I have an already setup view and want to wrap it in a subclass of UIView.
class ElementView: UIView {
var selected = false
}
The problem is that I cannot initialize ElementView with the already existing view.
Swift doesn't allow assigning to self too.
init(view: UIView) {
//here I would have to call super.init(coder: NSCoder?) or super.init(frame: CGRect)
//none of which actually properly initializes the object
}
Any ideas how to implement this?
Thanks :)
Clarification:
I will give you the larger context hoping it'd be more clear:
I am implementing a UIScrollView subclass. The scroll view contains an array of UIView objects which are added externally (from the user of the class).
In my custom UIScrollView class I want to implement a tap gesture recognizer for each object. That's already done:
let singleTap = UITapGestureRecognizer(target: self, action: Selector("handleTap:"))
singleTap.cancelsTouchesInView = false
addGestureRecognizer(singleTap)
and the handler:
func handleTap(recognizer: UITapGestureRecognizer) {
//some code to handle the tap
}
The problem arises when I want to handle the tap, I want to store the previous state of the view (was it tapped before or not) in order to toggle that state when the tap happens. I want to do different stuff to the UIView depending on its state.
recognizer.view returns the view to which the recognizer is attached to, which is what I need. But, this way I have no possibility of implementing a state for the UIView.
That's why I wanted to implement a custom wrapper for UIView which should contain the state information (which is also a problem). That's how I came up to asking this question...
In order to create a custom init for the UIView subclass you have to call the required init for the UIView superclass. Also while creating the custom init you have to send the Frame of the view to the superclass. Upon fulfilling these requirements you are free to pass on any arguments to the newly created init including the tap recognizer info.
Remember, if you are creating any variables - they have to be not nil upon the creation of the instance, thus in variable declaration you have to create some initial argument (for example - 0 for Int, etc.) for the initializer to work. Here is the example code:
var variableOne: Int = 0
var variableTwo: Int = 0
init(variableOne: Int, variableTwo: Int) {
self.variableOne = variableOne
self.variableTwo = variableTwo
super.init(frame: CGRectZero)
}
required init?(coder aDecoder: NSCoder) {
super.init(coder: aDecoder)
}
It sounds like you are trying to mimic a copy constructor, and it sounds like you are trying to build it in IB. The short answer is what you are doing doesn't make sense. It further sounds like that you wanted your code above to own or assume the identity of the argument view (your reference to not being able to assign to self). If this assumption is correct, your code would make even less sense - you would just need to assign to a variable.
Just create the view class, with the code that you have posted, you do not need to implement a constructor, since you have provided a default value for your selected variable. Your will with then be populated from IB via the coder based constructor.
If you are trying to clone or copy a given view, then refer to "Can UIView be copied"
Related
In a couple of my projects I think I'm not created a great structure in many cases.
It could be a game where I've created a game board (think about chess) with a grid of 8 * 8 cells. Each cell has a gesture recognizer that relies on a subclass (cell.swift), with the game logic in a parent ViewController.
For arguments sake, let us say we want to display to the user which square they have touched.
I've found out how to do this from the subclassed UIView (obvs. create the alert in the subclassed UIView / cell.swift in this example)
UIApplication.shared.keyWindow?.rootViewController?.present(alertController, animated: true, completion: nil)
but it seems to break the structure of the app - but wouldn't it be the same accessing an action in the parent ViewController? What is the best way of approaching this>
Your rootViewController is the VC on the bottom of your stack. It's not a safe way to access the visible VC, and is rarely useful, in general (there are cases, but I doubt your app would find them useful).
What you likely want to use is a delegate pattern. Let's say the parent VC that displays your chess board (let's call this MyBoardViewController), conforms to a protocol like the following. MyView is whatever custom UIView class you're using for the chess squares:
protocol SquareAlertHandler {
func handleSquarePressed(sender : myView)
}
And add the following property to your MyView class:
weak var delegate : SquareAlertHandler?
And replace whatever event handler you're currently using, with the following (I'm assuming you're using a UIButton in IB to handle the press, and have arbitrarily named the outlet 'didPress:'):
#IBAction didPress(sender : UIButton) {
delegate?.handleSquarePressed(self)
}
Now, add the protocol to your MyBoardViewController, and define the method:
class MyBoardViewController : UIViewController, SquareAlertHandler {
... ... ...
func handleSquarePressed(sender : myView) {
// Do something to handle the press, here, like alert the user
}
... ... ...
}
And finally, wherever you create the MyView instances, assign the MyBoardViewController instance as the delegate, and you're good to go.
Depending on your Swift literacy, this may be confusing. Adding code, so that I can at least match up the class names, would help to clarify things.
I have a
#IBDesignable
class Fancy:UIButton
I want to
addTarget(self, action:#selector(blah),
forControlEvents: UIControlEvents.TouchUpInside)
So where in UIButton should that be done?
Where is the best place for addTarget ?
1 - I have seen layoutSubviews suggested - is that right?
Note - experimentation shows that a problem with layoutSubviews is that, of course, it can be called often, whenever things move around. It would be a bad idea to "addTarget" more than once.
2 - didMoveToSuperview is another suggestion.
3 - Somewhere in (one of) the Inits?
Note - experimentation shows a fascinating problem if you do it inside Init. During Init, IBInspectable variables are not yet actually set! (So for example, I was branching depending on the "style" of control set by an IBInspectable; it plain doesn't work as #IBInspectable: won't work when running!)
4 - Somewhere else???
I tried to do it in Init, and it worked well. But it breaks designables from working in the Editor.
By thrashing around, I came up with this (for some reason both must be included?)
#IBDesignable
class DotButton:UIButton
{
#IBInspectable var mainColor ... etc.
required init?(coder decoder: NSCoder)
{
super.init(coder: decoder)
addTarget(self, action:#selector(blah),
forControlEvents: UIControlEvents.TouchUpInside)
}
override init(frame:CGRect)
{
super.init(frame:frame)
}
I don't know why that works, and I don't understand why there would be two different init routines.
What's the correct way to include addTarget in a UIButton?
tl;dr
override func endTrackingWithTouch(touch: UITouch?, withEvent event: UIEvent?) {
super.endTrackingWithTouch(touch, withEvent: event)
if let touchNotNil = touch {
if self.pointInside(touchNotNil.locationInView(self), withEvent: event) {
print("it works")
}
}
}
Why not use addTarget
addTarget method is part of action-target interface which is considered 'public'. Anything with reference to your button can, say, remove all of its actions, effectively breaking it. It is preffered to use some of 'protected' means, for instance endTrackingWithTouch which is accessible only to be overriden, not called directly. This way it will not interfere with any external objects using action-target mechanism.
(I know there is no strict 'public' or 'protected' in ObjC/UIKit, but the concept remains)
Your way
If you want to do it exactly your way then your example is all good, just copy addTarget call to init(frame:CGRect).
Or you can put addTarget in awakeFromNib (don't forget super) instead of init?(coder decoder: NSCoder), but you will be forced to implement init with coder anyway, so...
layoutSubviews and didMoveToSuperView both are terrible ideas. Both may happen more than once resulting in blah target-action added again. Then blah will be called multiple times for a single click.
By the way
The Apple way
By the Cocoa MVC (which is enforced by UIKit classes implmentation) you should assign that action to the object controlling that button, animations or not. Most often that object will be Cocoa MVC 'Controller' - UIViewController.
If you create button programmatically UIViewController should assign target to itself in overridden loadView or viewDidLoad. When button is loaded from nib the preffered way is to assign target action in xib itself.
Old Good MVC
As mentioned here in real MVC views do not send actions to themselves. The closest thing to real MVC Controller in UIKit is UIGestureRecognizer.
Be warned that it's pretty difficult to pull of real MVC with UIKit class set.
You should not add as target the same object that produces the action.
The target and its callback should be another object, usually a view controller.
There are 2 inits methods because the button can be instantiated by calling init or by the process of deserializion (NSCoder) from a nib/xib. Since you probably added the button to a storyboard the init method called is init?(_: NSCoder).
[UPDATE]
I agree about what you say in the comment, but I think that the action-target pattern should be used for communicating with other objects, I'm using conditional, because as far as I know I never seen something like what you wrote in Apple code or some other library. If you want to intercept and make some actions inside the button you should probably override some of the methods exposed in UIControl.
About designable, you are, again, correct. init(frame) is called if you are creating a button programmatically, init(coder) if the button comes from a xib.
The method init(frame) is also called during designable process. At this point I think that the best option is to debug directly your view.
Place some breakpoints inside you UIButton subclass
Select the view in your storyboard
Go to the Editor -> Debug selected views
Now you should be able to understand where the problem is.
How about implementing awakeFromNib and doing it there?
https://developer.apple.com/library/mac/documentation/Cocoa/Reference/ApplicationKit/Protocols/NSNibAwaking_Protocol/#//apple_ref/occ/instm/NSObject/awakeFromNib
You can also conditionally run (or not run) code when it is being run in the context of Interface Builder:
#if !TARGET_INTERFACE_BUILDER
// this code will run in the app itself
#else
// this code will execute only in IB
#endif
(see http://nshipster.com/ibinspectable-ibdesignable/)
Your initialize method is not correct, this will work:
```swift
override init(frame: CGRect) {
super.init(frame: frame)
self.loadNib()
}
required init?(coder aDecoder: NSCoder) {
super.init(coder: aDecoder)
self.loadNib()
}
private func loadNib() {
let nibView = NSBundle(forClass: self.classForCoder).loadNibNamed("yourView", owner: self, options: nil).first as! UIView
nibView.frame = self.bounds
nibView.autoresizingMask = [.FlexibleWidth, .FlexibleHeight]
self.button.addTarget(self, action: #selector(action), forControlEvents: .TouchUpInside)
self.addSubview(nibView)
}
```
I want to subclass UIButton and add to it a property isActive, which is a Bool that changes the look of the button (this is not related to the default .enabled property).
However, I also want to set the button type when initialized. Specifically, I want to initialize it much like it is in UIButton, like:
let button = MyButton(type: .Custom)
However, because the .buttonType is a read-only, I cannot just override it like:
convenience init(type buttonType: UIButtonType) {
buttonType = buttonType
self.init()
}
This spits out an error: Cannot assign to value: 'buttonType' is a 'let' constant (I already implemented the required init(coder:)).
So how can I set the .buttonType when initialized?
NOTE: The custom property shall be also used in other functions to change the logic within it, so I picked up a property, instead of defining two functions to change the look.
You don't need use it, according to documentation:
This method is a convenience constructor for creating button objects with specific configurations. If you subclass UIButton, this method does not return an instance of your subclass. If you want to create an instance of a specific subclass, you must alloc/init the button directly.
When creating a custom button—that is a button with the type UIButton.ButtonType.custom—the frame of the button is set to (0, 0, 0, 0) initially. Before adding the button to your interface, you should update the frame to a more appropriate value.
means button you created already custom, just use init()
buttonType is read-only, so you have to set the type via UIButton's convenience initializer:
convenience init(type buttonType: UIButtonType) {
self.init(type: buttonType)
// assign your custom property
}
I have a UIView with a TableView and a Button (Big Button). The TableView has a custom Cell. In this cell there is an "Add" button. I want to animate the first button when the user makes click on the Add button.
This is my schema:
This is my code:
class ProductsViewController: UIViewController, UITableViewDelegate, UITableViewDataSource {
#IBOutlet var tableView: UITableView!
#IBOutlet var bigButton: UIButton! <- I WANT TO ANIMATE THAT BUTTON
}
ProductCell class
class ProductCell: UITableViewCell {
#IBAction func addProduct(sender: AnyObject) {
//I WANT TO ACCESS THE BIG BUTTON FROM HERE
}
}
Screen example of my app
I've tried to get the parent controller or the superview to get the IBOutlet but the app is crashing allways
Add block properties to your cells which lets them notify your view controller when they have been clicked. In your view controller block code, you can then access the big button.
See my answer to a similar question. Simply replace the switch example with your button. So replace UISwitch with UIButton.
How can I get index path of cell on switch change event in section based table view
So rather than have the cell try and talk to another cell/button, have the cell notify the controller which can then manage the big button changes.
Although I made a comment about using alternate methods you could also employ a strategy below based on updates to a property stored in the current view controller class. You could just as well use property observation on the ProductsViewController but I assume you'd like to keep OOP focused and reduce the size of your controller.
Subclass the ViewController
One could subclass an existing UIViewController and then create a property in the super class that deals with the value that was changed (row tapped). In that subclass you could then do some animation. Because you would be subclassing you continue to obtain all the benefits and methods defined in your existing controller. In your identity inspector point your Class to the new subclass and create any functional updates to your UI using animation.
class ProductsViewController:... {
var inheritedProperty:UIView = targetView {
willSet {newValue } // is the newValue
didSet {oldValue} //is the old value
}
}
class AnimatedProductsViewController:ProductsViewController {
override var inheritedProperty:UIView {
//do something interesting if the property of super class changed
willSet {newValue } // is the newValue
didSet {oldValue} //is the old value
//you might want to call this method like so
// didSet { animate(newValue) }
}
func animate (view: UIView){
//do animation routine using UIView animation, UIDynamics, etc.
}
}
Property Observation
Whenever the didSelectCell... method is called just set a value to the inheritedProperty. Then add the property observers (see sample code) and react when the property changes (maybe pass a reference to the view you want to animate).
For example: Within the property observer you can just take that view and pass it to your animator function (whatever is going to do the animation). There are many examples on SO of how to animate a view so just search for (UIView animation, UIDynamics, etc).
The normal benefits of separation are encapsulation of functionality and reuse but Swift also guarantees that each set of property observers will fire independently. You'd have to give some more thought to this as to its applicability in this use case.
Do all this things in your viewController
Add target Method to cell's add button in cellForRowAtIndexPath Method
Like This
cell.add.addTarget(self, action: "addProduct:", forControlEvents: UIControlEvents.TouchUpInside)
Define method
func addProduct(button:UIButton)
{
// do button animation here
}
I have a custom UIView (called GridView) that I initialize and then add to a ViewController (DetailViewController). GridView contains several UIButtons and I would like to know in DetailViewController when those buttons are touched. I'm new to Swift and am wondering what is the best pattern to use to get those events?
If you want to do this with notifications, use 1:
func postNotificationName(_ notificationName: String,
object notificationSender: AnyObject?)
in the method that is triggered by your button. Then, in your DetailViewController, add a listener when it is initialized with 2:
func addObserver(_ notificationObserver: AnyObject,
selector notificationSelector: Selector,
name notificationName: String?,
object notificationSender: AnyObject?)
Both functions can be called from NSNotificationCenter.defaultCenter().
Another method would be to add callbacks which you connect once you initialize the GridView in your DetailViewController. A callback is essentially a closure:
var callback : (() -> Void)?
which you can instantiate when needed, e.g.
// In DetailViewController initialization
gridView = GridView()
gridView.callback = { self.doSomething() }
In GridView you can trigger the callback like this:
func onButton()
{
callback?()
}
The callback will only execute, if unwrapping succeeds. Please ensure, that you have read Automatic Reference Counting, because these constructs may lead to strong reference cycles.
What's the difference? You can connect the callback only once (at least with the method I've showed here), but when it triggers, the receiver immediately executes its code. For notifications, you can have multiple receivers but there is some delay in event delivery.
Lets assume your GridView implementation is like as follows:
class GridView : UIView {
// Initializing buttons
let button1:UIButton = UIButton(...)
let button2:UIButton = UIButton(...)
// ...
// Adding buttons to view
self.addSubview(button1)
self.addSubview(button2)
// ...
}
Now, we will add selector methods which will be called when a button is touched. Lets assume implementation of your view controller is like as follows:
class DetailViewController : UIViewController {
let myView:GridView = GridView(...)
myView.button1.addTarget(self, action: "actionForButton1:", forControlEvents: UIControlEvents.TouchUpInside)
myView.button2.addTarget(self, action: "actionForButton2:", forControlEvents: UIControlEvents.TouchUpInside)
// ...
func actionForButton1(sender: UIButton!) {
// Your actions when button 1 is pressed
}
// ... Selectors for other buttons
}
I have to say that my example approach is not a good example for encapsulation principles of Object-Oriented Programming, but I have written like this because you are new to Swift and this code is easy to understand. If you want to prevent duplicate codes such as writing different selectors for each button and if you want to set properties of your view as private to prevent access from "outside" like I just did in DetailViewController, there are much much better solutions. I hope it just helps you!
I think you better create a class called GridView that is inherited from the UIView. Then, you can connect all you UI element with you class as IBOutlet or whatever using tag something like that. Later on, you can ask the instance of GridView in DetailViewController so that you can connect as IBAction.
Encapsulation is one of the principles of OOP.