Section 1
Preview this deck
What does the iOS application lifecycle consist of?
Front
| 5 | 0 | ||
| 4 | 0 | ||
| 3 | 0 | ||
| 2 | 0 | ||
| 1 | 0 |
Active users
1
All-time users
1
Favorites
0
Last updated
6 years ago
Date created
Mar 1, 2020
Cards (126)
Section 1
(50 cards)
What does the iOS application lifecycle consist of?
The main point of entry to our apps is the UIApplicationDelegate, a protocol we conform to, to do thing. A number of methods are called... applicationWillLaunch, applicationDidLaunch, applicationWillEnterForeground, applicationDidEnterForeground, becomeActive Will launch is called first and provides storyboards, then didlaunch allows for updates after state restoration. Foreground methods are called next, and also after interruptions to the app such as phone calls. ResignActive is called when home button pressed or a phone call happens for example, DidEnterBackground is called when app is backgrounded, it gives a few seconds to do any clean up WillTerminate is called when app will be purged from memory. Launching options can tell you about push notifications or urls that directed to the app.
What is UserDefaults and when is it used?
UserDefaults is a simple key value store that is persistent. It is not secure and not persisted through app installs. Generally it is used to store simple user preferences and configs.
How do you download a file in swift?
Typically use the download task function on URLSession. This will download a file and store in a temporary file in your app. The completion handler will give you a url to the temp file. You must open the file or move it to permanent disk storage to keep it. Otherwise completion handler will get rid of it.
What options do you have with animation on iOS?
UIKit - gives some basic animations on UIView, using .animateWithDuration methods. You can animate frame, alpha, background color changes for example. CoreAnimation - manipulate the layers of view using things like CABasicAnimation to animate border transitions for example. UIKit Dynamics - used for more complex animations that a user could interact with potentially
How do you serialise and map JSON from the backend?
Apple introduced Codable protocols that allow us to fairly easily encode and decode to and from JSON. If we create a Struct that conforms to Decodable and who's parameters match the JSON's types and variables names, we can use a JSONDecoder to decode directly to that struct. We can also provide explicit codingkeys to change the JSON variable names to names we may prefer. We can also use the init with decoder to decode JSON into different or custom types. Alternatively, what we often see is a ResponseModel that matches the JSON and a domain model object that the responsemodel is mapped to. This gives clarity between the actual response and the mapped data we require.
What is KVO - Key Value Observation?
KVO provides mechanics for observing property changes. It allows many observers to subscribe to one emitter. KVO in Swift requires dynamic @objc on variables and @objc on class to work as it uses Objective-C under the hood. The notification center uses a similar architecture to KVO to allow observers to subscribe to `post` notifications. You should remember to remove subscribers when they are no longer required. KVO is used in Reactive frameworks such as RxSwift.
How do you typically implement networking in iOS?
If you think of an iOS architecture as a layered system, then networking sits on the Service Layer, so it should be seperate from the UI, business logic or storage layers. A basic implementation for networking would be to have a NetworkManager wrapping a URLSession or an open source network manager. An API Client who uses the NetworkManager to send requests to an API and handles things like user authentication and signing requests. And then individual services responsiblef for defining what Endpoints to hit and what models to return. A UserService or a PostService for example. This gives a good level of seperation of concerns such that if a change occured in the APIClient are it would not impact the other areas.
How do you save data to disk on iOS?
Typically you use the FileManager, which can create subdirectories and store large files to disk. You can store any object so lonage as it is serialiable.
What should go into the networking / service layer?
Service Layer is between your app and any external data that you require. So things in the service layer may be responsible for communicating with network, over bluetooth, GPS, gyroscope etc. For any given service you'd ideally create a layered interaction where the low-level implementation details are hidden behind high-level interfaces. For example, a network manager may use URL Session, an api client may use that network manager and post service may use the api client. Or a BLEClient may manager connection to bluetooth devices, a PeripheralClient may manage discover, then a DeviceService may manage discovery, connection and communication to a specific device, and finally a certain data service will use that DeviceService and map it to models used in the app. These designs seperate concerns and provide layers of indirection.
What are xibs and nibs?
Xibs are interface builder files that define UI, when compiled they become nibs. For the most part they've been replaced by storyboards.
Can a static library (.a) contain resources like images, sound files etc?
No, .a files just contain compiled code from the archive utility. Instead you can create a .framework bundle that can also contain resources.
Have you used sockets and/or pubsub systems?
Sockets and pub/sub systems are used for near real-time connection. By opening a connection to a socket or subscribing to a publisher you get the most recent changes emitted from the publisher and any new published data. There are libraries out there that do the heavy lifting, such as Starscream With starcream you open the connection to a web socket, then use delegate methods to react to the output and connection states
How can you unwrap an optional in swift?
Force unwrap using the `bang` operator. This will crash if the optional is nil. You can use this for implicit unwrapping too, so you can tell the compiler I know this will exist, then refer to it as a non-optional later, as with IBOutlets to labels for example. Optional binding / unwrapping using if let or guard let to check if the value exists and do something else if it does not. Optional chaining, it certain cases you can refer to methods or properties in an optional and if it is nil it will act as nil.
What is VIPER?
VIPER is an architecture for building apps based on Clean Architecture. It focusing on seperating responsibilities. It is an alternative MVC and MVVM. In VIPER you have a View, Interactor, Presenter, Entity and Router. Each feature tends to be built around a Use Case and each Use Case fits into these building blocks. The basic flow of VIPER is a user is brought to a View from a Router. The View displays and receives User Input. The Presenter translates this User Input and tells the Interactor. The Interactor tends to contain the business logic and the Entities. It tells manages updates to Entities and tells the Presenter what to do next. Who in turns tells the View or the Router. Benefits: easier to test, Clean, SRP, good for big teams. Negatives: lots of boilerplate (code generation used), higher learning curve, need full buy-in and to stick tightly to it.
What is the storage layer in an iOS application?
The storage layer contains the classes responsible for saving, loading and persisting data in your app. This may be in-memory classes, UserDefaults, FileManager stores or databases like CoreData or Realm. A typical set of classes may be wrappers for Keychain, UserDefaults, FileManager as well as a Repository that reads and writes to disk. There are also NSManagedObjects that persist domain models to CoreData.
What are Singletons? What are they used for?
Singleton is a class that returns only one instance of itself, no matter how many times requested. Example - UIApplication.shared. Singletons are often thought as an anti-pattern. They store global state that may be changed somewhere in the app unexpectedly resulting in unexpected changes elsewhere. They are difficult to test as you cannot inject an alternative singleton unless you wrap your singleton in a protocol. Sometimes they can be useful, such as UIApplication.shared or Configuration.shared, when it is not likely to change.
How do you work with UICollectionView?
UICollectionView works very similarly to UITableView. It also requires a UICollectionViewLayout - and this defines the layout such as horizontal or vertical scrolling, differing cells sizes and heights. etc.
What are extensions and categories in Swift and Objective-C?
They are both ways of extending functionality. In Swift, extensions allow you to declare additional properties (not stored) and methods to a class or struct. They can be used to add conformance to protocols. In Objective-C categories are typically used to extend functionality of 3rd Party or Apple frameworks. A drawback is that extensions are globally applied by default so a declaration in one place is available everywhere, unless made private or fileprivate.
What are the advantages and disadvantages of creating AutoLayouts in code versus using storyboards?
Advantages of Storyboards: Quicker to build initial UI, can be easy to visualise, Apple provides many examples Disadvantages of Storyboards: Harder to debug, merge conflicts are more difficult to handle Advantages of Code: Explicit, easier to debug, possible merge conflicts are more clear Disdvantages of Code: Must compile app to see changes, constraint api can be challenging, SnapKit can help.
Compare GCD and OperationQueues?
GCD is an Apple API that allows you to dispatch blocks of code to be executed on different disptach queues, which can be serial queues or concurrent queues, and can be synchronous or asynchronous. GCD then handles thread management for you, you just know you're block will be executed at some point. OperationQueues are similar in that you can specificy Operations to happen on given queues. OperationQueues and Operations give you more control over your processing. You can set the priority of operations, check the status of existing operations, cancel operations and specify operation dependencies.
How is memory management handled in iOS?
Using Automatic Reference Counting. When an instance of a class references another instance, it increases that instances reference count by 1. As long as an instances reference count is more than 0 it remains in memory. When it becomes 0 it is deallocated memory. By default references between classes a strong. Value types (structs) do not increase reference counts as they are copied rather than referenced. Memory leaks can occur due to Strong Reference Cycles. For example, a UIViewController references a ViewModel. The VC passes a closure to the ViewModel referencing itself. Now we have a strong reference cyle that will prevent deallocation. Appending weak or unowned fixes this. Weak creates a weak reference that does not increase reference count and can be nil. Unowned explicitly states self will exist when this closure calls. Weak is used more often but unowned may be used for example, when you know a UIButton tap's closure will call the parent UIViewController.
What are the concerns and limitations of networking on iOS?
Battery and Bandwidth. Try issue as few requests as possible and retry if network suddenly drops. Large downloads and uploads should warn user. iOS 13 allows for scheduling background network tasks for times of high battery and good connection. We can also ensure wi-fi is used over network for big uploads downloads.
What is URLSession and how does it work?
URLSession is Apples API for networking. You can create URLSessions and configure them or use the default URLSession.shared. (A browser might use multiple sessions for example) Configuration can include cookie and caching setting, and even mocking. Within a session you can create multiple dataTasks, downloadTasks or uploadTasks. You can pause, start and cancel them. You can use a delegate to get updates on your session, or you can use a completion to handle a completed task.
How do you handle multi-threading with networking iOS?
You should not block the main thread, which would block the UI for a user when doing a network request. By default, URLSession dispatches network requests to a background thread, and when your completion handler is called, it is called on a background thread unless you specify otherwise. We can handle the threading with DispatchQueues or NSOperations. Most typically, if a network call finishes on a background thread, we call DispatchQueue.main.async to get back to the main thread to do UI Updates. Not doing this can cause funky UI, runtime errors and crashes.
How do you mix AutoLayout with Frames?
Autolayout and frames can coexist so long as they aren't mixed directly. For example, you could have a superview use autolayout, then have it's subvies use frames. The subviews would override layoutSubviews() to calculate their frames correctly.
What is REST?
Rest stands for Representation State Transfer. It is an API architecture built on top of HTTP. It defines how APIs should be designed. It should align with the CRUD acronym. Ideally a RESTful API would mean a resource has one endpoint to which you can GET all resources, or GET a particular resource with an ID. You could POST to that endpoint or delete an ID. It should return standard HTTP status codes such as 200 for a good GET or 201 for a good POST. Rest responses tend to JSON Rest imposes statelessness. I.e. the server need know nothing about the state of the client and visa versa.
What is Serialization?
Serializing means converting in-memory objects into raw data (i.e. bits) to store on disk. Deserializing is the opposite.
What is MVC?
Model View Controller. It is an architecture pattern defining how the app works. Model represents the data, such as array of structs of Characters. View represents the view shown to the user, responsible for showing information and responding to taps. Controller sits between the two translating user interaction to model updates or view updates. MVC can lead to Massive View Controller syndrom when one ViewController owns all the views and the model. MVVM is a pattern designed to address this.
What are CGRect Frames? When and where would you use them?
Frames are rectangles you can create that define where UI elements will appear on the screen. They use the coordinate system of their superviews. You define the origin x and y point then speific height and width attributes for you UI elements. They are rarely used as auto layout and constraints work better for varying screen sizes. If they are used tend to define positions based on the frames of other objects.
What are properties and instance variables in Objective-C and Swift? What are type properties?
Properties in Objective-C are used to store data in instances of classes. They define memory management using strong, weak or assign. When you declare a property they store their values in instances that by convention have the property name preceded by underscore. A feature of Objective-C properties is that you can use KVO to functionally react to property changes. Swift properties are typically stored properties defined as constants or variables. They default to strong but can be weak or unowned. Swift properties must be defined at initialisation of class unless they are optional. Swift properties do not provide KVO but do provide Property Observers i.e. willSet and didSet. KVO can be achieved by using the @objc modifier. Class or type properties are defined in Swift using the static or class keyword. They are defined for the entire class or type rather than particular instances. Type properties are lazy by default and are guaranteed by initialised only once. Lazy can be used on stored properties to ensure they are initialised only when needed.
What is Codable?
Codable is a combination of Encodable and Decodable. It is a protocol. Conforming to the protocol means your object or data can be encoded or decoded to different types such as data or json.
What is the difference between `struct` and `class` in Swift? When would use either?
A struct is a value type. It is copied when assigned to a new instance. It does not increase reference counts. It cannot inherit. It's thread-safe. A class is a reference type. When a variable is assigned an instance it copies a reference to that instance. It must be made thread-safe. It can implement a deinit. A struct tends to be used for simple data types or wrappers of data. As that data can copied and passed around. In a multi-threaded environment such as when opening a database connection on a different thread, structs are safer. A class tend to be used when inheritance is required, when copying an instance doesn't make sense such as a UIWindow or UIViewController, or when the class is acting like a conduit such as a APIClient or ConfigurationManager. It makes sense to have single instances.
What is the difference between `let` and `var`? What differences does this have to Objective-C?
A let defines a constant. It must have a value when initialised. It cannot be changed. A var defines a variable. It can change and can be unassigned i.e. nil. The difference with Objective-C is that everything is dynamic and can be nil. Nil can also be handled by throwing exceptions, whereas Swift optionals must be unwrapped.
What are protocols in Objective-C and Swift?
Protocols define the interface that any conforming types must implement in order to conform to that protocol. In Swift protocols can define what instance methods, type methods, properties, operators and subscripts must be available. By default swift protocol parameters are required, you'd have to add @objc to allow optional protocol implementations. Protocols abstract away the implementation detail so that you focus in the interface rather than types or inheritance trees. Protocols allow you swap implementations at runtime. Protocols make testing easier as you can easily declare mocks and stubs conforming to protocols. Protocols are a core part of the delegation pattern used frequently in iOS.
How do you download images on iOS?
Use a dataTask from a URLSession to fetch the data from a URL. Convert the Data into a UIImage, then you're done. Things to consider are download size and time, large images may take longer. Caching, simple caching could use a dictionary or NSCache to store images in memory against URLs you requested them from. Could use a third-party framework such as Alamofire.
What is Keychain and when is it used?
UserDefaults is a simple key value store that is persistent. It is encrypted by the OS and is persisted between app installs. Generally it used to store sensitive things like passwords or tokens.
How do you do animation with Frames and AutoLayout?
Use UIView.animateWithDuration. In the animate block change the frame or constraints. If changing constraints, call layoutIfNeeded after on the views changed.
What is UIStackView? When would you use it and why?
UIStackView is used as a container and makes it easy to align and resize views dynamically. It effectively manages constraints given a grid layout. It is also handing because when removing and adding views at runtime, constraints are handled.
How would you cache images?
Three ways. 1. In-memory cache using something like a Dictionary or more optimall NSCache, storing an image against URL to retrieve it later if it exists. 2. In database like Core Data or Realm, though not really used for larger images 3. In file system using iOS built-in File-Manager to write to a file and retrieve later
What is NSCoding?
NSCoding is Objective-C style Codable. It is a protocol that your objects can conform to make them easily serializable to store in NSUserDefaults, NSFileManager or Keychain. Conforming to NSCoding means you must implement an init with decode and encode from method.
How would you approach major database/storage migration in your application?
You'd have to ensure whatever your current db solution is, it is not tightly coupled. That is there should be a layer of classes and protocols that other objects interact with to store and get data from the database. This layer will then be injected with Core Data, for example. So long as the classes follow SRP correctly you should be able to inject Realm in place of Core Data with relatively little effort. Then you just have to code a copy from one database to another before you run the migration.
What is AFNetworking/Alamofire? How do you use it?
AFTNetworking / Alamofire are third-party libraries that wrap URLSession and make it easier to work with. You can use them off the bat or configure them. They handle all the things URLSession does like authentication, cookies, download types, serialisation etc. With introduciton of Codable I see them less.
What are closures / blocks and how are they used?
Closures and blocks are independent snippets of code that can be passed around. They can be declared inline or given there type names. They are the building blocks of async and multithreading as you can pass a closure to be ran when some other thing has finished, such as an API. They risk memory leaks by causing strong reference cycles and a conscious use of weak or unowned must be used to mitigate this risk.
What are the types of iOS binaries you can create using XCode? (.app, .ipa, .a, .framework)
.app - a bundle designed to represent an app. Tend to contain an executable, info.plist, resources and supporting frameworks .ipa - iOS App Store Package. The file you are given when you archive projects to be submitted to app store. .a - static libraries generated by the archive tool .framework - a bundle designed represent a framework that you can import into your project to use
What is MVVM?
MVVM - Model View ViewModel - attempt to solve Massive View Controller. Model owned by ViewModel, business logic in View Model. ViewController focused on view presentation. View's independent of models. Things become easier to test. ViewModel can hold dependencies to other protocols that provide functionality
How do you test network requests?
Generally we don't test actual network requests, unless you know you can rely on a robust integration server that gives you expected results everytime. Instead we write unit tests and mock network responses. If you build your service in layers and use completions, or promises or some sort of wrappe for the response, you can create mock responses and fire them as and when you need in tests.
What is HTTP?
HTTP is Hyper Text Transfer Protocol. It is the backbone of today's internet. It defines how API requests are made and received. Typically we use HEAD, GET, POST, PATCH, PUT and DELETE to talk out APIs. HEAD gets you the headers such as 200 OK or 404 Not Found. GET fetches data of a sort. POST posts data. PATCH updates a subsegment of data. PUT updates a whole part of data. DELETE deletes. They largely align with CRUD. HTTP requests achieve the majority of what we require other than real-time updates for which we can use sockets.
What is the Delegate pattern in iOS?
The delegate pattern is a varation of the observer pattern. It defines a one-to-one relationship between the delegator and delegatee. An example of the pattern in use is a tableView who requires a datasource and delegate. But only one of each. Protocols are often used to implement the delegation pattern. You define that your class requires a delegate that conforms to x protocol so you can send messages to it. Often delegates are declared as weak to avoid strong reference cycles.
What are ViewControllers? What is the lifecycle for one?
A ViewController is a class that subclasses UIViewController and must have a view to draw on. Any marginally complicated UI should use a ViewController. UIViewController gives us access to lifecycle methods so we know what the view is doing such as loading, appearing, disappearing. It also gives us easy access to presentation logic if embedded in a navigation controller or similar. Typical lifecycle event are viewDidLoad (do setup), viewWillAppear (do things that may need updating as view was already loaded), viewDidApper, layoutSubviews, viewWillDisappear and viewDidDisappear. LoadView can be used if not using nibs to load programmatically.
How should consider using optionals?
Be wary of using them too much due to cognitive strain and effort to unwrap and consider the optional case. Favour optional binding over force unwrapping though to avoid runtime errors.
Section 2
(50 cards)
How you will store user info (username, password or token) securely in iOS?
Use the keychain.
How we can wait for some thread to finish before starting another?
Use operation queues and add dependencies between tasks. Use a dispatchGroup, dispatch an activity to happen, call wait on the group to wait till the group finishes.
Why create a custom view?
To encapsulate visual appearance for a specific component in your app, that can then be reused.
What does static analyser do?
The static analyser runs through your code and tries to identify bugs before running.
Difference between bounds and frame?
Bounds is the size and location of a view relative to it's own coordinate space, frame is the size and location relative to it's superview
How do you work with UIScrollView?
UIScrollView is used to display content too big for the screen. Typically you constrain your scroll view to the size of the device and define scrolling behaviour. You can also become it's delegate and use methods like scroll view did scroll to define more complcated behaviours.
How do you work with UITableView?
UITableView is class designed to show a list of scrollable content. UITableViewCells define the content and how it's displayed. Cells and are dequeued and reused to enhance performance and reduce memory impact. UITableViewDataSource is the source of data for the view defining things like how many rows, which cell to dequeue for given indexPath UITableViewDelegate determines things like cell heights, and lifecycles such as willDisplay and interactions like didSelectRowat.
Difference between viewDidLoad and viewDidAppear?
viewDidLoad is called when a view was loaded, it may only be called once while viewDidAppear may be called multiple times when a view has been loaded. For example, if an app is backgrounded then reopened, viewDidLoad may not be called but viewDidAppear would be.
What database options are there for iOS?
Core Data - the most common. API provided by Apple. SQL under the hood. Uses an object graph which can be confusing. No explicitly thread-safe. Realm - fast, secure, cloud platform backup, reactive, queriable. Thread safe. SQL - you can store in SQL directly without using Core Data though this is rar.
Why UIControl is provided if we can create custom UIView?
Makes it easier to create more interactive elements by overriding control methods and responding to UIControl.events.
What are the challenges in working with UI on iOS?
Different device screen sizes and orientations. Safe areas of the X + screens vs iPhones with a button. Use on iPads and split screen views. UI performance with dynaimc cell sizes for TableViews and CollectionViews. Autolayout tries to resolve this using constraints.
How do you optimize table views performance for smooth, fast scrolling?
The main issue is calculating the correct cell height. You can use frames or constraints to get the correct cell height, but it can take time to calculate which impacts performance. You can solve this by calculating the cell height yourself and using UITableViewAutomaticDimension on your tableview or keeping a prototype cell that you fill with content and use to calculate cell height. A radical approach would be to use something like ComponentKit, a framework from Facebook designed for performant scrolling.
How we can layout subviews in a view?
Adding subViews will call layoutSubviews on the subview, superview and it's other subviews. Otherwise we can call layoutIfNeeded to layout straightaway or setNeedsLayout to layout on next drawing cycle.
How to animate view with constraint?
Use UIView.animateWithDuration. In the animate block change the constraints. Call layoutIfNeeded after on the views changed.
What do you use to lay out your views correctly on iOS?
You can use frames and the coordinate system. I tend to use Autolayout with programmatic constraints and snapkit. You could use an alternative such as LayoutKit created by LinkedIn to increase performance for complicated view hierarchies in scrollable views. Like lots of dynaimc content Or componentkit
Difference between UIWindow and UIView.
UIWindows don't have content themselves but are a container for displaying content. UIViews display content to a particular portion of the app.
Explain ~/Documents, ~/Library, ~/tmp. What directory is ~ on iOS?
tmp folder is mainly for stuff not required to be persisted. Library folder can store things the user need not know about or have access to. Documents folder should be available to user so they can share or access through iTunes.
What alternative ways of working with UI do you know?
React is an popular framework provided by Facebook for declartive UI, but it uses JavaScript under the hood. Alternatively Facebook released ComponentKit as an Objective-C++ native solution. LayoutKit is a LinkedIn version you could try.
What is NSUserDefaults? What type of data can we store there?
UserDefaults is a simple key-value store that is persisted on the device. It is not secure. It can store any codables. It is mainly used for simple user preferences.
How do you work with storyboards in a large team?
The biggest problem comes from merge conflicts on a large storyboard. Alternatives are to define UI in code or split storyboard to smaller sections of app, often one per viewcontroller.
How do you check if your code has memory leaks?
Could use the memory graph debugger and include live allocations to see what objects are referencing what. Could use instruments, it has Allocations tool, can be used to view allocations over time.
What are layers?
CALayers are simple rectangles on the screen where you may draw and display things.
How do you unit and integration test UI?
Standard Unit Test can test some basic UI, such as, does the view have expected sub-views, are they in the correct state, enabled etc. correct labels. Snapshot Testing is a recent way. You record a snapshot image of your view, then test whether your code continues to match the snapshot. XCUITesting is another way to test UI but more typically test user flows and element existence.
Why do we override drawRect: method? How to explicitly call it?
Don't call it. Call setNeedsDisplay. Override to implement custom drawing logic. For example, you may start drawing bezierpaths or arcs.
Notifications remote? How?
Enabled push notification capability in app. Creates a new App-ID. Request permissions from user for notifications. Call UIApplication.shared.registerForRemoteNotifications. Use AppDelegate didRegisterForRemoteNotifications to get device token. Notifications are sent. AppDelegate launchOptions will contain infor or didReceiveRemoteNotification. Payload is simple json telling you about the notification. You may have to create a new key in your certificate to test notifications.
What is bundle?
Bundles are simply directories, but Apple use them with the .app and .framework extensions to bundle up app executables and frameworks and resources. The bundle can be used a referenced in code.
What design patterns are commonly used in iOS apps?
MVC - Model View Controller - design pattern for splitting UI, interaction, and data. Advocated by Apple. Good for small apps. Can result in Massive View Controllers. Singleton - one and only instance of objects. OK for somethings. Used for UIApplication instance for example. Can be overused. Dangers of global state, race conditions, difficulty testing, lack of dependency injection. Delegation - variant of the Observer pattern where one and only one delegate is responsible for dealing with activity. Used in things like TableViews and Delegates. Observer - observers can subscribe to observees it you like and respond when the emit events. Can have multiple observers. Used in the Notification Center and if you use Key-Value Observing. Also a prominent feature of RxSwift if you use it. MVVM - Model View ViewModel - attempt to solve Massive View Controller. Model owned by ViewModel, business logic in View Model. ViewController focused on view presentation. View's independent of models. Coordinators - split responsibility for navigation and user flows out of View Controllers and parent Navigation Controllers, useful for deep linking. Easier to test.
What is MVC? What are benefits and negatives? Alternatives?
MVC - Model View Controller. Model is the data - domain model objects Views - are views showing things to users and interpreting input Controller - is business logic, handling data changes and telling views to change. Advantage - easy to understand, quick to implement, good for small apps and new developers Disadvantage - iOS ViewController becomes too large with too many responsibilities, violates SRP, becomes more difficult to test Alternatives - MVVM-C, general conscious use of SRP to layer the architecture more, so network clients and database clients
What are the compression resistance and content hugging properties?
They define how likely your view will be squeezed or expand given other constraints. A high compression resistance means your view is less likely to shrink if another constraint forces it to. A high content hugging property means you view is more likely to hugh it's content rather than grow. Often these are used with labels as you may want a label's size to stick to it's intrinisic content.
What is AutoLayout? When and where would you use it?
Autolayout is Apple's way of defining layouts based on constraints and relative positions of views. Constraining the width height or position of elements means Autolayout can determine the correct frame to give the view. Constraints can be defined in code, storyboards or nibs. Constraints have a number of different options including should they be equalTo, lessthan or greater than, what is their constant and what is their priority To make working with constraints easier you could use something like SnapKit or PureLayout.
Can we have multiple UIWindows in iOS?
Yes, if we have multiple screens for example.
What is Wildcard App IDs?
Wildcard app id can be used to build the majority of apps whereas explicit App IDs are required for push notifications and in-app payments. Xcode automatic signing tends to handle this.
Have you uploaded app on Appstore? What is the process?
Create a provisioning profile and certificate for distribution. This can be automatically handled for you by Xcode and App Store Connect if you want. Archive a version on Xcode, may have to increment version and build for an updated app. Create or create new version of app in app store connect. Fill in data. Add your build. Submit your build for review. Alternatively submit to TestFlight for beta testers.
How we can execute some code when app is in background?
Yes. Either use a backgroundTask to gain sometime when the app is backgrounded. Or use Background Fetch capability, this will launch you app in Background to perform activity. Can also achieve this by subscribing to certain API's like significant location changes.
Explain concept of notification center, local?
The Notification Center uses the Observer pattern. Objects can register themselves as observers or listeners for a particular notification and target a method. When that notification is posted the observer can respond and process. Used locally it's as simple as post notification and observing. Remember to remove observers before deallocation.
How do you make a pixel-perfect UI according to a designer's specs?
Ideal world the designer would have considered ever edge-case, screen size, data possibilities, but the reality is pixel-perfect design is difficult. Best solution is a quick feedback loop between you and designers, stakeholders and testers so you can find edge cases and deal with them quickly
How we can do multithreading in iOS?
Use Grand Central Dispatch to dispatch blocks to background threads. Alternatively use OperationQueues for more control, e.g. adding dependencies between queues, monitoring queue progress, controlling number of concurrent tasks. Or a DisptachGroup to process multiple things concurrently and wait for all to finish.
Explain application life cycle or application states.
An App is either not running or backgrounded. At some bottom it launches and becomes active. It can then become inactive, be backgrounded and finally terminated. All this lifecycle states are received by our AppDelegate. Methods like appWillLaunch and didLaunch gives us time to setup and prepare our app. The App may resignActive for phone calls and then becomeActive again. The App can be backgrounded by user. The App can be terminated by user or purged from memory by the OS.
What can you use to store data in iOS?
In memory constructs like arrays, dictionaries and sets. They can't store lots of data and aren't readily persistable. UserDefaults / Keychain. Simple key-value stores, one insecure, the other secure. Easy, fast and persistent. Can't handle lots of data or extensive querying. File / Disk Storage. Stor files and data to disk given a URL. Persistent. Not query or database like. Databases like Core Data or Realm. Bigger setup but more querying, storage, complicated relationships.
What is the difference between static and dynamic libraries?
Static libraries are compiled and included in the executable. Static libraries are linked at compile time. Every app using the library gets its own copy, increasing app size. Dynamic libraries are outside the executable and can change. Dynamic libraries are linked when loading or at runtime. Apps can use a single copy.
How is data mapping important when you store data?
The main thing to consider is that when you load from the database, data will go from database to raw data objects to your custom doman models and visa versa. As an example querying Core Data will get you back NSManagedObjects, you could skip a step and use these across your application, but then you are tightly couple to Core Data, plus you lose thread-safety. So it's better to map data to and from NSManagedObjects. Likewise, when using UserDefaults or the keychain, you have to grab the raw data, then somewhere deserialize that data to your custom objects for use. So data mapping should maybe it's on layer between db retrieval and saves to help loosely couple.
What are benefits of collections views?
CollectionViews allow you to have TableView style functionality with added customisations. Such as varying scroll directions. Grid style layouts. Different cell shapes and sizes. Can implement drag and drop. Multiple cells in one row.
Difference between Developer and Enterprise Developer accounts?
Developer accounts intended for app store, enterprise accounts intended for distributing apps through an enterprise
Common reasons for app rejection from Appstore review process?
Crashes, bugs, placeholder content, misleading users, poor user interface against guidelines.
Difference between linking and embedding a framework?
Embedding a framework will include the compiled framework in your app or framework.
Is UIKit thread safe?
Drawing is apparently thread-safe, but not all UIKit is, hence it tends to be used on the mainThread.
Difference in Xib/storyboard vs nib?
Xibs are interface builder files that define UI, when compiled they become nibs. For the most part they've been replaced by storyboards.
What are various singleton instances provided by frameworks?
UIApplication.shared, FileManager.default, UserDefaults.standard, URLSession.shared.
What are different Instruments Xcode supports for app profiling?
CPU profiling, memory profiling and time profiling.
How does Autolayout work with multi-threading?
Every UI change must be done on the main thread. If you change a constraint then do it on the main thread. Call setNeedsLayout to make UIKit aware it needs to update.
Section 3
(26 cards)