msicc

Sending push notifications to your Xamarin app from WordPress with Azure, Part II – the Function

Sending push notifications to your Xamarin app from WordPress with Azure, Part II – the Function

First, let’s have a look at the lineup of this series once again:

  • Preparing your WordPress (blog/site)
  • Preparing the Azure Function and connect the Webhook (this post)
  • Preparing the Notification Hub
  • Send the notification to Android
  • Send the notification to iOS
  • Adding in Xamarin.Forms

Creating a new Azure Function in Visual Studio

The most simple approach to create a new Azure Function (if you already have an Azure account) is adding a new project to your Xamarin solution:

After the project is loaded, double click on the .csproj file in the Solution Explorer to open the file for editing it. Make sure you have the following two PropertyGroup entries:

  <PropertyGroup>
    <TargetFramework>net461</TargetFramework>
    <AzureFunctionsVersion>v1</AzureFunctionsVersion>
  </PropertyGroup>
  <ItemGroup>
    <!--DO NOT UPDATE THE AZURE PACKAGES, IT WILL BREAK EVERYTHING!!!!-->
    <PackageReference Include="Microsoft.Azure.NotificationHubs" Version="1.0.9" />
    <PackageReference Include="Microsoft.Azure.WebJobs.Extensions.NotificationHubs" Version="1.3.0" />
    <PackageReference Include="Microsoft.NET.Sdk.Functions" Version="1.0.31" />
    <PackageReference Include="Newtonsoft.Json" Version="9.0.1" />
  </ItemGroup>

You may notice that I made an all caps comment into the second PropertyGroup entry. As I am using a v1 Function, these are the latest packages that I am able to use. They are doing their job, and allow us to use an easy way to bind the Function to the Azure NotificationHub , which we are going to implement in the next post. I delayed updating the whole setup to use a v2 function intentionally at this point.

Processing the Webhook payload

In order to be able to process the payload (remember, we are getting a JSON string) from our WordPress Webhook, we need to deserialize it. Let’s create the class that holds all information about it:

using Newtonsoft.Json;

namespace NewPostHandler
{
    public class PublishedPostNotification
    {
        [JsonProperty("id")]
        [JsonConverter(typeof(StringToLongConverter))]
        public long Id { get; set; }

        [JsonProperty("title")]
        public string Title { get; set; }

        [JsonProperty("status")]
        public string Status { get; set; }

        [JsonProperty("featured_media")]
        public string FeaturedMedia { get; set; }
    }
}

The class gets it pretty straight forward, we will use this implementation as-is for the payload we are sending to Android later on. The use of the StringToLongConverter is optional. For completeness, here is the implementation:

using Newtonsoft.Json;
using System;

namespace NewPostHandler
{
    public class StringToLongConverter : JsonConverter
    {
        public override bool CanConvert(Type t) => t == typeof(long) || t == typeof(long?);

        public override object ReadJson(JsonReader reader, Type t, object existingValue, JsonSerializer serializer)
        {
            if (reader.TokenType == JsonToken.Null) return default(long);
            var value = serializer.Deserialize<string>(reader);
            if (long.TryParse(value, out var l))
            {
                return l;
            }

            return default(long);
        }

        public override void WriteJson(JsonWriter writer, object untypedValue, JsonSerializer serializer)
        {
            if (untypedValue == null)
            {
                serializer.Serialize(writer, null);
                return;
            }
            var value = (long)untypedValue;
            serializer.Serialize(writer, value.ToString());
            return;
        }

        public static readonly StringToLongConverter Instance = new StringToLongConverter();
    }
}

Now that we prepared our data transferring object, it is time to finally have a look at the processor code.

[FunctionName("HandleNewPostHook")]
public static async Task<HttpResponseMessage> Run([HttpTrigger(AuthorizationLevel.Function, "post", Route = null)]HttpRequestMessage req, TraceWriter log)
{
    _log = log;
    _log.Info("arrived at 'HandleNewPostHook' function trigger.");

    //ignoring any query parameters, only using POST body

    PublishedPostNotification result = null;

    try
    {
        _jsonSerializerSettings = new JsonSerializerSettings()
        {
            MetadataPropertyHandling = MetadataPropertyHandling.Ignore,
            DateParseHandling = DateParseHandling.None,
            Converters =
            {
                new IsoDateTimeConverter { DateTimeStyles = DateTimeStyles.AssumeUniversal },
                StringToLongConverter.Instance
            }
        };

        _jsonSerializer = JsonSerializer.Create(_jsonSerializerSettings);

        using (var stream = await req.Content.ReadAsStreamAsync())
        {
            using (var reader = new StreamReader(stream))
            {
                using (var jsonReader = new JsonTextReader(reader))
                {
                    result = _jsonSerializer.Deserialize<PublishedPostNotification>(jsonReader);
                }
            }
        }

        if (result == null)
        {
            _log.Error("There was an error processing the request (serialization result is NULL)");
            return req.CreateResponse(HttpStatusCode.BadRequest, "There was an error processing the post body");
        }

       //subject of the next post
      //await TriggerPushNotificationAsync(result);
    }
    catch (Exception ex)
    {
        _log.Error("There was an error processing the request", ex);
        return req.CreateResponse(HttpStatusCode.BadRequest, "There was an error processing the post body");
    }

    if (result.Id != default)
    {
        _log.Info($"initiated processing of published post with id {result.Id}");
        return req.CreateResponse(HttpStatusCode.OK, "Processing new published post...");
    }
    else
    {
        _log.Error("There was an error processing the request (cannot process result Id with default value)");
        return req.CreateResponse(HttpStatusCode.BadRequest, "There was an error processing (postId not valid)");
    }
}

Let’s go through the code. The first thing I want to know is if we ever enter the Function, so I log the entrance. The second step is setting up the JsonSerializer to deserialize the payload into the DTO class I created before.

There are several scenarios that I am handling and returning different responses. Ideally, we would run through and arrive at the TriggerPushNotificationAsync call, followed by a jump the ‘OK‘- response if the post id received from our Webhook is valid. During testing, however, I ran into other situations as well, where I return a ‘Bad Request‘ response with a hint that something went wrong.

The implementation of the TriggerPushNotificationAsync method is not shown in this post as it will be subject of the next post in this series.

Testing the code locally

One of the reasons I chose to start the Function in Visual Studio is its ability to debug it locally. If you don’t have the necessary tools installed, Visual Studio will prompt you to do so. After installing them, you’ll be able to follow along.

Once the service is running, we will be able to test our function. If you haven’t already heard about it, Postman will be the easiest tool for that. Copy the function url and paste it into the url field in Postman. Next, add a sample JSON payload to the body (settings: raw, JSON) and hit the ‘Send’ button:

If all goes well, Postman will give you a success response:

The Azure CLI will also write some output:

As you can see, all of our log entries were written to the CLI, plus some additional information from Azure itself. Don’t worry for the moment about the anonymous authorization state, this is just because we are running locally. In theory, we could already publish the function to Azure now. As we know that we will extend the Function in the next post, however, we will not do this right now.

Conclusion

As you can see, writing an Azure Function isn’t as complicated as it sounds. Visual Studio brings all the tools you need to get started pretty fast. The ability to test the Function code locally is another big advantage that comes with Visual Studio.

In the next post, we will configure the NotificationHub on Azure and extend our Function to call into it and fire the notifications.

Until the next post, happy coding, everyone!
Posted by msicc in Android, Azure, Dev Stories, iOS, Xamarin, 1 comment
Sending push notifications to your Xamarin app from WordPress with Azure, Part I [new series]

Sending push notifications to your Xamarin app from WordPress with Azure, Part I [new series]

Overview

Choosing the “right” solution for sending push notifications isn’t easy if you have a WordPress blog. There are quite a bunch of options to choose from, and the right one for you might differ from my decision. I am using the most generic solution – a Webhook that triggers an Azure Function, which triggers the notification via Azure Notification Hubs. This series will grow as follows:

  • Preparing your WordPress (blog/site) (this post)
  • Preparing the Azure Function and connect the Webhook
  • Preparing the Notification Hub
  • Send the notification to Android
  • Send the notification to iOS
  • Adding in Xamarin.Forms

The app implementations are very platform-specific, but it is quite easy to integrate the post notifications in a Xamarin.Forms app (which will be the last post in this series). If you want to see the whole integration already in action, feel free to download my blog reader app:

WordPress plugins for the win

If you have a self-hosted blog like I do, you may know that the plugin ecosystem is there to help you with a lot of things that WordPress hasn’t out of the box. While a WordPress-hosted site as Webhook integration without an additional plugin, we need one to create such a Webhook on a self-hosted WordPress blog. The plugin I am using is simply called “Notification” and can be found here.

To install the plugin, follow these simple steps:

  • choose “Plugins” on your WordPress dashboard
  • select “Add New” and type in “notification”
  • Hit the “Install Now” button
  • Activate the plugin

Exploring the options

Once you have installed and activated the plugin, you will have a new option in the dashboard menu. Let’s have a look at the options.

  • Notifications – this shows you a list of your currently active notifications
  • Add New Notification – lets you create a new notification
  • Extensions – the plugin allows you to extend your notifications with external services like Slack, Twitter or SendGrid to engage even more users. We do not need these for the webhook, however.
  • Settings – the control panel for the plugin – this is where we will be for the rest of this blog

Enabling the Webhook

On the Settings page, select the “CARRIERS” option. The plugin uses so-called carriers to send out the notifications. By default, the Email carrier is active. I do not need this one for the moment, so I deactivated it an activated the Webhook carrier instead:

Setting Post Triggers

The next step is to verify we have the trigger for posts active:

You can modify the other triggers as well, but for the moment, I am focusing just on posts. I am thinking about integrating comments in the future, which will allow even more interaction from within my app.

Add a new notification

Let’s create our first notification. Select the “Add New Notification” action, which will bring up this page:

Select the “Add New Carrier” option and add the Webhook carrier:

Next, select the Trigger for the Webhook. During development, I am using the saving draft option as it allows me to easily trigger a notification without annoying anyone:

This will enable the “Merge Tags” list on the right-hand side. To create the Webhook payload, we need to add some arguments (using the “Add argument” button). Tip: you can copy the merge tag by just clicking on it and paste it into the “Value” box:

Don’t forget to activate the JSON format – we do not want it to be sent as XML. Make sure the Carrier is enabled and hit the save button on the upper right.

Testing the Webhook

Now that we finished the setup of our Webhook, let’s test it. To do so, go to the “Settings” page again and select “DEBUGGING”. Check the “Enable Notification logging” box and click the “Save Changes” button:

To test the notification, just create a new blog post and save the draft. Go back to the “DEBUGGING” setting, where you will be presented a new Notification log entry. Expanding this log entry, you will see some common data about the notification:

If you scroll a bit further down, you will see the payload of the Webhook. Sadly, you won’t get the raw JSON string, but a structured overview of the payload:

Verify that the payload contains all the data you need and adjust the settings if necessary. Once that is done, we are ready to go to the next blog post (coming soon).

Conclusion

In this post, I showed you how to create a Webhook that will trigger our upcoming Azure Function. Thanks to the “Notification” plugin, the process is pretty straight forward. In the next post, we will have a look at the Azure Function that will handle the Webhook.

Until the next post, happy coding, everyone!

Posted by msicc in Android, Azure, Dev Stories, iOS, Xamarin, 1 comment
MSicc’s Blog version 1.6.0 out now for Android and iOS

MSicc’s Blog version 1.6.0 out now for Android and iOS

Here are the new features:

Push Notifications

With version 1.6.0 of the app, you can opt-in to receive push notifications once I publish a new blog post. I use an Azure Function (v1, for the ease of bindings – at least for now), and of course, an Azure NotificationHub. The Function gets called from a WebHook (via a plugin on WordPress), which triggers it to run (the next blog posts I write will be about how I achieved this, btw.)

New Design using Xamarin.Forms Shell

I also overhauled the design of the application. Initially, it was a MasterDetail app, but I never felt happy with that. Using Xamarin.Forms.Shell, I optimized the app to only show the last 30 posts I wrote. If you need older articles, you’ll be able to search for them within the app. The new design is a “v1” and will be constantly improved along with new features.

Bugs fixed in this release
  • fixed a bug where code snippets were not correctly displayed
  • fixed a bug where the app did not refresh posts after cleaning the cache
  • other minor fixes and improvements

I hope some of you will use the application and give me some feedback.

You can download the app using these links:
iOS | Android

Until the next post, happy coding, everyone!

Posted by msicc in Android, Azure, Dev Stories, iOS, Xamarin, 0 comments
WordPressReader (Standard) updated to version 2.1.0

WordPressReader (Standard) updated to version 2.1.0

Currently, I am working on an update for my blog reader app on iOS to bring it on par with the Android version (which snagged push notifications for new posts already). As my WordPressReader library sits at the core of the application, I made also some updates to this library.

Here is what’s new:

  • based on .netStandard 2.1 (still works across UWP, WPF, Xamarin)
  • activated nullable reference types
  • because of the nullable reference types, updated the entity classes to always include null values (for type safety)
  • fixed a bug where comments in response to pingbacks lead to a crash
  • renamed CreateAnonymousComment to CreateAnonymousCommentAsync
  • other minor fixes

The update to the app is already available via Nuget, while you can check the source code in the Github repo.

What’s next:

I am currently exploring further improvements using C# 8.0, most notably IAsyncEnumerable<T> and ValueTask.

If you have any feedback for the library, feel free to sound off in the comments, open a Github issue or contact me via my social media accounts.

Until the next post, happy coding, everyone!

Title Image by Kevin Phillips from Pixabay

Posted by msicc in Dev Stories, 1 comment
#XfEffects: Forms Effect to automatically scale FontSize on Label

#XfEffects: Forms Effect to automatically scale FontSize on Label

Why do I need this?

When working with text, we often have to deal with some or all of the following:

  • dynamic text with different length on every instance
  • multiple devices with different screen resolutions
  • limited number of lines

As the amount of places where I need to automatically scale the FontSize is steadily increasing within my apps, I had to come up with a solution – the AutoFitFontSizeEffect.

The shared code

Of course, every Effect has a shared code part. Like in my first post, there are two classes for this – the Effect wrapper and the static parameter class on top of it. The wrapper is pretty straight forward:

    public class AutoFitFontSizeEffect : RoutingEffect
    {
        #region Protected Constructors

        public AutoFitFontSizeEffect() : base($"XfEffects.{nameof(AutoFitFontSizeEffect)}")
        {
        }

        #endregion Protected Constructors
    }

Like with all effects, we are deriving from RoutingEffect and initializing the class with the effect’s id. As I wanted the effect to be configurable with a minimum and a maximum size, I added a static class that takes these two parameters:

    public static class AutoFitFontSizeEffectParameters
    {
        #region Public Fields

        public static BindableProperty MaxFontSizeProperty = BindableProperty.CreateAttached("MaxFontSize", typeof(NamedSize), typeof(AutoFitFontSizeEffectParameters), NamedSize.Large, BindingMode.Default);
        public static BindableProperty MinFontSizeProperty = BindableProperty.CreateAttached("MinFontSize", typeof(NamedSize), typeof(AutoFitFontSizeEffectParameters), NamedSize.Default, BindingMode.Default);

        #endregion Public Fields

        #region Public Methods

        public static NamedSize GetMaxFontSize(BindableObject bindable)
        {
            return (NamedSize)bindable.GetValue(MaxFontSizeProperty);
        }

        public static NamedSize GetMinFontSize(BindableObject bindable)
        {
            return (NamedSize)bindable.GetValue(MinFontSizeProperty);
        }

        public static double MaxFontSizeNumeric(BindableObject bindable)
        {
            return Device.GetNamedSize(GetMaxFontSize(bindable), typeof(Label));
        }

        public static double MinFontSizeNumeric(BindableObject bindable)
        {
            return Device.GetNamedSize(GetMinFontSize(bindable), typeof(Label));
        }

        public static void SetMaxFontSize(BindableObject bindable, NamedSize value)
        {
            bindable.SetValue(MaxFontSizeProperty, value);
        }

        public static void SetMinFontSize(BindableObject bindable, NamedSize value)
        {
            bindable.SetValue(MinFontSizeProperty, value);
        }

        #endregion Public Methods
    }

Let’s break that class down. First, I created two attached BindableProperty objects of type NamedSize. The NamedSize enumeration makes it easy for us to determine the minimum and maximum sizes. If you want to know the values behind the enum entries, check this table in the docs.

To get and set those values out of the BindableProperty, I implemented corresponding methods. As we will see later on, I implemented also two methods that get the numeric values, which will be used in our platform-specific implementations.

Android implementation

Android has a built-in method on TextView to achieve the auto-scaling functionality we desire (read more about it on the Android docs). This makes the implementation pretty straight forward:

    public class AutoFitFontSizeEffect : PlatformEffect
    {
        #region Protected Methods

        protected override void OnAttached()
        {
            if (this.Control is TextView textView)
            {
                if (AutoFitFontSizeEffectParameters.GetMinFontSize(this.Element) == NamedSize.Default &&
                    AutoFitFontSizeEffectParameters.GetMaxFontSize(this.Element) == NamedSize.Default)
                    return;

                var min = (int)AutoFitFontSizeEffectParameters.MinFontSizeNumeric(this.Element);
                var max = (int)AutoFitFontSizeEffectParameters.MaxFontSizeNumeric(this.Element);

                if (max <= min)
                    return;

                textView.SetAutoSizeTextTypeUniformWithConfiguration(min, max, 1, (int)ComplexUnitType.Sp);
            }
        }

        protected override void OnDetached()
        {
        }

        #endregion Protected Methods
    }

Before using the SetAutoSizeTextTypeUniformWithConfiguration method on the TextView, I am running two checks: one if both parameters are set to NamedSize.Default, and the other one if the minimum value is bigger than the maximum value. If we pass past these check, we are using the above mentioned method. That is already everything it needs to make the text scaling automatically within the bounds of the TextView on Android.

iOS implementation

Like Android, also iOS has a pretty easy way to automatically scale the FontSize:

    public class AutoFitFontSizeEffect : PlatformEffect
    {
        #region Protected Methods

        protected override void OnAttached()
        {
            if (this.Control is UILabel label)
            {
                if (AutoFitFontSizeEffectParameters.GetMinFontSize(this.Element) == NamedSize.Default &&
                    AutoFitFontSizeEffectParameters.GetMaxFontSize(this.Element) == NamedSize.Default)
                    return;

                var min = (int)AutoFitFontSizeEffectParameters.MinFontSizeNumeric(this.Element);
                var max = (int)AutoFitFontSizeEffectParameters.MaxFontSizeNumeric(this.Element);

                if (max <= min)
                    return;

                label.AdjustsFontSizeToFitWidth = true;
                label.MinimumFontSize = (float)min;
                label.Font = label.Font.WithSize((float)max);
            }
        }

        protected override void OnDetached()
        {
        }

        #endregion Protected Methods
    }

We are running the same checks as on Android before we are effectively changing the properties on the UILabel that will make the text scale automatically. With setting AdjustsFontSizeToFitWidth to true and setting the MinimumFontSize to our min value as well as the max value as FontSize, we have already done everything it needs on iOS.

Conclusion

The checks we run before using the codes are not random. It may happen that you only add the effect to your Xamarin.Forms.Label without setting the MinFontSize and MaxFontSize. In this case, I am just returning.

Besides mixing up the sizes, the main reason for the second check is that the platform-specific size values are different between platforms. Also in this case, I am just returning.

Besides that, we are able to use all other properties of the default Xamarin.Forms.Label implementation, with MaxLines and LineBreakMode being the two most important ones.

As always, I hope this post will be helpful for some of you. Of course, the sample project for this series is updated on GitHub.

Happy coding, everyone!

Posted by msicc in Android, Dev Stories, iOS, Xamarin, 3 comments
Visual Studio Extensions that will make your life easier in 2020

Visual Studio Extensions that will make your life easier in 2020

Xaml Styler

Download Link: https://marketplace.visualstudio.com/items?itemName=TeamXavalon.XAMLStyler

Project Site: https://github.com/Xavalon/XamlStyler

In Visual Studio 2019, writing XAML normally ends up in very long lines because it does not automatically break into new lines. In May 2019, I discovered the Xaml Styler extension by watching a video on Channel9.

Xaml Styler, even in its default configuration (which I did never change, tbh), breaks your XAML code into new lines once the count of properties of an element exceeds a certain amount. The extension as tons of options to configure, but for me, the default config always worked well.

Here is a sample:

As you can see, the XAML file is a by far more readable now. The different colors on the namespace declarations and closing tags are coming from the next extension on my list.

Viasfora

Download Link: https://marketplace.visualstudio.com/items?itemName=TomasRestrepo.Viasfora

Project Site: https://viasfora.com/

Viasfora aims to make your code more readable. To achieve that goal, it is coloring the code inside Visual Studio. The coloring makes it easy to see the scopes of each line, method, and even classes. As I discovered the extensions only a few weeks ago, I decided to not change any of the default settings, but it is already making my code a whole lot more readable.

Here is a sample from one of my recent projects:

VSColorOutput

Download Link: https://marketplace.visualstudio.com/items?itemName=MikeWard-AnnArbor.VSColorOutput

Project Site: https://mike-ward.net/vscoloroutput/

Ever tried to search the Visual Studio output for warnings, build errors or exceptions? Of course, you did. VSColorOutput makes this search a whole lot easier, as it colors errors/exceptions in red, warning in yellow, and build success messages in green. You see everything at a glance, which can save you a lot of time. See yourself:

CodeMaid

Download Link: https://marketplace.visualstudio.com/items?itemName=SteveCadwallader.CodeMaid

Project Site: http://www.codemaid.net/

CodeMaid is a very powerful extension. I am pretty sure I am not even using half of its functions, to be honest. I mainly use it to let it automatically sort my code files to my gusto and let it generate regions around it. This way, my code stays always organized in the same way and I don’t even have to think about it. I will explore more of the other functions moving onwards.

Here is a sample of a code file after CodeMaid cleaned it up (collapsed regions):

Material icons generator

Download Link: https://marketplace.visualstudio.com/items?itemName=nikainteristi.Materialiconsgenerator

Project Site: https://github.com/interisti/vs-material-icons-generator

If you are developing mobile applications, chances are high you will need one or another icon within an app. Material icons generator tries to replicate the popular function of Android Studio, with the added bonus of making it also available for iOS and UWP apps. The extension sadly does not create Image Asset folders and .json files for iOS correctly, but it creates the images so one can create an Image Asset from it in just a minute. The project does seem to be actively developed according to Github – I put this extension on the list because I did not find a better alternative until now.

Markdown Editor

Download Link: https://marketplace.visualstudio.com/items?itemName=MadsKristensen.MarkdownEditor

Project Site: https://github.com/madskristensen/MarkdownEditor

If you have to deal with Markdown files, there are floating quite a few helper apps around the web. I prefer to work on them in Visual Studio, as I use Markdown most of the time on Github when writing readme or similar files. Markdown Editor makes it easy to do so, and I almost immediately see the result in the preview window, which is quite helpful from time to time.

The extension is written by the creator of Markdig, which is by far your best option if you have to deal with Markdown in your C# app. Here is a sample:

Regex Editor

Download Link: https://marketplace.visualstudio.com/items?itemName=GeorgyLosenkov.RegexEditorLite

Project Site: N/A

If you ever had to deal with validations, chances are high you solved it with Regex. The extension provides a bunch of options that will help you to write Regex-patterns. You can test it in the same Window with some sample data to verify it will do what it is supposed to do. As a bonus, it is able to create a sample method of how to use it if you want to. Here is what it looks like with an email verification pattern:

Conclusion

To kick off my blogging year 2020, I showed you some of the most helpful Visual Studio extensions I installed on my machine. I hope some of you will find the one or another extension as helpful as I do. If you have an extension that is not on my list and you consider it useful, feel free to sound off in the comments or via social media.

Until the next post, happy coding!

Posted by msicc in Dev Stories, 4 comments
My 2019 (a short review)

My 2019 (a short review)

Bitcoin, Crypto payments & Linux (moving out of my comfort zone)

At the beginning of 2019, I was working on the C#-SDK for AtomicPay, a crypto payment provider. I was also working on a mobile administration app for AtomicPay but during the year, the regulatory rules changed in Switzerland for KYC (know your customer). I had to stop my efforts in this area because, for an indie developer, they are just impossible to handle.

That didn’t stop me to discover other crypto/blockchain stuff, and so I did move out of my comfort zone to play around with a Linux Server and implementing a Bitcoin full node on it. You can read more on this in the Crypto&Blockchain section of my blog.

Even though my efforts have shifted back towards mobile applications, the blockchain area is still of interest to me.

Xamarin, Xamarin, Xamarin

All of my mobile side projects this year were deeply tied into the Xamarin universe. The biggest milestone I finally achieved this year was to finish the porting of my very first application ever from Windows Phone to Android and iOS. I wrote about some of the things I learned right here:

Barcode Scanning

The last few weeks I was busy to write a barcode scanning control for Xamarin Forms that uses native capabilities on iOS and Firebase ML on Android. It was only last week that I finally got both variants working.

Blogging

I attempted to write more regularly this year, but I clearly failed to keep up with it. One of my new year’s resolution is to keep a steady schedule of writing, as the list of what I want(ed) to write about is growing. I will go through this list during my upcoming vacation and try to find a better rhythm for my blogging efforts besides work and working on my side projects.

This year’s review is a bit different from the years before. I kept it short, focusing on the important parts (from my point of view). There is just one thing left to do in this post – to wish you all

A Merry Xmas and a Happy New Year!

Title-Image by Johan L from Pixabay

Posted by msicc in Dev Stories, Editorials, 0 comments
What I’ve learned from porting my first app ever to Android and iOS with Xamarin

What I’ve learned from porting my first app ever to Android and iOS with Xamarin

What’s the app about?

The app is about fishing knots. It sounds boring for most people, but for me, this app made me becoming a developer. So I have a somewhat emotional connection to it. It was back in time when Windows Phone 7 was new and hot. A new shiny OS from Microsoft, but clearly lacking the loads of apps that were available on iOS and Android. At that time, I also managed to get my fishing license in Germany.

As I had a hard time to remember how to tie fishing knots, I searched the store and found… nothing. I got very angry over that fact, partly because it meant I had to use one of the static websites back then, but more about the fact that there was this damn app gap (WP7 users will remember). So I finally decided to learn how to write code for Windows Phone and wrote my first app ever after some heavy self-studying months.

Why porting it?

Writing code should soon become my favorite spare-time activity, effectively replacing fishing. And so the years went on, I made some more apps (most of them for Windows Phone) and also managed to become employed as a developer. After some time, S. Nadella became the CEO of Microsoft, and Windows for mobile phones was dead. So I had created all my “babies”, and they were now set to die as Windows Phone/Mobile did. Not accepting this, I started to create a plan to port my apps over to the remaining mobile platforms. After Facebook effectively killed my most successful app (UniShare – that’s another story, though), I stopped porting that one and started with Fishing Knots +.

Reading your own old code (may) hurt

When I was starting to analyze which parts of the code I could reuse, I was kind of shocked. Of course, I knew that there was this code that I wrote when I didn’t know it better, but I refused to have a look into it and refactor it (for obvious reasons from today’s perspective). I violated a lot of best practices back then, the most prominent ones

  • No MVVM
  • Repeating myself over and over again
  • Monster methods with more than 100 lines within

In the end, I did the only right thing and did not reuse any line of my old code.

Reusing the concept without reusing old code

After I took the right decision to not use my old codebase, I needed to abstract the concept from the old app and translate it into what I know now about best practices and MVVM. I did not immediately start with the implementation, however.

The first thing I did was drawing the concept on a piece of paper. I used a no-code language in that sketch and asked my family if they understand the idea behind the app (you could also ask your non-tech friends). This approach helped me to identify the top 3 features of the app:

  • Controllable animation of each knot
  • Easy-to-follow 3-step instructions for each knot
  • Read-Aloud function of the instructions

Having defined the so-called “Minimum Viable Product“, I was ready to think about the implementation(s).

The new implementation

Finding the right implementation isn’t always straight forward. The first thing I wrote was the custom control that powers the controllable animation behind the scenes. I wrote it out of the context in a separated solution as I packed it into a NuGet package after finishing. It turned out to be also the most complex part of the whole app. It uses a common API in Xamarin.Forms, and custom renderers for Android and iOS. I had to go that route because of performance reasons – which is one of the learnings I took away from the porting.

It was also clear that I will use the MVVM pattern in the new version. So I was setting up some basic things using my own Nuget packages that I wrote during working on other Xamarin based projects.

When it came to the overall structure of my app, I thought a Master/Detail implementation would be fine. However, somehow this never felt right, and so I turned to Shell (which was pretty new, so I tried to give it a shot). In the end, I went with a more custom approach. The app uses a TabbedPage with 3 tabs, one being for the animation, the second for the 3-Step tutorial, and last but not least the Settings/About page. The first two pages share a custom top-down menu implementation, bound to the same ViewModel for its items and selection.

More Xamarin.Forms features I learned (to love)

Xamarin and Xamarin.Forms itself are already powerful and have matured a lot since the time I used it to write my first Xamarin app for Telefonicá Germany. Here is a (high level) list of features I started to use:

  • Xamarin.Essentials – the one library that kickstarts your application – seriously!
  • Xamarin Forms Animations – polish the appearance of your app with some nice looking visual activity within the UI
  • Xamarin Forms Effects – easily modify/enhance existing controls without creating a full-blown custom renderer
  • Xamarin Forms VisualStateManager – makes it (sometimes) a whole lot easier to change the UI based on property changes
  • Xamarin.Forms Triggers – alternative approach to modify the UI based on property changes (but not limited to that)

The three musketeers

Because of Xamarin and Xamarin.Forms are such powerful tools, you may run into the situation of needing help/more information. My three musketeers to get missing information, implementation help or solution ideas:

  • Microsoft Xamarin Docs – the docs for Xamarin are pretty extensive and by reading them (even again), I often had one of these “gotcha!”- moments
  • Github – if the docs don’t help, Github may. Be it in the issues of Xamarin(.Forms) or studying the renderers, Github has been as helpful as the docs to me.
  • Web Search – chances are high that someone had similar problems/ideas solved and wrote a blog about it. I don’t blindly copy those solutions. First I read them, then I try to understand them and finally, I implement my own abstraction of them. This way, I am in a steady learning process.

Learn to understand native implementations

I guarantee you will run into a situation where the musketeers do not help when focusing solely on Xamarin. Accept the situation that Xamarin is sitting on top of the native code of others and does the heavy conversion for us. Learn to read Objective-C, Swift, Java and Kotlin code and translate it into C# code. Once you found possible solutions in one of the native samples, blog posts or docs, you will see that most of them are easy to translate into Xamarin code. Do not avoid this part of Xamarin development, it will help you in future, trust me.

Conclusion

Porting over my first app ever to Android and iOS has provided me not only a lot of fun but also huge learnings/practicing. Some of them are of behavioral nature, some of them are code implementations. This post is about the behavioral part – I will write about some of the implementations in my upcoming blog posts.

I hope you enjoyed reading this post. If you have questions or have similar experiences and want to discuss, feel free to leave a comment on this post or connect to me via social media.

Until the next post, happy coding!

Helpful links:

Posted by msicc in Dev Stories, Xamarin, 1 comment

How to work around IDE freezes in Visual Studio 2019 when switching between Build configurations

After a lot of project unloading, deleting our custom configurations back and forth and project file modifications, we finally found a workaround to continue our work by browsing other/similar issues to ours in the VS feedback forums. The freeze of the solution is caused by a feature that is meant to fasten up solution loading: parallel project initialization.

It seems to be an issue that some others already had with 16.2.x versions, so it may be a regression (as it is flagged as fixed).

The workaround

The workaround is pretty easy, just follow these simple steps:

  • Open Tools / Options in Visual Studio 2019 and find the Projects and Solutions node
  • Unselect ‘Allow parallel project initialization’
  • Click ‘OK’ and close the solution

After that, we need to delete the .vs folder of Visual Studio 2019 within the local solution folder. This folder contains a SQLite DB that corresponds with some behind the scenes stuff for the parallel project initialization (and more).

Now open the solution – switching between Build configurations should now work again.

As always, I hope this post is helpful for some of you.

Until the next, happy coding!

Posted by msicc in Dev Stories, 0 comments
[Updated] #XfEffects: Xamarin.Forms Effect to change the TintColor of ImageButton’s image – (new series)

[Updated] #XfEffects: Xamarin.Forms Effect to change the TintColor of ImageButton’s image – (new series)

The documentation recommends using Effects when we just want to change properties on the underlying native control. I have begun to love Effects as they make my code more readable. With Effects, I always know that there is a platform-specific implementation attached, while that is not obvious when using a custom renderer. Nowadays, I always try to implement an Effect before a Renderer.


Update: I updated the sample project to also respect changes in the ImageButton‘s source. I recently ran into the situation to change the Source (Play/Pause) via my ViewModel based on its state and realized that the effect needs to be reapplied in this case. Please be aware.


The basics

Effects work in a similar way to Renderers. You implement the definition in the Xamarin.Forms project, which attaches it to the control that needs the change. The PlatformEffect implementation needs to be exported to be compiled into the application. Like in a Renderer, the platform implementation also supports property changes. In this new series #XfEffects, I am going to show you some Effects that have been useful for me.

Effect declaration

Let’s turn over to this post’s Effect. We will change the TintColor of an ImageButton to our needs. Let’s start with creating the class for our Effect:

    public class ImageButtonTintEffect : RoutingEffect
    {
        public ImageButtonTintEffect() : base($"XfEffects.{nameof(ImageButtonTintEffect)}")
        {
        }
    }

All Xamarin.Forms Effect implementations need to derive from the RoutingEffect class and pass the Effect‘s name to the base class’ constructor. That’s pretty much everything we have to do for the Effect class itself.

Effect extension for passing parameters

The easiest way for us to get our desired TintColor to the platform implementation is an attached BindableProperty. To be able to attach the BindableProperty, we need a static class that provides the container for the attached property:

public static class ImageButtonTintEffectExtensions
{
    public static readonly BindableProperty TintColorProperty = BindableProperty.CreateAttached("TintColor", typeof(Color), typeof(ImageButtonTintEffectExtensions), default, propertyChanged: OnTintColorPropertyChanged);
    public static Color GetTintColor(BindableObject bindable)
    {
        return (Color)bindable.GetValue(TintColorProperty);
    }
    public static void SetTintColor(BindableObject bindable, Color value)
    {
        bindable.SetValue(TintColorProperty, value);
    }
    private static void OnTintColorPropertyChanged(BindableObject bindable, object oldValue, object newValue)
    {
    }
}

Of course, we want to set the TintColor property as Xamarin.Forms.Color as this will make it pretty easy to control the color from a Style or even a ViewModel.

We want our effect to only being invoked if we change the default TintColor value. This makes sure we are running only code that is necessary for our application:

private static void OnTintColorPropertyChanged(BindableObject bindable, object oldValue, object newValue)
{
    if (bindable is ImageButton current)
    {
        if ((Color)newValue != default)
        {
            if (!current.Effects.Any(e => e is ImageButtonTintEffect))
                current.Effects.Add(Effect.Resolve(nameof(ImageButtonTintEffect)));
        }
        else
        {
            if (current.Effects.Any(e => e is ImageButtonTintEffect))
            {
                var existingEffect = current.Effects.FirstOrDefault(e => e is ImageButtonTintEffect);
                current.Effects.Remove(existingEffect);
            }
        }
    }
}

Last but not least in our Xamarin.Forms application, we want to use the new Effect. This is pretty easy:

<!--import the namespace-->
xmlns:effects="clr-namespace:XfEffects.Effects"
<!--then use it like this-->
<ImageButton Margin="12,0,12,12"
    effects:ImageButtonTintEffectExtensions.TintColor="Red"
    BackgroundColor="Transparent"
    HeightRequest="48"
    HorizontalOptions="CenterAndExpand"
    Source="ic_refresh_white_24dp.png"
    WidthRequest="48">
    <ImageButton.Effects>
        <effects:ImageButtonTintEffect />
    </ImageButton.Effects>
</ImageButton>

We are using the attached property we created above to provide the TintColor to the ImageButtonTintEffect, which we need to add to the ImageButton‘s Effects List.

Android implementation

Let’s have a look at the Android-specific implementation. First, let’s add the platform class and decorate it with the proper attributes to export our effect:

using Android.Content.Res;
using Android.Graphics;
using System;
using System.ComponentModel;
using Xamarin.Forms;
using Xamarin.Forms.Platform.Android;
using AWImageButton = Android.Support.V7.Widget.AppCompatImageButton;
[assembly: ResolutionGroupName("XfEffects")]
[assembly: ExportEffect(typeof(XfEffects.Droid.Effects.ImageButtonTintEffect), nameof(XfEffects.Effects.ImageButtonTintEffect))]
namespace XfEffects.Droid.Effects
{
    public class ImageButtonTintEffect : PlatformEffect
    {
        protected override void OnAttached()
        {
            
        }
        protected override void OnDetached()
        {
        }
        protected override void OnElementPropertyChanged(PropertyChangedEventArgs args)
        {
        }
    }
}

Remember: the ResolutionGroupName needs to be defined just once per app and should not change. Similar to a custom Renderer, we also need to export the definition of the platform implementation and the Forms implementation to make our Effect working.

Android controls like buttons have different states. Properties on Android controls like the color can be set based on their State attribute. Xamarin.Android implements these states in the Android.Resource.Attribute class. We define our ImageButton‘s states like this:

static readonly int[][] _colorStates =
{
    new[] { global::Android.Resource.Attribute.StateEnabled },
    new[] { -global::Android.Resource.Attribute.StateEnabled }, //disabled state
    new[] { global::Android.Resource.Attribute.StatePressed } //pressed state
};

Good to know: certain states like ‘disabled‘ are created by just adding a ‘-‘ in front of the matching state defined in the OS states list (negating it). We need this declaration to create our own ColorStateList, which will override the color of the ImageButton‘s image. Add this method to the class created above:

private void UpdateTintColor()
{
    try
    {
        if (this.Control is AWImageButton imageButton)
        {
            var androidColor = XfEffects.Effects.ImageButtonTintEffectExtensions.GetTintColor(this.Element).ToAndroid();
            var disabledColor = androidColor;
            disabledColor.A = 0x1C; //140
            var pressedColor = androidColor;
            pressedColor.A = 0x24; //180
            imageButton.ImageTintList = new ColorStateList(_colorStates, new[] { androidColor.ToArgb(), disabledColor.ToArgb(), pressedColor.ToArgb() });
            imageButton.ImageTintMode = PorterDuff.Mode.SrcIn;
        }
    }
    catch (Exception ex)
    {
        System.Diagnostics.Debug.WriteLine(
            $"An error occurred when setting the {typeof(XfEffects.Effects.ImageButtonTintEffect)} effect: {ex.Message}\n{ex.StackTrace}");
    }
}

This code works above the Android SDK 23, as only then the ability to modify the A-channel of the defined color was added. The Xamarin.Forms ImageButton translates into a AppCompatImageButton on Android. The AppCompatImageButton has the ImageTintList property. This property is of type ColorStatesList, which uses the states we defined earlier and the matching colors for those states.

Last but not least, we need to set the composition mode. If you want to get a deeper understanding of that, a great starting point is this StackOverflow question. To make things not too complicated, we are infusing the color into the image. The final step is to call the method in the OnAttached override as well as in the OnElementPropertyChanged override.

The result based on the sample I created looks like this:

iOS implementation

Of course, also on iOS, we have to attribute the class, similar to the Android version:

using System;
using System.ComponentModel;
using UIKit;
using Xamarin.Forms;
using Xamarin.Forms.Platform.iOS;
[assembly: ResolutionGroupName("XfEffects")]
[assembly: ExportEffect(typeof(XfEffects.iOS.Effects.ImageButtonTintEffect), nameof(XfEffects.Effects.ImageButtonTintEffect))]
namespace XfEffects.iOS.Effects
{
    public class ImageButtonTintEffect : PlatformEffect
    {
        protected override void OnAttached()
        {
        }
        protected override void OnDetached()
        {
        }
        protected override void OnElementPropertyChanged(PropertyChangedEventArgs args)
        {
        }
    }
}

The underlying control of the Xamarin.Forms ImageButton is a default UIButton on iOS. The UIButton control has an UIImageView, which can be changed with the SetImage method. Based on that knowledge, we are going to implement the UpdateTintColor method:

private void UpdateTintColor()
{
    try
    {
        if (this.Control is UIButton imagebutton)
        {
            if (imagebutton.ImageView?.Image != null)
            {
                var templatedImg = imagebutton.CurrentImage.ImageWithRenderingMode(UIImageRenderingMode.AlwaysTemplate);
                //clear the image on the button
                imagebutton.SetImage(null, UIControlState.Normal);
                imagebutton.ImageView.TintColor = XfEffects.Effects.ImageButtonTintEffectExtensions.GetTintColor(this.Element).ToUIColor();
                imagebutton.TintColor = XfEffects.Effects.ImageButtonTintEffectExtensions.GetTintColor(this.Element).ToUIColor();
                imagebutton.SetImage(templatedImg, UIControlState.Normal);
            }
        }
    }
    catch (Exception ex)
    {
        System.Diagnostics.Debug.WriteLine($"An error occurred when setting the {typeof(ImageButtonTintEffect)} effect: {ex.Message}\n{ex.StackTrace}");
    }
}

Let’s review these code lines. The first step is to extract the already attached Image as a templated Image from the UIButton‘s UIImageView. The second step is to clear the Image from the Button, using the SetImage method and passing null as UIImage parameter. I tried to leave out this step, but it does not work if you do so.

The next step changes the TintColor for the UIButton‘s UIImageView as well as for the button itself. I was only able to get the TintColor changed once I changed both TintColor properties.

The final step is to re-attach the extracted image from step one to the UIImageButton‘s UIImageView, using once again the SetImage method – but this time, we are passing the extracted image as UIImage parameter.

Of course, also here we need to call this method in the OnAttached override as well as in OnElementPropertyChanged.

The result should look similar to this one:

Conclusion

It is pretty easy to implement extended functionality with a Xamarin.Forms Effect. The process is similar to the one of creating a custom renderer. By using attached properties you can fine-tune the usage of this code and also pass property values to the platform implementations.

Please make sure to follow along for the other Effects I will post as part of this new series. I also created a sample app to demonstrate the Effects (find it on Github). As always, I hope this post will be helpful for some of you.

Until the next post, happy coding, everyone!
Posted by msicc in Android, Dev Stories, iOS, Xamarin, 0 comments