined with the let keyword must have a value by the time the object is initialized. The window isn't created during initialisation, so it can't be a let.
It's the same situation for view controllers - none of the outlets, or indeed the view properties, can be lets, because they don't exist at initialisation, but are created later when the view is required.
let properties don't really work well for UIKit and view controllers - by design, view controllers don't do a lot of work on initialisation, and you can only set up lets at that point. You are absolutely correct in that let properties would be great for things like a location manager or a managed object context, but as it turns out lazy var is often a better bet.
In my (limited, like everyone!) Swift experience, let properties are great for making immutable model classes, and local lets are the default for creating references in-code.
You can use let properties and define them right there, so for a location manager:
let locationManager = CLLocationManager()
Think more about what information you need to create whatever's going into the property. If it can be made from scratch without any context, do it like the example above. If it needs to have a delegate or any other properties passed in, you'll probably be creating or setting it after initialisation so a let isn't appropriate.

If you define window property as let, then UIKit framework will not be able to set window property of your app delegate when instantiating your default view controller from main storyboard.
So the answer is:
If you want to do everything in code manually, you can make window property to be defined with let. This way you must initialize it in init(...) method of your AppDelegate.
Otherwise, if you want to use storyboards and have them instantiated automatically with default view controller then you must define window as var and enjoy routines that apple perform for you.

It is ought to be initialized inside the application:didFinishLaunchingWithOptions: function. That way you can create different windows that suit your needs. For example:
The application may be being launched in the background to perform a task - no need to initialize a window here.
Application may be being launched by a notification - a different window object may be created.
Also, as pointed out by #Keenle, if you're using interface builder to shape your application, the window is created and assigned at runtime. In no way this could be done if the window property was constant.

Related

Why can't an #IBOutlet be assigned let instead of var in Swift?

I'm following a tutorial on Swift and I noticed that the author uses var instead of let when declaring an #IBOutlet variable. So I became curious as to why I can't use let instead since an object's properties are still mutable even if the object is constant or is this not the case?
The error Xcode shows when using let is
#IBOutlet attribute requires property to be mutable
but I'm confused because questionLabel is a UILabel object and not necessarily a property of an object. Or is the questionLabel object a property of the current viewController?
import UIKit
class ViewController: UIViewController {
#IBOutlet let questionLabel: UILabel!
}
Thank you in advance if I'm over analyzing.
The #IBOulet marked properties are generally properties of a ViewController that are connected using the interface builder. The view you create in the interface builder has to actually connect the interface elements in it to the properties during your actual application runtime.
For that reason it firstly creates a new ViewController using some init without connecting any interface elements. They only get connected at a later stage. For the runtime to be able to hook the properties up to the view elements after the object creation has completed they cannot be constants, they have to be mutable. Because they do not have a value after the initializer has finished they have to be optionals. And to not make using the properties cumbersome afterwards they are implicitly unwrapped optionals, so that you do not have to write label!.property but label.property suffices.
That is why your code crashes as soon as you try to do something with an IBOutlet variable which you failed to connect and that is also the reason why you cannot use / change / manipulate those fields in the initializer.
Regarding your actual var / let confusion. Yes, the object itself that is referenced using let can be changed, e.g. the text of a UILabel BUT the actual object reference cannot be changed. That would mean that if you do not give the constant a specific value in the initializer it would forever remain nil.
For the simple reason that it is not assigned during initialization (in the initXXX methods) but later, when the view is being loaded.
The compiler actually cannot be even sure that the variable is ever assigned because the view loading is comletely dynamic.
In swift, all vars and lets can be thought of as properties.
A property is immutable (a constant) if it's declared with let. It's mutable (a variable) if it's declared using the var keyword. That is the defining difference between let and var.
Outlets must be mutable because their value does not get set until after the object is initialized. (The view controller gets initialized and it's outlets don't get loaded right away.)
You are right questionLabel is an object of type UILabel but used as a property of your class ViewController. That's why you have #IBOutlet attribute requires property to be mutable. If you use var you are saying that a property is mutable. If you use let you are saying that the property is immutable.
Try to create questionLabel without #IBOutletand see what's going on. Probably, you can put let in front.
First the ViewController is created, then the view tree is built. That means that when the ViewController finished it's init these views don't exist yet. They will be added just before viewDidLoad by parsing the XML-like data of the storyboard or XIB.
I know that it's Apple's default way of doing things but I would always write my outlets like:
#IBOutlet let questionLabel: UILabel?
For the simple reason it's absolutely not proven this label will really exist at run time. For example when reusing ViewControllers over multiple screens, changing the layout after setting the connections and so on this outlet might not be set. If you would use the questionLabel defined as UILabel! and it would be nil your application will crash. I don't think applications in production should ever crash over something silly like that.
For real application safety the only way to know really for sure it exists is to build your UI's in code. But the ease of use of Storyboards for one off screens is really tempting, I still use them a lot.

Wrapping my mind around Objects in iOS

I'm having trouble understanding the idea of objects. From what I've read, they're instances of a class. When learning swift, they're quite easy to understand. Simply create a class and create an instance of it, and from there, you can modify it's properties and call its methods:
class ExampleClass {
let ExampleProperty = "rabbit"
}
let exampleInstance = ExampleClass()
But I don't see how that translates when using iOS, since I haven't seen any objects being created explicitly yet:
var example = Wss()
So my questions are:
Are things like buttons, labels, and sliders objects?
-If so, where's the "code" behind them? Why do buttons, labels, etc. display even before they're connected through outlets and actions to the View Controller? Is there a hidden "var thisButton = ThisViewController()" embedded into each of those sliders and buttons?
If my assumptions are wrong, can someone explain to me how objects work?
"Is there a hidden "var thisButton = ThisViewController()" embedded into each of those sliders and buttons?"
No, and this is exactly where interface builder excels. Much of Xcode's modern Interface Builder comes from NeXTSTEP. When you drag out a new UI component like NSButton and place it on your story board, Xcode is instantiating a new object of the NSButton class for you. When you save your file, Xcode serializes all the objects of your story board into a .nib file. At the time when this was invented, it was quite revolutionary, all made possible because of the dynamism of Objective C. It made GUI programming much simpler and dynamic. Every object in your story board is aware of its class. For example, when you instantiate a new NSButton, you can open the inspector and see for yourself that its class is NSButton. When you add custom views to your application, they keep track of their class in the same way. Whenever a nib file is loaded, these views are instantiated from their classes. You might have noticed that you never override the initializer of your views. Instead, you override methods like awakeFromNib. This is because there's a lot of behind the scenes work being done for you, from the time the object is first instantiated, to the time. During this time IBOutlets and IBActions are bound for you.
Competitors tried to make similar interface building applications, but they ultimately resorted to doing code generation behind the scenes. In these systems, when you saved your interface file, the program would generate a source file that contains code that instructs how to instantiate these objects anew whenever the interface is loaded. However, it proved significantly more complex a task then just serializing the objects, so these systems were error prone, and significantly harder to debug (because you'd be trying to debug machine generated source files).
Answering your questions:
Yes. Your objects are just being created from a NIB, or Storyboard. So the NIB, or Storyboard, will create those visual (UI) elements for you, which you can then be accessed via the IBOutlets
Your assumptions, are not completely wrong as in, there is in fact something allocating those objects for you. The NIB, or Storyboard, just describe a way for those objects to be created. Also some other customisations, like frames, colors, etc.
More about how this ties up can be found here.
Building on Alexander's answer:
UIView objects have a method init(frame:) that lets you create a new UIView object with a specified frame.
Other UIView subclasses might have init methods that take additional parameters.
UIView objects also support an init method init(coder:) that knows how to create an object from a stream of stored data. This is known as "deserializing" the object, or converting it from a byte-stream back into a running object.
When you build an object in a Storyboard or XIB file in Interface Builder, the system serializes the object into a byte stream and saves it into your Storyboard/XIB.
Then when you invoke the storyboard scene/XIB, the system reads the data stream and uses it to recreate (deserialize) the objects that are described in the storyboard/XIB.
The effect is essentially the same as if you wrote a bunch of code that created and configured all your views, but instead of writing all that code you are able to build your interface in Interface Builder, which is faster and easier to create, and MUCH faster and easier to update and maintain than a bunch of custom code.
But I don't see how that translates when using iOS, since I haven't seen any objects being created explicitly yet
There's no difference between the objects in iOS and what you understand objects to be. Objects are instances of a class. What you need to understand is that your own code is not the only place where objects can be created, and your own code will often interact with objects created outside your code. Here's a simple example:
let defaults = NSUserDefaults.standardUserDefaults()
Here defaults gets a reference to a user defaults object that the system provides. You never need to instantiate NSUserDefaults yourself.
Are things like buttons, labels, and sliders objects?
Yes, those are instances of UIButton, UILabel, and UISlider, respectively.
If so, where's the "code" behind them?
It's in the UIKit framework. You don't get to see the source code for those classes, but you can still use them by linking the framework into your app.
Why do buttons, labels, etc. display even before they're connected through outlets and actions to the View Controller?
You're talking about storyboards here. When you set up a view in Xcode's storyboard editor, the data that's stored in the storyboard file is essentially an archive containing serialized objects. When a view controller is instantiated from a storyboard, the objects in the storyboard are recreated from that data and then connected to the view controller's outlets. You can start this process yourself by instantiating a new view controller like this:
let storyboard = UIStoryboard(name: "MyStoryboard", bundle: nil)
let controller = storyboard.instantiateViewController(withIdentifier: "MyViewController")
You don't usually need to do that, though, because the segues in your storyboard provide for transitioning between scenes, including creating the view controller that's the destination of a given segue.

Calling functions from separate view controllers in swift

I think the solution to this is going to need to use delegation, but I'm unfamiliar with how to use them.
So in my project, I have my main viewcontroller/storyboard that contains a UIScrollView. That UIScrollview calls another storyboard (xib file) as a subview. The other storyboard (which is an xib file) is controlled with another swift file.
My question is, when I call an action inside of my other storyboard, how can I call a function from the main viewcontroller. Like say the viewdidload from the first viewcontroller.
I can't make the whole thing a global function, it needs to stay inside its class. So if I try to do ViewController.viewDidLoad() it needs (I think) an instance variable or something.
Thanks.
You can try:
Using weak variable (property) in the other class with type UIViewController
Assign the parent view controller to that property after the other view is initialized
Good reads about weak, strong, unowned references Here And Here
Firstly, if you want to call it with class name as you said above declare your method with "class". So its just like static in Java. It makes it generic to call it anywhere in your project. Make a separate extension.
class func myfunc(){
}
if you want to send data from B to A controller. You use what is called delegation. You give the work of B to A. Make a protocol above B for functions that you want to do or send with them. Call them in B. And then in A write code for those functions. So that you have the data from B to A
Else you demand something like common data. Create a singleton class and initialize properties methods there. You can use objects for that and call it in other controller to modify or make different instances.
You dont call viewDidLoad(). As the name says it loads once. If you want something that modify everytime you screen appears, use viewWillAppear

Swift 2 / XCode 7 - "EXC_BAD_ACCESS (code=2...)" when calling "ViewController().view" in another class

why is this error happening and what can i do to fix/prevent in the future? thanks!
NOTE: my other class is set up as such:
class Other {
//then all relevant funcs called
}
am i missing some basic setup information in order for this to run?
The formal explanation would be:
You're trying to access the view property before it was initialized. Another way to look at it is that you're trying to access the view property before it was loaded (in viewDidLoad).
Solution:
Depends what you're using that view for. I've never had to access another view controller's property like that. Consider exploring other strategies such as delegation, weak references to another controller, and passing variables in prepareForSegue if you need a reference from a view controller from another.

Massive Parent-Child and delegate pattern

I'm facing with a complex design problem. Due to a hard designed graphic I can't use Apple navigation pattern as UINavigationController or other ones.
This is the app diagram
Black arrow: protocols direction
Blue arrow: parent-child pattern direction
I created this flow because I want to maintain the code inside the single ViewController clear and isolated: I thought about them as modules that they can be used somewhere in other apps.
The RootViewController is the MainVCs position manager. It decides which is the current view-controller that it must be showed and it configures it based on its status (modified by delegate methods).
The single MainVC doesn't know that RootVC exists and it call Root using protocols. The single ChildViewController doesn't know that its MainVC exists and it call the Main using protocols and so on.
Code is clear and much easy to understand, my purpose, but when I have to apply a modify to the skeleton (RootVC) from the ChildViewControllerN, child of the umpteenth ChildViewController, child of the umpteenth MainViewController, I have to propagate the protocol until the RootViewController.
My first question is: is this pattern correct? What do you think about it?
My second question is: is there a limit point where I haven't to use delegate pattern but something like KVO?
ADDING
I read a university lecture about the composite pattern that it says:
A composite object knows its contained components, that is, its children. Should components maintain a reference to their parent component?
Depends on application, but having these references supports the Chain of Responsibility pattern
So, according to the Chain, I can maintain a reference of the parent but I have to declare a custom interface for this kind of reference. However, doing this I will decrease the number of protocols, but the second question still be unanswered
Opinion:
When I go beyond a single level of parent/child relationships, I tend to stop using delegates and move to NSNotification. Frequently, I go directly to NSNotification to reduce dependencies. I prefer that to KVO because it is explicit, whereas KVO is harder to debug as the project progresses.
(Example: what looks like a simple variable assignment on a background thread results in a hard-to-diagnose crash if a listener is deallocated on the main thread between the moment of assignment and the KVO delivery.)

Resources