I understand that ViewModel shouldn't have any knowledge of View, but how can I call MediaElement.Play() method from ViewModel, other than having a reference to View (or directly to MediaElement) in ViewModel?
Other (linked) question: how can I manage View's controls visibility from ViewModel without violating MVVM pattern?
1) Do not call Play() from the view model. Raise an event in the view model instead (for instance PlayRequested) and listen to this event in the view:
view model:
public event EventHandler PlayRequested;
...
if (this.PlayRequested != null)
{
this.PlayRequested(this, EventArgs.Empty);
}
view:
ViewModel vm = new ViewModel();
this.DataContext = vm;
vm.PlayRequested += (sender, e) =>
{
this.myMediaElement.Play();
};
2) You can expose in the view model a public boolean property, and bind the Visibility property of your controls to this property. As Visibility is of type Visibility and not bool, you'll have to use a converter.
You can find a basic implementation of such a converter here.
This related question might help you too.
For all the late-comers,
There are many ways to achieve the same result and it really depends on how you would like to implement yours, as long as your code is not difficult to maintain, I do believe it's ok to break the MVVM pattern under certain cases.
But having said that, I also believe there is always way to do this within the pattern, and the following is one of them just in case if anyone would like to know what other alternatives are available.
The Tasks:
we don't want to have direct reference from the ViewModel to any UI elements, i.e. the the MediaElement and the View itself.
we want to use Command to do the magic here
The Solution:
In short, we are going to introduce an interface between the View and the ViewModel to break the dependecy, and the View will be implementing the interface and be responsible for the direct controlling of the MediaElement while leaving the ViewModel talking only to the interface, which can be swapped with other implementation for testing purposes if needed, and here comes the long version:
Introduce an interface called IMediaService as below:
public interface IMediaService
{
void Play();
void Pause();
void Stop();
void Rewind();
void FastForward();
}
Implement the IMediaService in the View:
public partial class DemoView : UserControl, IMediaService
{
public DemoView()
{
InitializeComponent();
}
void IMediaService.FastForward()
{
this.MediaPlayer.Position += TimeSpan.FromSeconds(10);
}
void IMediaService.Pause()
{
this.MediaPlayer.Pause();
}
void IMediaService.Play()
{
this.MediaPlayer.Play();
}
void IMediaService.Rewind()
{
this.MediaPlayer.Position -= TimeSpan.FromSeconds(10);
}
void IMediaService.Stop()
{
this.MediaPlayer.Stop();
}
}
we then do few things in the DemoView.XAML:
Give the MediaElement a name so the code behind can access it like above:
<MediaElement Source="{Binding CurrentMedia}" x:Name="MediaPlayer"/>
Give the view a name so we can pass it as a parameter, and
import the interactivity namespace for later use (some default namespaces are omitted for simplicity reason):
<UserControl x:Class="Test.DemoView"
xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
xmlns:ia="http://schemas.microsoft.com/expression/2010/interactivity"
x:Name="MediaService">
Hookup the Loaded event through Trigger to pass the view itself to the view model through a Command
<ia:Interaction.Triggers>
<ia:EventTrigger EventName="Loaded">
<ia:InvokeCommandAction Command="{Binding LoadedCommand}" CommandParameter="{Binding ElementName=MediaService}"></ia:InvokeCommandAction>
</ia:EventTrigger>
</ia:Interaction.Triggers>
last but not least, we need to hookup the media controls through Commands:
<Button Command="{Binding PlayCommand}" Content="Play"></Button>
<Button Command="{Binding PauseCommand}" Content="Pause"></Button>
<Button Command="{Binding StopCommand}" Content="Stop"></Button>
<Button Command="{Binding RewindCommand}" Content="Rewind"></Button>
<Button Command="{Binding FastForwardCommand}" Content="FastForward"></Button>
We now can catch everything in the ViewModel (I'm using prism's DelegateCommand here):
public class AboutUsViewModel : SkinTalkViewModelBase, IConfirmNavigationRequest
{
public IMediaService {get; private set;}
private DelegateCommand<IMediaService> loadedCommand;
public DelegateCommand<IMediaService> LoadedCommand
{
get
{
if (this.loadedCommand == null)
{
this.loadedCommand = new DelegateCommand<IMediaService>((mediaService) =>
{
this.MediaService = mediaService;
});
}
return loadedCommand;
}
}
private DelegateCommand playCommand;
public DelegateCommand PlayCommand
{
get
{
if (this.playCommand == null)
{
this.playCommand = new DelegateCommand(() =>
{
this.MediaService.Play();
});
}
return playCommand;
}
}
.
. // other commands are not listed, but you get the idea
.
}
Side note: I use Prism's Auto Wiring feature to link up the View and ViewModel. So at the View's code behind file there is no DataContext assignment code, and I prefer to keep it that way, and hence I chose to use purely Commands to achieve this result.
I use media element to play sounds in UI whenever an event occurs in the application. The view model handling this, was created with a Source property of type Uri (with notify property changed, but you already know you need that to notify UI).
All you have to do whenever source changes (and this is up to you), is to set the source property to null (this is why Source property should be Uri and not string, MediaElement will naturally throw exception, NotSupportedException I think), then set it to whatever URI you want.
Probably, the most important aspect of this tip is that you have to set MediaElement's property LoadedBehaviour to Play in XAML of your view. Hopefully no code behind is needed for what you want to achieve.
The trick is extremely simple so I won't post a complete example. The view model's play function should look like this:
private void PlaySomething(string fileUri)
{
if (string.IsNullOrWhiteSpace(fileUri))
return;
// HACK for MediaElement: to force it to play a new source, set source to null then put the real source URI.
this.Source = null;
this.Source = new Uri(fileUri);
}
Here is the Source property, nothing special about it:
#region Source property
/// <summary>
/// Stores Source value.
/// </summary>
private Uri _Source = null;
/// <summary>
/// Gets or sets file URI to play.
/// </summary>
public Uri Source
{
get { return this._Source; }
private set
{
if (this._Source != value)
{
this._Source = value;
this.RaisePropertyChanged("Source");
}
}
}
#endregion Source property
As for Visibility, and stuff like this, you can use converters (e.g. from bool to visibility, which you can find on CodePlex for WPF, SL, WP7,8) and bind your control's property to that of the view model's (e.g. IsVisible). This way, you control parts of you view's aspect. Or you can just have Visibility property typed System.Windows.Visibility on your view model (I don't see any pattern breach here). Really, it's not that uncommon.
Good luck,
Andrei
P.S. I have to mention that .NET 4.5 is the version where I tested this, but I think it should work on other versions as well.
My MVVM application contains two views:
AllStrategiesView
StrategyView
When user click certain strategy in AllStrategiesView StrategyView with this strategy is created. I use such code to notify application that StrategyView should be created:
.............
public void OpenStrategyView()
{
OnPropertyChanged("OpenStrategy");
}
.................
private void OnWorkspacePropertyChanged(object sender, PropertyChangedEventArgs e)
{
const string openStrategyString = "OpenStrategy";
if (e.PropertyName == openStrategyString)
{
AllStrategiesViewModel vm = (sender as AllStrategiesViewModel);
OpenStrategy(vm.SelectedStrategy);
}
}
However another part of the program shows error message because there are no such property "OpenStrategy":
/// <summary>
/// Warns the developer if this object does not have
/// a public property with the specified name. This
/// method does not exist in a Release build.
/// </summary>
[Conditional("DEBUG")]
[DebuggerStepThrough]
public void VerifyPropertyName(string propertyName)
{
// Verify that the property name matches a real,
// public, instance property on this object.
if (TypeDescriptor.GetProperties(this)[propertyName] == null)
{
string msg = "Invalid property name: " + propertyName;
if (this.ThrowOnInvalidPropertyName)
throw new Exception(msg);
else
Debug.Fail(msg);
}
The question is:
Is it right or wrong to use OnPropertyChanged to notify application that something need to be done? Should I rewrite my code to not to use OnPropertyChanged or should I disable VerifyPropertyName code?
This is bad practice. The PropertyChanged event on INotifyPropertyChanged should be used to notify subscribers that a property on the object instance has changed. This is typically used in WPF to notify the UI that it needs to update itself with the new property value.
In MVVM, you should use some kind of commanding or alternative viewmodel/view communication mechanism to invoke verbs (methods) on your view model from the view. The commanding provided by WPF has limitations, so I would recommend using an MVVM framework and the mechanisms that they provide.
Well it depends on what you want it to do. In your case it looks like you have a property "Workspace" which indicates which VM you should be looking at. This doesn't seem too bad of a usage IMHO.
If you were doing something completely unrelated to the property that was changed then it might work, but it's certainly not what I'd expect it to do (see Principle of Least Astonishment). OnPropertyChanged is intended to indicate that a property that has been bound to has changed and should be re-fetched.
You can of course just have another event on your ViewModel, like:
public event Action<String> OpenStrategy;
One more thing... This code is completely redundant:
const string openStrategyString = "OpenStrategy";
if (e.PropertyName == openStrategyString)
the following is exactly the same, from the compiler's perspective, and much more readable:
if (e.PropertyName == "OpenStrategy")
There's nothing wrong in asking your application to do something in the PropertyChanged event, however do not raise a PropertyChanged event just to ask the application to do something.
PropertyChanged is used to indicate that a property has changed, and should be used for that only.
Devdigital's answer gives a good example, that the UI uses the PropertyChange notification to know when it should update. Other objects can also subscribe to receive change notifications, and they should only be notified when a value changes, not when you want to run some application code.
Using your example, I would rewrite it like this:
public void OpenStrategyView()
{
OpenStrategy(this.SelectedStrategy);
}
private void OnWorkspacePropertyChanged(object sender, PropertyChangedEventArgs e)
{
if (e.PropertyName == "SelectedStrategy")
{
OpenStrategyView();
}
}
I have control in SL4. I want data validation on button click. Big problem is normally SL4 give validation using binding property.
like example given shown in this example
http://weblogs.asp.net/dwahlin/archive/2010/08/15/validating-data-in-silverlight-4-applications-idataerrorinfo.aspx
<TextBox Text="{Binding Name,Mode=TwoWay,ValidatesOnDataErrors=true}"
Height="23"
Width="120"
HorizontalAlignment="Left"
VerticalAlignment="Top" />
BUT I WANT TO SHOW ERROR MESSAGE LIKE THIS ....
using my own code like on button click i check
(textbox1.text == null ) then set this style of error to textbox1
One way of deferring validation is to set the property UpdateSourceTrigger=Explicit in the bindings. If you do this, the bindings won't update the source objects, and hence won't cause validation errors, until you explicitly tell the bindings to do so. When your button is clicked, you force an update on the bindings, using a line such as the following for each control:
someTextBox.GetBindingExpression(TextBox.TextProperty).UpdateSource();
You then have your property setters throwing exceptions for invalid data.
This approach can be a bit of a pain if there are quite a lot of controls to force binding updates on.
Also, forcing an update on the bindings has to be done in the code-behind of a control. If you're using a Command with the button as well then you might run in to an issue. Buttons can have both a Command and a Click event handler, and both will execute when the button is clicked on, but I don't know the order in which this happens or even if an order can be guaranteed. A quick experiment suggested that the event handler was executed before the command, but I don't know whether this is undefined behaviour. There is therefore the chance that the command will be fired before the bindings have been updated.
An approach to programmaticaly creating validation tooltips is to bind another property of the textbox and then deliberately cause an error with this binding.
'sapient' posted a complete solution, including code on the Silverlight forums (search for the post dated 07-08-2009 4:56 PM). In short, he/she creates a helper object with a property whose getter throws an exception, binds the Tag property of the textbox to this helper object and then forces an update on the binding.
'sapient's code was written before Silverlight 4 was released. We'll 'upgrade' his/her code to Silverlight 4. The class ControlValidationHelper becomes the following:
public class ControlValidationHelper : IDataErrorInfo
{
public string Message { get; set; }
public object ValidationError { get; set; }
public string Error
{
get { throw new NotImplementedException(); }
}
public string this[string columnName]
{
get { return Message; }
}
}
It's easy enough to knock up a quick demo application to try this out. I created the following three controls:
<TextBox x:Name="tbx" Text="{Binding Path=Text, ValidatesOnDataErrors=True, NotifyOnValidationError=True, Mode=TwoWay}" />
<Button Click="ForceError_Click">Force error</Button>
<Button Click="ClearError_Click">Clear error</Button>
The Text property and the event handlers for the two buttons live in the code-behind and are as follows:
public string Text { get; set; }
private void ForceError_Click(object sender, RoutedEventArgs e)
{
var helper = new ControlValidationHelper() { Message = "oh no!" };
tbx.SetBinding(Control.TagProperty, new Binding("ValidationError")
{
Mode = BindingMode.TwoWay,
NotifyOnValidationError = true,
ValidatesOnDataErrors = true,
UpdateSourceTrigger = UpdateSourceTrigger.Explicit,
Source = helper
});
tbx.GetBindingExpression(Control.TagProperty).UpdateSource();
}
private void ClearError_Click(object sender, RoutedEventArgs e)
{
BindingExpression b = tbx.GetBindingExpression(Control.TagProperty);
if (b != null)
{
((ControlValidationHelper)b.DataItem).Message = null;
b.UpdateSource();
}
}
The 'Force error' button should make a validation error appear on the textbox, and the 'Clear error' button should make it go away.
One potential downside of this approach occurs if you are using a ValidationSummary. The ValidationSummary will list all validation errors against ValidationError instead of against the name of each property.
Although my answer wasn't regarded as preferable, I'm still sure that the MVVM pattern is the best choice to perform validation.
In my code you should use the model validator from this post about validation and any mvvm framework, for example MVVM Light.
It is much easier to add validation rules using the view model and model validator classes:
public class PersonViewModel : ViewModelBase, INotifyDataErrorInfo
{
private ModelValidator _validator = new ModelValidator();
public PersonViewModel()
{
this._validator.AddValidationFor(() => this.Age)
.Must(() => this.Age > 0)
.Show("Age must be greater than zero");
}
}
And you can validate the model if and only if a user explicitly clicks a button:
#region INotifyDataErrorInfo
public IEnumerable GetErrors(string propertyName)
{
return this._validator.GetErrors(propertyName);
}
public bool HasErrors
{
get { return this._validator.ErrorMessages.Count > 0; }
}
public event EventHandler<DataErrorsChangedEventArgs> ErrorsChanged = delegate { };
protected void OnErrorsChanged(string propertyName)
{
ErrorsChanged(this, new DataErrorsChangedEventArgs(propertyName));
this.RaisePropertyChanged("HasErrors");
}
#endregion
public bool Validate()
{
var result = this._validator.ValidateAll();
this._validator.PropertyNames.ForEach(OnErrorsChanged);
return result;
}
As everyone can see, there is nothing difficult here, just 20-30 lines of code.
Moreover, the MVVM approach is much more flexible and you can reuse some common validation scenaries among several view models.
I've seen a library that allows me to do this inside my XAML, which sets the visibility of the control based on whether or not the user is in a role:
s:Authorization.RequiresRole="Admin"
Using that library with my database requires a bunch of coding that I can't really do right now. Ultimately here's what I want to know...
I have received the authenticated users role from my SPROC, and its currently stored in my App.xaml.cs as a property (not necessary for the final solution, just FYI for now). I want to create a property (dependency property? attached property?) that allows me to say something very similar to what the other library has: RequiresRole="Admin", which would collapse the visibility if the user is not in the Admin role. Can anyone point me in the right direction on this?
EDIT
After building the authorization class, I get the following error:
"The property 'RequiredRole' does not exist on the type 'HyperlinkButton' in the XML Namespace clr-namespace:TSMVVM.Authorization"
I'm trying to add this xaml:
<HyperlinkButton x:Name="lnkSiteParameterDefinitions"
Style="{StaticResource LinkStyle}"
Tag="SiteParameterDefinitions"
Content="Site Parameter Definitions"
Command="{Binding NavigateCommand}"
s:Authorization.RequiredRole="Admin"
CommandParameter="{Binding Tag, ElementName=lnkSiteParameterDefinitions}"/>
When I started typing the s:Authorization.RequiredRole="Admin", intellisense picked it up. I tried setting the typeof(string) and typeof(ownerclass) to HyperlinkButton to see if that would help, but it didn't. Any thoughts?
Attached property is the way to implement it. You should define a property like this:
public class Authorization
{
#region Attached DP registration
public static string GetRequiredRole(UIElement obj)
{
return (string)obj.GetValue(RequiredRoleProperty);
}
public static void SetRequiredRole(UIElement obj, string value)
{
obj.SetValue(RequiredRoleProperty, value);
}
#endregion
// Using a DependencyProperty as the backing store for RequiredRole. This enables animation, styling, binding, etc...
public static readonly DependencyProperty RequiredRoleProperty =
DependencyProperty.RegisterAttached("RequiredRole", typeof(string), typeof(Authorization), new PropertyMetadata(RequiredRole_Callback));
// This callback will be invoked when some control will receive a value for your 'RequiredRole' property
private static void RequiredRole_Callback(DependencyObject source, DependencyPropertyChangedEventArgs e)
{
var uiElement = (UIElement) source;
RecalculateControlVisibility(uiElement);
// also this class should subscribe somehow to role changes and update all control's visibility after role being changed
}
private static void RecalculateControlVisibility(UIElement control)
{
//Authorization.UserHasRole() - is your code to check roles
if (Authentication.UserHasRole(GetRequiredRole(control)))
control.Visibility = Visibility.Visible;
else
control.Visibility = Visibility.Collapsed;
}
}
PS: Have noticed too late that you were asking about Silverlight. Though I believe it works in the same way there, but I've tried it only on WPF.
I have run into an issue with WPF and Commands that are bound to a Button inside the DataTemplate of an ItemsControl. The scenario is quite straight forward. The ItemsControl is bound to a list of objects, and I want to be able to remove each object in the list by clicking a Button. The Button executes a Command, and the Command takes care of the deletion. The CommandParameter is bound to the Object I want to delete. That way I know what the user clicked. A user should only be able to delete their "own" objects - so I need to do some checks in the "CanExecute" call of the Command to verify that the user has the right permissions.
The problem is that the parameter passed to CanExecute is NULL the first time it's called - so I can't run the logic to enable/disable the command. However, if I make it allways enabled, and then click the button to execute the command, the CommandParameter is passed in correctly. So that means that the binding against the CommandParameter is working.
The XAML for the ItemsControl and the DataTemplate looks like this:
<ItemsControl
x:Name="commentsList"
ItemsSource="{Binding Path=SharedDataItemPM.Comments}"
Width="Auto" Height="Auto">
<ItemsControl.ItemTemplate>
<DataTemplate>
<StackPanel Orientation="Horizontal">
<Button
Content="Delete"
FontSize="10"
Command="{Binding Path=DataContext.DeleteCommentCommand, ElementName=commentsList}"
CommandParameter="{Binding}" />
</StackPanel>
</DataTemplate>
</ItemsControl.ItemTemplate>
</ItemsControl>
So as you can see I have a list of Comments objects. I want the CommandParameter of the DeleteCommentCommand to be bound to the Command object.
So I guess my question is: have anyone experienced this problem before? CanExecute gets called on my Command, but the parameter is always NULL the first time - why is that?
Update: I was able to narrow the problem down a little. I added an empty Debug ValueConverter so that I could output a message when the CommandParameter is data bound. Turns out the problem is that the CanExecute method is executed before the CommandParameter is bound to the button. I have tried to set the CommandParameter before the Command (like suggested) - but it still doesn't work. Any tips on how to control it.
Update2: Is there any way to detect when the binding is "done", so that I can force re-evaluation of the command? Also - is it a problem that I have multiple Buttons (one for each item in the ItemsControl) that bind to the same instance of a Command-object?
Update3: I have uploaded a reproduction of the bug to my SkyDrive: http://cid-1a08c11c407c0d8e.skydrive.live.com/self.aspx/Code%20samples/CommandParameterBinding.zip
I was having this same issue while trying to bind to a command on my view model.
I changed it to use a relative source binding rather than referring to the element by name and that did the trick. Parameter binding didn't change.
Old Code:
Command="{Binding DataContext.MyCommand, ElementName=myWindow}"
New Code:
Command="{Binding DataContext.MyCommand, RelativeSource={RelativeSource AncestorType=Views:MyView}}"
Update: I just came across this issue without using ElementName, I'm binding to a command on my view model and my data context of the button is my view model. In this case I had to simply move the CommandParameter attribute before the Command attribute in the Button declaration (in XAML).
CommandParameter="{Binding Groups}"
Command="{Binding StartCommand}"
I have found that the order in which I set Command and CommandParameter makes a difference. Setting the Command property causes CanExecute to be called immediately, so you want CommandParameter to already be set at that point.
I have found that switching the order of the properties in the XAML can actually have an effect, though I'm not confident that it will solve your problem. It's worth a try, though.
You seem to be suggesting that the button never becomes enabled, which is surprising, since I would expect the CommandParameter to be set shortly after the Command property in your example. Does calling CommandManager.InvalidateRequerySuggested() cause the button to become enabled?
I stumbled upon a similar problem and solved it using my trusty TriggerConverter.
public class TriggerConverter : IMultiValueConverter
{
#region IMultiValueConverter Members
public object Convert(object[] values, Type targetType, object parameter, System.Globalization.CultureInfo culture)
{
// First value is target value.
// All others are update triggers only.
if (values.Length < 1) return Binding.DoNothing;
return values[0];
}
public object[] ConvertBack(object value, Type[] targetTypes, object parameter, System.Globalization.CultureInfo culture)
{
throw new NotImplementedException();
}
#endregion
}
This value converter takes any number of parameters and passes the first of them back as the converted value. When used in a MultiBinding in your case it looks like the following.
<ItemsControl
x:Name="commentsList"
ItemsSource="{Binding Path=SharedDataItemPM.Comments}"
Width="Auto" Height="Auto">
<ItemsControl.ItemTemplate>
<DataTemplate>
<StackPanel Orientation="Horizontal">
<Button
Content="Delete"
FontSize="10"
CommandParameter="{Binding}">
<Button.Command>
<MultiBinding Converter="{StaticResource TriggerConverter}">
<Binding Path="DataContext.DeleteCommentCommand"
ElementName="commentsList" />
<Binding />
</MultiBinding>
</Button.Command>
</Button>
</StackPanel>
</DataTemplate>
</ItemsControl.ItemTemplate>
</ItemsControl>
You will have to add TriggerConverter as a resource somewhere for this to work. Now the Command property is set not before the value for the CommandParameter has become available. You could even bind to RelativeSource.Self and CommandParameter instead of . to achieve the same effect.
I've come up with another option to work around this issue that I wanted to share. Because the CanExecute method of the command gets executed before the CommandParameter property is set, I created a helper class with an attached property that forces the CanExecute method to be called again when the binding changes.
public static class ButtonHelper
{
public static DependencyProperty CommandParameterProperty = DependencyProperty.RegisterAttached(
"CommandParameter",
typeof(object),
typeof(ButtonHelper),
new PropertyMetadata(CommandParameter_Changed));
private static void CommandParameter_Changed(DependencyObject d, DependencyPropertyChangedEventArgs e)
{
var target = d as ButtonBase;
if (target == null)
return;
target.CommandParameter = e.NewValue;
var temp = target.Command;
// Have to set it to null first or CanExecute won't be called.
target.Command = null;
target.Command = temp;
}
public static object GetCommandParameter(ButtonBase target)
{
return target.GetValue(CommandParameterProperty);
}
public static void SetCommandParameter(ButtonBase target, object value)
{
target.SetValue(CommandParameterProperty, value);
}
}
And then on the button you want to bind a command parameter to...
<Button
Content="Press Me"
Command="{Binding}"
helpers:ButtonHelper.CommandParameter="{Binding MyParameter}" />
I hope this perhaps helps someone else with the issue.
I'll add what worked for me for a DataGridTemplateColumn with a button.
Change the binding from:
CommandParameter="{Binding .}"
to
CommandParameter="{Binding DataContext, RelativeSource={RelativeSource Self}}"
Not sure why it works, but it did for me.
I recently came across the same problem (for me it was for the menu items in a context menu), nad while it may not be a suitable solution for every situation, I found a different (and a lot shorter!) way of solving this problem:
<MenuItem Header="Open file" Command="{Binding Tag.CommandOpenFile, IsAsync=True, RelativeSource={RelativeSource AncestorType={x:Type ContextMenu}}}" CommandParameter="{Binding Name}" />
Ignoring the Tag-based workaround for the special case of context menu, the key here is to bind the CommandParameter regularly, but bind the Command with the additional IsAsync=True. This will delay the binding of the actual command (and therefore its CanExecute call) a bit, so the parameter will already be available. This means, though, that for a brief moment, the enabled-state might be wrong, but for my case, that was perfectly acceptable.
You may be able to use my CommandParameterBehavior that I posted to the Prism forums yesterday. It adds the missing behaviour where a change to the CommandParameter cause the Command to be re-queried.
There's some complexity here caused by my attempts to avoid the memory leak caused if you call PropertyDescriptor.AddValueChanged without later calling PropertyDescriptor.RemoveValueChanged. I try and fix that by unregistering the handler when the ekement is unloaded.
You'll probably need to remove the IDelegateCommand stuff unless you're using Prism (and want to make the same changes as me to the Prism library). Also note that we don't generally use RoutedCommands here (we use Prism's DelegateCommand<T> for pretty much everything) so please don't hold me responsible if my call to CommandManager.InvalidateRequerySuggested sets off some sort of quantum wavefuntion collapse cascade that destroys the known universe or anything.
using System;
using System.ComponentModel;
using System.Windows;
using System.Windows.Input;
namespace Microsoft.Practices.Composite.Wpf.Commands
{
/// <summary>
/// This class provides an attached property that, when set to true, will cause changes to the element's CommandParameter to
/// trigger the CanExecute handler to be called on the Command.
/// </summary>
public static class CommandParameterBehavior
{
/// <summary>
/// Identifies the IsCommandRequeriedOnChange attached property
/// </summary>
/// <remarks>
/// When a control has the <see cref="IsCommandRequeriedOnChangeProperty" />
/// attached property set to true, then any change to it's
/// <see cref="System.Windows.Controls.Primitives.ButtonBase.CommandParameter" /> property will cause the state of
/// the command attached to it's <see cref="System.Windows.Controls.Primitives.ButtonBase.Command" /> property to
/// be reevaluated.
/// </remarks>
public static readonly DependencyProperty IsCommandRequeriedOnChangeProperty =
DependencyProperty.RegisterAttached("IsCommandRequeriedOnChange",
typeof(bool),
typeof(CommandParameterBehavior),
new UIPropertyMetadata(false, new PropertyChangedCallback(OnIsCommandRequeriedOnChangeChanged)));
/// <summary>
/// Gets the value for the <see cref="IsCommandRequeriedOnChangeProperty"/> attached property.
/// </summary>
/// <param name="target">The object to adapt.</param>
/// <returns>Whether the update on change behavior is enabled.</returns>
public static bool GetIsCommandRequeriedOnChange(DependencyObject target)
{
return (bool)target.GetValue(IsCommandRequeriedOnChangeProperty);
}
/// <summary>
/// Sets the <see cref="IsCommandRequeriedOnChangeProperty"/> attached property.
/// </summary>
/// <param name="target">The object to adapt. This is typically a <see cref="System.Windows.Controls.Primitives.ButtonBase" />,
/// <see cref="System.Windows.Controls.MenuItem" /> or <see cref="System.Windows.Documents.Hyperlink" /></param>
/// <param name="value">Whether the update behaviour should be enabled.</param>
public static void SetIsCommandRequeriedOnChange(DependencyObject target, bool value)
{
target.SetValue(IsCommandRequeriedOnChangeProperty, value);
}
private static void OnIsCommandRequeriedOnChangeChanged(DependencyObject d, DependencyPropertyChangedEventArgs e)
{
if (!(d is ICommandSource))
return;
if (!(d is FrameworkElement || d is FrameworkContentElement))
return;
if ((bool)e.NewValue)
{
HookCommandParameterChanged(d);
}
else
{
UnhookCommandParameterChanged(d);
}
UpdateCommandState(d);
}
private static PropertyDescriptor GetCommandParameterPropertyDescriptor(object source)
{
return TypeDescriptor.GetProperties(source.GetType())["CommandParameter"];
}
private static void HookCommandParameterChanged(object source)
{
var propertyDescriptor = GetCommandParameterPropertyDescriptor(source);
propertyDescriptor.AddValueChanged(source, OnCommandParameterChanged);
// N.B. Using PropertyDescriptor.AddValueChanged will cause "source" to never be garbage collected,
// so we need to hook the Unloaded event and call RemoveValueChanged there.
HookUnloaded(source);
}
private static void UnhookCommandParameterChanged(object source)
{
var propertyDescriptor = GetCommandParameterPropertyDescriptor(source);
propertyDescriptor.RemoveValueChanged(source, OnCommandParameterChanged);
UnhookUnloaded(source);
}
private static void HookUnloaded(object source)
{
var fe = source as FrameworkElement;
if (fe != null)
{
fe.Unloaded += OnUnloaded;
}
var fce = source as FrameworkContentElement;
if (fce != null)
{
fce.Unloaded += OnUnloaded;
}
}
private static void UnhookUnloaded(object source)
{
var fe = source as FrameworkElement;
if (fe != null)
{
fe.Unloaded -= OnUnloaded;
}
var fce = source as FrameworkContentElement;
if (fce != null)
{
fce.Unloaded -= OnUnloaded;
}
}
static void OnUnloaded(object sender, RoutedEventArgs e)
{
UnhookCommandParameterChanged(sender);
}
static void OnCommandParameterChanged(object sender, EventArgs ea)
{
UpdateCommandState(sender);
}
private static void UpdateCommandState(object target)
{
var commandSource = target as ICommandSource;
if (commandSource == null)
return;
var rc = commandSource.Command as RoutedCommand;
if (rc != null)
{
CommandManager.InvalidateRequerySuggested();
}
var dc = commandSource.Command as IDelegateCommand;
if (dc != null)
{
dc.RaiseCanExecuteChanged();
}
}
}
}
There's a relatively simple way to "fix" this problem with DelegateCommand, though it requires updating the DelegateCommand source and re-compiling the Microsoft.Practices.Composite.Presentation.dll.
1) Download the Prism 1.2 source code and open the CompositeApplicationLibrary_Desktop.sln. In here is a Composite.Presentation.Desktop project that contains the DelegateCommand source.
2) Under the public event EventHandler CanExecuteChanged, modify to read as follows:
public event EventHandler CanExecuteChanged
{
add
{
WeakEventHandlerManager.AddWeakReferenceHandler( ref _canExecuteChangedHandlers, value, 2 );
// add this line
CommandManager.RequerySuggested += value;
}
remove
{
WeakEventHandlerManager.RemoveWeakReferenceHandler( _canExecuteChangedHandlers, value );
// add this line
CommandManager.RequerySuggested -= value;
}
}
3) Under protected virtual void OnCanExecuteChanged(), modify it as follows:
protected virtual void OnCanExecuteChanged()
{
// add this line
CommandManager.InvalidateRequerySuggested();
WeakEventHandlerManager.CallWeakReferenceHandlers( this, _canExecuteChangedHandlers );
}
4) Recompile the solution, then navigate to either the Debug or Release folder where the compiled DLLs live. Copy the Microsoft.Practices.Composite.Presentation.dll and .pdb (if you wish) to where you references your external assemblies, and then recompile your application to pull the new versions.
After this, CanExecute should be fired every time the UI renders elements bound to the DelegateCommand in question.
Take care,
Joe
refereejoe at gmail
After reading some good answers to similar questions I changed in your example the DelegateCommand slightly to make it work. Instead of using:
public event EventHandler CanExecuteChanged;
I changed it to:
public event EventHandler CanExecuteChanged
{
add { CommandManager.RequerySuggested += value; }
remove { CommandManager.RequerySuggested -= value; }
}
I removed the following two methods because I was too lazy to fix them
public void RaiseCanExecuteChanged()
and
protected virtual void OnCanExecuteChanged()
And that's all... this seems to ensure that CanExecute will be called when the Binding changes and after the Execute method
It will not automatically trigger if the ViewModel is changed but as mentioned in this thread possible by calling the CommandManager.InvalidateRequerySuggested on the GUI thread
Application.Current?.Dispatcher.Invoke(DispatcherPriority.Normal, (Action)CommandManager.InvalidateRequerySuggested);
In .NET 7.0 RC1 this has been fixed.
(At least in a sense...)
It should now automatically reevaluate CanExecute() whenever CommandParameter changes, including upon initialization.
Although that does not prevent the initial call to CanExecute() when CommandParameter is still null, many ICommand implementations should already be handling that and it does render obsolete the obscure and problematic XAML attribute ordering workaround / hack.
As intimated by #Daniel-Svensson in a GitHub comment:
the actual issue here is that ICommand.CanExecute is not reevaluated
when the value bound to CommandParameter changes. Doing so would
obviously be the correct behavior, since the command parameter is
passed to CanExecute, so everyone expects this behavior intuitively.
And that is what is being fixed.
According to #pchaurasia14, a Sr. Engineering Manager for WPF at Microsoft:
This has been fixed in RC1 release. You may try it out. ... I meant, .NET 7 RC1.
The GitHub tracking issue #316 in the dotnet/wpf project is listed as closed. The code change CommandParameter invalidates CanExecute #4217 has been included in .NET 7.0 RC1. It was merged on July 21, 2022 and is included in the list of commits (scroll way down) for the RC1 release.
I've logged this as a bug against WPF in .Net 4.0, as the problem still exists in Beta 2.
https://connect.microsoft.com/VisualStudio/feedback/ViewFeedback.aspx?FeedbackID=504976
Some of these answers are about binding to the DataContext to get the Command itself, but the question was about the CommandParameter being null when it shouldn't be. We also experienced this. On on a hunch, we found a very simple way to get this to work in our ViewModel. This is specifically for the CommandParameter null problem reported by the customer, with one line of code. Note the Dispatcher.BeginInvoke().
public DelegateCommand<objectToBePassed> CommandShowReport
{
get
{
// create the command, or pass what is already created.
var command = _commandShowReport ?? (_commandShowReport = new DelegateCommand<object>(OnCommandShowReport, OnCanCommandShowReport));
// For the item template, the OnCanCommand will first pass in null. This will tell the command to re-pass the command param to validate if it can execute.
Dispatcher.BeginInvoke((Action) delegate { command.RaiseCanExecuteChanged(); }, DispatcherPriority.DataBind);
return command;
}
}
not sure if this will work in a data template, but here is the binding syntax I use in a ListView Context menu to grab the current item as a command parameter:
CommandParameter=
"{Binding RelativeSource={RelativeSource AncestorType=ContextMenu},
Path=PlacementTarget.SelectedItem,
Mode=TwoWay}"
Its a long shot. to debug this you can try:
- checking the PreviewCanExecute event.
- use snoop/wpf mole to peek inside and see what the commandparameter is.
HTH,
The commandManager.InvalidateRequerySuggested works for me as well. I believe the following link talks about similar problem, and M$ dev confirmed the limitation in the current version, and the commandManager.InvalidateRequerySuggested is the workaround. http://social.expression.microsoft.com/Forums/en-US/wpf/thread/c45d2272-e8ba-4219-bb41-1e5eaed08a1f/
What important is the timing of invoking the commandManager.InvalidateRequerySuggested. This should be invoked after the relevant value change is notified.
Beside Ed Ball's suggestion on setting CommandParameter prior to Command, make sure your CanExecute method has a parameter of object type.
private bool OnDeleteSelectedItemsCanExecute(object SelectedItems)
{
// Your goes heres
}
Hope it prevents someone spending the huge amount of time I did to figure out how to receive SelectedItems as CanExecute parameter