WinRT and application view wide Pointer event monitoring #wpdev #windev

It seems like every year I go through same technical difficulties. Last year I blogged about WinRT and pointer manipulation #win8dev #winrtdev

One of the underlying classes detects manipulation by user by means of Touch class. This class never made it to WinRT XAML and I have been messing around for a few days now… until something finally sunk in..

CoreWindow exposes following events

PointerEntered Occurs when a pointer moves into the bounding box of the Windows Store app.
PointerExited Occurs when the pointer moves outside the bounding box of the Windows Store app.
PointerMoved Occurs when a pointer moves within the bounding box of the Windows Store app.
PointerPressed Occurs when a mouse button is clicked, or a touch or pen contact is detected, within the bounding rectangle of the Windows Store app.
PointerReleased Occurs when a mouse button is released, or a touch or pen contact is lifted, within the bounding rectangle of the Windows Store app.

These events work exactly like Touch.TouchFrameReported and get bubbled up no matter which UIElement is tapped.

Time to continue porting

WindowsPhone 8.1, NavigationBar and Silverlight apps #wpdev

With Windows Phone 8.1 Microsoft drastically changed on of the basic chassis requirements in Windows Phone since day 0. The hardware buttons at the bottom of the phone. There were many reasons and one of them was the ability to use same chassis for both Android and Windows Phone devices. This many cited was very important smaller OEMs that recently signed up to Windows Phone ecosystem.

WP7 resolutions:

  • 800 x 480

WP8 resolutions (and scaling factors as reported  by Silverlight apps):

  • 800 x 480     (1.0)
  • 1280 x 720   (1.5)
  • 1280 x 768   (1.6)
  • 1920 x 1080 (2.25)

For Silverlight Windows Phone app (most of the apps that is) the resolution defaults to 800 x 480. This fits nicely with Scaling factor 1.6 which was used by many early Nokia devices. The scaling factor 1.5 however gave 853 x 480. This primarily affected HTC 8 S/X devices and Samsung ATIV S devices only. Later when Lumia 1520 came out, it uses 1020 and that used 2.25 and that also had 853 x 480 res.

What this meant was that WP7 and any WP8 apps that fix their resolution would leave 53 px gap. With my 1Shot camera app, I have a complicated control structure and I ended up fixing many heights and widths.. Not only that I also move certain objects on OrientationChanged using specific values of TranslateX and TranslateY. Last week after a user complaint, I added code to detect 720p and 1080p resolutions and to stretch those by additional 53px. However I never tested this on newer 720p devices that no longer have hardware buttons.

What Windows Phone Dev team did was added another system control like ApplicationBar called NavigationBar. Unlike ApplicationBar, NavigationBar isn’t accessible from within application. Playing with emulator set to show NavigationBar did not show any change in size reported by

double width = Application.Current.Host.Content.ActualWidth;
double height = Application.Current.Host.Content.ActualHeight;

After much complaining around twitter I came across windows phone 8.1 hide NavigationBar post on MSDN Forums. One of the suggestions was using PhoneApplicationFrame.FullScreen. Setting it to true or false made no difference.. none at all. After another round of complains, I came across this post Layout and the Windows Phone navigation bar.. it suggest that SizeChanged event should take care of it and ActualWidth and ActualHeight properties will sort it out.

Being not particularly clever, I thought that Page’s SizeChanged would be fired.. sure it was firing but ActualWidth and ActualHeight were 0.0. Width and Height were NaN so that was useless. Thinking that I could do better than that, I tried using SizeChanged exposed by Application.Current.Host.Content.. nope no difference.. it wasn’t a particularly bright idea.. Yesterday I spent time deliberating whether I should use hardcoded lists to indicate NavigationBar availability or not.. I already do so for high resolution sensors.. nah too much hassle

So today I said.. surely there must be something… on SizeChanged event handler of page, I scanned App.RootFrame.. I found something.. a private field called _visibleRegion correctly showed that 53 px at the botton were not available on 720p emulator when NavigationBar was visible. I think I am clever than most…

private Thickness GetMargin()
{
    Thickness t = new Thickness();

    var field = App.RootFrame.GetType().GetField("_visibleRegion", BindingFlags.Instance | BindingFlags.NonPublic);

    if (field != null)
    {
        object visibleRegion = field.GetValue(App.RootFrame);

        if (visibleRegion != null && visibleRegion is Thickness)
        {
            t = (Thickness)visibleRegion;
        }
    }

    return t;
}

This would get me the right margin and I could just just correct a few bits and bingo.. hahah yah right.. FieldAccessExpcetion eat that you clever git.. Apps don’t run under full trust on phone and I couldn’t reflect to this level. Sigh.. This was hard.. why had Windows Phone team gone way out of their way to piss developers off ?? Having almost lost all hope, I thought, let me try subscribe to Page’s LayoutRoot grid’s SizeChanged event… had a look there.. event handler exposed NewValue and they showed the correct values!!! yay

double gridWidth = e.NewSize.Width > e.NewSize.Height ? e.NewSize.Width : e.NewSize.Height;
double gridHeight = e.NewSize.Width > e.NewSize.Height ? e.NewSize.Height : e.NewSize.Width;

double w = Application.Current.Host.Content.ActualWidth;
double h = Application.Current.Host.Content.ActualHeight;
double hostWidth = w > h ? w : h;
double hostHeight = w > h ? h : w;

bool hasSoftButtons = hostWidth > gridWidth;

Just for your sanity.. I always want Landscape mode so width is always greater than height based on supported resolutions.. so there you have it.. subscribe to LayoutRoot’s SizeChanged and get the right size.. OnResize, the dimensions reflect 800 x 480 resolution and not 853 x 480.

Finally!

WinRT License API, Silverlight 8.1 apps and sideload detection #wpdev

Since early 2011, I have been using the method described here.
Recently I have been upgrading all my apps to target Windows Phone 8.1 API. Most of previous API is still accessible in WP 8.1 SL8.1 apps. However a great deal more is available under WinRT API.

With my Alarm Clock app, I have started using WinRT API to pin and update Tiles. I have used WinRT API for background tasks as I can run them more frequently and I can detect Timezone change. I have added IAP to the app.
The app however is available in 2 flavours Free (with ads and restrictions) and Paid. So far I have only been updating the paid app. Of course I use sideload detection for Paid app however while considering update of Free version, I decided to use CurrenApp to execute License check and potentially offer to convert Free users in-place as opposed to getting them to buy another app!!

In good old SL API, I would have used

var license = new Microsoft.Phone.Marketplace.LicenseInformation();
IsTrial = license.IsTrial();

With WinRT that becomes

IsTrial = CurrentApp.LicenseInformation.IsTrial;

While developing and deploying if you observe the behaviour, the IsTrial never returns false. You cannot test Trial scenario without explicit pre-processor directives!
Of course CurrentAppSimulator is only available for WinRT app and not SL8.1 apps. The same behaviour would be exposed by the sideloaded apps. Not downloaded from store equates to full version. Like in good old days, WMAppPRHeader.xml is still attached after certification. So lets modify the code

try
{
   XDocument doc = XDocument.Load("WMAppPRHeader.xml");
 
   IsTrial = CurrentApp.LicenseInformation.IsTrial;
}
catch
{
   IsTrial = true;
}

You treat all apps without WMAppPRHeader.xml as trial. Job done!
Note that IAP are not affected. If you had TestIAP, it is not active by default.

CurrentApp.LicenseInformation.ProductLicenses["TestIAP"].IsActive;

Happy coding!

Image scaling related crashes on 1080p devices #wpdev

Back in March one of the 1Shot camera app users emailed me stating that the app crashes when using he uses digital zoom on Lumia 1520. I couldn’t reproduce the issue on my 1020. I then borrowed a friend’s 1520 and hmm sure I could reproduce it but I couldn’t catch the exception.

So we left it at that.. fast forward on Friday I get a mail back from chap in question and he states that he’s running latest greatest version of preview and the freaky crash is still happening. Luckily I have the 1520 at my desk and decided to put it to good use over the weekend.

So I put VS configuration in Debug mode and I set VS to catch all thrown exceptions. Run app and swipe left right and top bottom and every so often the app crashes.. no exception being caught.. the app just disappears and we are back to home screen. The only clue was the output of debug mode.. The program ‘[4880] AgHost.exe’ has exited with code -1073741819 (0xc0000005) ‘Access violation’.

The sad reality, the debug output is littered with messages starting ‘AgHost.exe’ (CoreCLR: Silverlight AppDomain): A quick check on the net showed that AgHost.exe is used to manage access through identity verification.. great. The error code wasn’t particularly useful either.

In this particular case, the error only occurred when zooming in. My implementation of digital zoom was scaling the preview image for the live view. The only action associated with crash was zooming in and out. As it happens, I detect manipulation delta and compute scale..

private void LivePreviewTapTarget_ManipulationDelta(object sender, System.Windows.Input.ManipulationDeltaEventArgs e)
{
    if (!this.viewModel.CaptureInitiated && AppSettings.InternalAppSettings.DragZoom)
    {
        SetDragToScale(e.CumulativeManipulation.Translation);
        e.Handled = true;
    }
}

void SetDragToScale(System.Windows.Point pScale)
{
    double x = Math.Abs(pScale.X);
    double y = Math.Abs(pScale.Y);

    double manipulation = x > y ? pScale.X : pScale.Y;

    bool zoomOut = manipulation >= 0 ? false : true;

    double scale = (this.viewModel.HasHighResolutionCamera ? 0.01 : 0.005) * Math.Sqrt(Math.Pow(x, 2) + Math.Pow(y, 2));

    if (zoomOut)
        this.viewModel.ViewScale = initialScale + scale;
    else
        this.viewModel.ViewScale = initialScale - scale;

    System.Diagnostics.Debug.WriteLine("zoomout {0}, scale {1}", zoomOut, scale);
}

My initial thought was that it being a newer SoC, the Xaml Renderer was somehow being proactive somewhere and getting itself in a mess.

Test 1: show less frames and show larger jumps so there is less rendering. For High Res camera I used constant of 0.01.. increase it to 0.05.

The resultant build crashed very very fast. hardly took a swipe or two to head back to home screen.

So I wanted to try the reverse but by laptop decided to shutdown and I was too knackered to continue yesterday.
Earlier today I had a chat with Bill Reiss.. he mentioned that he had some problems with 1520 too and that it had something to do with how large the image becomes on 1080p screen.

So I decided to continue where I left yesterday but with a reduction only for 1080p devices.
Test 2: reduce the multiplication constant to 0.0015 if a 1080p device is detected.
Result: Significant reduction in crashes.. I can still get it to crash but not that often. Here is the changed version

int DeviceScaleFactor = -1;
private void GetScaleFactor()
{
    object physicalScreenResolutionObject;

    if (DeviceExtendedProperties.TryGetValue("PhysicalScreenResolution", out physicalScreenResolutionObject))
    {
        var physicalScreenResolution = (Size)physicalScreenResolutionObject;

        DeviceScaleFactor = (int)(physicalScreenResolution.Width / 4.8);
    }
    else
    {
        DeviceScaleFactor = Application.Current.Host.Content.ScaleFactor;
    }
}

void SetDragToScale(System.Windows.Point pScale)
{
    double x = Math.Abs(pScale.X);
    double y = Math.Abs(pScale.Y);

    double manipulation = x > y ? pScale.X : pScale.Y;

    bool zoomOut = manipulation >= 0 ? false : true;

    if (this.DeviceScaleFactor == -1)
    {
        this.GetScaleFactor();
    }

    double scale = (this.viewModel.HasHighResolutionCamera ? 0.01 : 0.005) * Math.Sqrt(Math.Pow(x, 2) + Math.Pow(y, 2));

    if (DeviceScaleFactor == 225)
    {
        scale = 0.0015 * Math.Sqrt(Math.Pow(x, 2) + Math.Pow(y, 2));
    }

    if (zoomOut)
        this.viewModel.ViewScale = initialScale + scale;
    else
        this.viewModel.ViewScale = initialScale - scale;

    System.Diagnostics.Debug.WriteLine("zoomout {0}, scale {1}", zoomOut, scale);
}

if the Device scale factor is 225, its a 1080p device and one should exercise caution.

Where did the SystemTray vanish with #WinPRT #wpdev?

With Silverlight WP8 API, SystemTray is exposed in PhoneApplicationPage through Microsoft.Phone.Shell.

<phone:PhoneApplicationPage
    shell:SystemTray.IsVisible="True" 
    shell:SystemTray.BackgroundColor="AliceBlue"
    shell:SystemTray.ForegroundColor="Black"
    shell:SystemTray.Opacity="0.75">
    <shell:SystemTray.ProgressIndicator>
        <shell:ProgressIndicator />
    </shell:SystemTray.ProgressIndicator>
</phone:PhoneApplicationPage>
SystemTray.BackgroundColor = Colors.Cyan;
SystemTray.ForegroundColor = Colors.White;
SystemTray.IsVisible = true;
SystemTray.Opacity = 0.8;
SystemTray.ProgressIndicator = new ProgressIndicator() { IsIndeterminate = true, IsVisible = true };

it was very easily done through xaml itself. You could do it from codebehind by accessing static properties of SystemTray class.
However with WinPRT, the Page class does not expose SystemTray and now it lives as Windows.UI.ViewManagement.StatusBar.
No xaml declarative business here.

var statusBar = StatusBar.GetForCurrentView();
statusBar.BackgroundColor = Colors.Black;
statusBar.BackgroundOpacity = 0.8;
statusBar.ForegroundColor = Colors.White;
statusBar.ProgressIndicator.Text = "loading";
await statusBar.ShowAsync();

If I find anything declarative, I will post those at a later date.

Photo / Camera API – 1 #wpdev

My memory of Windows Phone NoDo API set is pretty bleak. I distinctly remember presence of CameraCaptureTask and PhotoChooserTask. I think I used PhotoChooserTask with my first app Scribble. So what’s this post about ? well Photo / Camera options available on Windows Phone 8.

Since I started with the two tasks, they still exist in Windows Phone 8.

CameraCaptureTask: this fires the stock camera app – ability to navigate to camera roll and use other lens is blocked but otherwise it is the stock windows phone app. Tap to focus / capture or use dedicated shutter button. On capture you are presented with a preview and you have the option to accept or retake the capture.

wp_ss_20140130_0005 wp_ss_20140130_0006 wp_ss_20140130_0007

PhotoChooserTask: this is a different beast. This opens photos hub and presents user with list of albums allowing user to select a photo. The Task has a parameter that toggles visibility of camera icon – ability to not just select a picture but also initiate CameraCaptureTask from within.

While both of these are very basic, let me point out that Instagram BETA still uses a layout very similar to Photochooser task and obviously uses Camera Capture Task.

Its very likely that Instagram could get away with murder. Most of us – 3rd party developers however do not have such luxury. We would be termed as lazy if not worse. And you might not want the stock camera UI. So what can you do ? I am still in old API mode so let me mention an API that was part of Mango – I don’t know if NoDo had it or not – I only used it after Mango public release in my Cool Camera app.

PhotoCamera Class: this class provides raw access to camera and exposes a few options like resolution, flash and focus. Apps render live view by means of using a video brush whose source is set to an instance of photo camera class. How to create a base camera app for Windows Phone provides a great sample to kick start camera app development using this class

PhotoCamera

For a camera app, call CaptureImage. Others might not need a full blown capture – something like QR scanning app would just use focus and can do with GetPreviewBufferArgb32.

Whilst this is a good start, the API is not extensive enough for anything beyond basic. In the next post, I will discuss PhotoCaptureDevice class. This class though simple, exposes enough API to create a full blown camera app allowing you to control every single aspect of camera. If you hardware supports it, you should be able to utilise it

Immutable Collections Performance #dotnet

Yesterday ayende blogged and tweeted about serious issues with Immutable collections performance. whilst I didn’t have time to really look into it, I did retweet the link and got involved in a discussion with ex-softie Brandon Paddock.

When I woke up, I saw some more discussion and thought, let me try it myself. Now we all know every chef has their own way and some ways are better than others. FYI, I never used ImmutableCollections before this date and it took me a few mins to get the hang of it.

so what did I do ? I kept it simple. Here’s the source and the results. As you can see creating immutable collection is an added option and that takes time but its not very high

class Program
{
    static int MaxIterations = 10 * 1000 * 1000;
    static int maxRuns = 5;
    static void Main(string[] args)
    {
        for (int i = 0; i < maxRuns; i++ )
        {
            GC.Collect();

            TestGenericList();

            GC.Collect();

            TestImmutableList();
        }
    }

    public static void PreRunWarmup()
    {
        List<int> genericDataList = new List<int>();

        for (int i = 0; i < MaxIterations; i++)
        {
            genericDataList.Add(i);
        }

        genericDataList.Clear();
        genericDataList = null;
    }

    public static void TestGenericList()
    {
        List<int> genericDataList = new List<int>();
                        
        var sw = Stopwatch.StartNew();
        for (int i = 0; i < MaxIterations; i++)
        {
            genericDataList.Add(i);
        }

        sw.Stop();
        Console.WriteLine("Generic List took {0} ms", (double)sw.ElapsedTicks * (double)1000 / (double)Stopwatch.Frequency);

        genericDataList.Clear();
        genericDataList = null;
    }

    public static void TestImmutableList()
    {
        List<int> genericDataList = new List<int>();

        var sw = Stopwatch.StartNew();
        for (int i = 0; i < MaxIterations; i++)
        {
            genericDataList.Add(i);
        }

        ImmutableList<List<int>> immutableList = ImmutableList.Create(genericDataList);

        sw.Stop();
        Console.WriteLine("Immutable List took {0} ms", (double)sw.ElapsedTicks * (double)1000 / (double)Stopwatch.Frequency);

        genericDataList.Clear();
        genericDataList = null;
    }
}

and the results are

Generic List took 85.5857446744927 ms
Immutable List took 88.5216447923091 ms
Generic List took 89.7324092764301 ms
Immutable List took 88.4739507537922 ms
Generic List took 88.1856734956681 ms
Immutable List took 90.4921934216639 ms
Generic List took 85.6533615392255 ms
Immutable List took 88.9282515637155 ms
Generic List took 87.9722577663554 ms
Immutable List took 84.6940472708294 ms
Press any key to continue . . .

Note: I added a pre-test warm up as the first iteration was taking unusually long.
So for a few more milli-seconds per million entries what do we get ?

If you have a threaded environment and you pass collection and iterate them, you should have come across the infamous InvalidOperationException stating something like “Collection was modified; enumeration operation may not execute.” At least I have resorted to using lock around collection enumeration code in past.
With immutable collection, you never have to worry about the output received from Immutable.CreateList. No more unnecessary locks around iterations.

Note: I used GC.Collect excessively and even resorted to cleaning and setting collections to null so as to not impact the next call. Just pure perf test.