Skip to content

Month: July 2016

What is a Custom Renderer?!

One of the most initially confusing, yet essential components to Xamarin.Forms are custom renderers. If you look at almost any advanced feature or customization tutorial for Forms, you are likely to run into the term custom renderer. But what the heck is a custom renderer?!? Everyone seems to reference the use of them, yet no one seems to take the time to thoroughly explain what they are and how they work.

Custom renderers are really simple once you understand the foundation behind how Forms renderers the user interface. Forms maps every UI element to each of the native platforms’ UI elements. In other words, Forms has generic views/controls that are referenced in XAML and are transformed into the native platforms equivalent view. For example, Forms has a Xamarin.Forms.Button class, which gets rendered to UIKit.UIButton on iOS, and android.widget.Button on Android. Essentially, a renderer is simply a mapping between a Forms element, and a platform’s native element. A custom renderer goes one step further; the developer, can override or extend Xamarin’s renderer.

Why would we ever want to change Xamarin’s default renderer? While Xamarin implements a majority of our UI/UX needs, we might need more granular control over the look or behavior of our views.

How do we create a custom renderer? A simple scenario (yet annoying) that requires the use of a custom renderer is when you want to prevent autocorrect on an Entry field when rendered on iOS. An Entry field is Forms’ version of a textbox. Unfortunately, by default on iOS, the textbox has autocorrection and first letter capitalization turned on, yet Forms does not provide a simple way to disable these settings. This is an example of a situation where a custom renderer is required to accomplish our goal.

Suppose you have a simple login page that uses an Entry field for the username and password:

Notice that we are using the default Entry tags in the LoginPage’s XAML. The magic does not happen within the portable class library (PCL); rather, all of our customization will be handled in each platforms’ native project.

  1. For this example, create a folder in the iOS project called: Renderers
  2. Within that new folder, create a new class called: CustomEntryRenderer
  3. Add the following line to the class before the namespace: [assembly: ExportRenderer(typeof(Entry), typeof(MyCoolApp.iOS.CustomEntryRenderer))]
  4. Have the class inherit from: EntryRenderer
  5. Override: OnElementChanged

Within a custom renderer class, there are two important properties to be aware of, Control and Element. The Control property refers to the platform’s native view. The Element is the Forms’ mapping to the Control. For this simple example, we only need to manipulate two settings on the native Control, AutocapitalizationType and AutocorrectionType.

Now, whenever we use an Entry element within our PCL project, this custom renderer will override the default iOS renderer due to the assembly reference we added before the namespace. One nice design aspect to note is that the our views are blissfully unaware of the custom renderer (i.e. they are loosely coupled). That also means that since we did not implement an Android custom renderer, Android will continue to use the default renderer.

Navigation using MVVM Light

MVVM Light is one of many free MVVM frameworks available today. As the name implies, MVVM Light is a lightweight framework that allows the developer to utilize as much or as little of the framework as is needed.

Conversely, other MVVM frameworks are more of an “all or nothing” design. The problem with the “all or nothing” design is that they essentially replace the base system that you are developing on. For instance, if you are developing with Xamarn.Forms, and you adopt one of beefier MVVM frameworks, the framework will completely override a large percentage of how Forms was intended to operate. This can cause major frustration when reading blogs, tutorials, and other coding examples.

As we discussed in the previous post, Forms utilizes XAML for UI implementations and thus is intended to follow the MVVM design pattern. That being said, a large percentage of Xamarin developers (including myself) agree that Forms does not by default correctly handle navigation using the MVVM pattern because the Views dictate the navigation logic not the ViewModels.

For anyone that is just starting out with Forms, navigation can be very frustrating. Fortunately, there is an easy solution to move navigation logic from the Views to the ViewModels using MVVM Light.

We first need to add the MVVM Light package to each project in our solution (PCL, Android, iOS, and any flavor of Windows that you are supporting).

  1. Right click Packages in the solution explorer
  2. Click Add Packages
  3. Use NuGet’s search function to find MVVM Light libraries only package (Id: MvvmLightLibs, Author: Laurent Bugnion (GalaSoft))
  4. Add Package (to each project)

MVVM Light will add the dependency CommonServiceLocator package by Microsoft. Before we implement our navigation service, we need a way to globally access this service so that all ViewModels have access. This is where the Locator comes into play.

In the PCL project, create a folder called Services. Within the Services folder create a class called Locator. We only want one Locator for the entire application, so we will follow the singleton pattern and make the Locator constructor static.

There are three main components to this class, the string constants at the top, the constructor, and various properties with getters. The constants act as keys for each page registered for the navigation service. The constructor handles all the registration of services and ViewModels; ViewModels do not have to be created within the locator, they can be instantiated by the view. Finally we expose the Navigation Service and ViewModel instances with property getters.

Next, let’s implement the NavigationService. In the same folder as the Locator, create a class called NavigationService. This class will implement the MVVM Light (GalaSoft) INavigationService interface. Add the following methods and properties:

At this point, we still do not have a globally accessible NavigationService. Open the class code-behind that implements Application; this is usually called App.xaml.cs. Add the following locator property:

Our ViewModels are now able to completely handle our navigation logic using App.Locator.NavigationService.NavigateTo(PAGE KEY).

Example:  App.Locator.NavigationService.NavigateTo(Locator.SecondPage);

This tutorial is based off of Mark’s Blog.

Xamarin.Forms, MVVM, and Navigation

Coming from a Xamarin native background, I was not entirely familiar with the MVVM pattern prior to switching to Xamarin.Forms. The native designs for iOS and Android typically follow the MVC pattern. MVVM was designed by Microsoft specifically for applications that utilize XAML to design their UIs. If you are familiar with MVC, MVVM is not wildly different. MVVM stands for Model-View-ViewModel.

MVVM Pattern
Source: Microsoft.com

A very rudimentary way to describe MVVM is that you have your typical data models which make up the Model part of the acronym. These data models represent the objects and business logic within your application. Next, we have the View; this is exactly what it sounds like, essentially everything that is rendered on the screen is part of the View.  The ViewModel is the unique (or not so unique) aspect of the pattern. This is where all of your view logic and states are handled. Additionally, the ViewModels act as the middleman between the Views and the Models.

The main objective to using the MVVM pattern is to create loosely coupled code between the Views and the ViewModels. In other words, the Views know about the ViewModels; however, the ViewModels should not know about the Views. Additionally, the Views should not know about the (data) Models and the Models should not know about the Views, or the ViewModels. The purpose behind such a design allows developers to quickly and easily swap out Views with other Views without having to change the business and application logical. Furthermore, this design allows separate teams to work on different components of the project. For example, on a large team, designers could work on the Views of the application (in XAML without any other programming knowledge), while developers could handle the ViewModels, and Models.

Xamarin.Forms was designed with MVVM in mind due to the adoption of XAML. Most of the framework has implemented the tools necessary to utilize the MVVM pattern. However, in my opinion (and the opinion of others) one aspect of the framework design completely breaks the MVVM pattern, and that is navigation.

Xamarin.Forms handles navigation using the INavigation interface available on any Page type (aka the Views). INavigation acts as a globally available navigation stack for your entire application. You might be thinking, that’s great, what’s the problem? The main problem, is that the Views are in charge of handling navigation, not the ViewModels. This technically breaks the MVVM pattern because the Views are dictating how your application operates, not the ViewModels. Under this design, it becomes very difficult to replace a particular View without fixing the logical structure of your application.

So what’s the solution? The solution is to move all of your navigation logic into your ViewModels. In my next post, I will go into detail of how to do this using a framework called MVVM Light.

Custom App Header in Forms

iOS Navigation Bar NestedIt is safe to assume that any business with a customer facing application will want to brand their application with at least their logo. From the very onset of using Xamarin.Forms, I found myself annoyed with the lack of customizability of the navigation bar for iOS. To be fair, it is a pain to customize on the native side; however, the Forms’ abstraction removes most of your options.

NOTE: the navigation bar shows up when a Page is nested within a NavigationPage; this creates a navigation stack.

The equivalent implementation for Android is a toolbar or action bar. Within the Android project, you can customize the toolbar using the native Toolbar.axml layout file. Unfortunately, in order to customize the iOS navigation bar to the same extent as Android requires a custom renderer and a fair amount of hacking. Even with the ability to manipulate the look of the iOS navigation bar using a custom renderer, both platforms are now diverging from each other (not to mention if you support Windows). This messy solution wasn’t going to work for me. Two simple features resolved this dilemma.

I came to the realization that I could not use the navigation bar (iOS) or toolbar (Android) and maintain a unified solution between platforms. Simply put, I wanted the platform specific bars to disappear. Thankfully, Xamarin provides a simple way to turn them off. Using the static reference to NavigationPage provides the method SetHasNavigationBar. This method takes two parameters, the page the setting applies to and the boolean value.

At this point my navigation stack no longer has a header. This allows me to add a completely custom implementation that will work for all platforms. An easy way to create such a header is with the combination of a boxview, a back button, and a logo. The specifics are not important because it will vary greatly between applications. An example could look like the following:

But how do I apply this XAML to each of my pages?! This can be accomplished with a ControlTemplate. A ControlTemplate is a UI template that can be applied to any XAML page. You’ll define the template in the ResourceDictionary located in the App.xaml file. You will also include a ContentPresenter within the template. A ContentPresenter is the content area or body of the pages the template is being applied to. The ControlTemplate Key will be used to reference the template in your pages.

Once the ControlTemplate is created, we need to apply it to our Pages. This is accomplished by adding a StaticResource reference of the template Key inside our pages’ ControlTemplate property in the XAML.

iOS Custom App HeaderYou now have a fully customizable header for your application. You can add EventHandlers, Commands, or GestureRecognizers to interact with your header as needed. You will have to implement the back button logic as well.

Why should software developers have a blog?

Of the many conference sessions I attended at Xamarin Evolve16 this year, there was one that was radically different than the others. When you go to a developer conference, you assume that all the sessions are focused entirely around the latest technology trends and improving your development skills. What I did not expect was a session on skills to improve the quality of life and long term success as a software developer. As software developers, we have very analytical minds, and thus we only focus on our quantitative technical skills and ignore the qualitative. In the session, 5 Soft Skills Every Software Developer Should Know, I was introduced to John Sonmez, and his mission is to change that.

John Sonmez is a software developer who was so successful in his day job, he was able to retire at the age of 33. Since then, he has become a software developer coach who has made it his life’s mission to helping developers succeed in their careers and surpass mediocre lives. John has built a unique business revolving around this core mission, called Simple Programmer.

Simple Programmer provides a multitude of services for software developers including blog posts, podcasts, YouTube videos, published books, and training courses. A large majority of this content is completely free and worth exploring.

One course in particular that is worth investigating is his free email course on building a blog. In this course, John sends you weekly emails explaining the simple process of how to build a blog correctly and the reasons for it.  While the content of his emails might first appear to be common sense, if they were so common, why do some many developers do it incorrectly, or not at all?

That’s nice, but why should I even care about blogging? The premise behind developers maintaining a blog is that it acts as a conduit to creating a brand for yourself.  Developers without a brand (or name for themselves) are simply code monkeys that can be easily replaced with another developer. Adversely, developers that have both the technical skills and name recognition within an industry are sought out by employers. In essence, you want to be name not a number in a stack of resumes.

Beyond simply name recognition within an industry, blogging provides a channel of communication between other developers. While Stack Overflow is the usual go to for quick technical issues, blogs are the second resource type that ALL developers depend on. Not only are they another tool in our tool-belt, they are also better mediums for more in depth topics; the format is more conducive for tutorials and longer articles than Stack Overflow.

Most developers usually have a healthy level of skepticism. Why should we trust this John guy whom you might not have ever heard of before? Don’t just listen to him, listen to the advice of other major names in the industry like Scott Hanselman. While John’s advice isn’t some magic get rich quick pill, his advice is a compilation of life skills for developers that, with hard work, could completely change the course of your career.