I have an application that display different datasets (users, nationality, etc) on the screen using radOutlookbar.
I have manage to load the required views in each item to display the data with no problem.
I then built views for each dataset (users, nationality, etc) to display the details about each selected item (i.e:user) within the displayed datasets.
Case:
First, I need to display the respective view for each dataset when I click on it's item.
Second, The displayed view will have an option to edit/add the displayed details.
I want to achieve this scenario using state-base-navigation.
So,
I have a PRISM region inside ItemsControl with ItemsPanelTemplate of grid to host the loaded views, basically I load the views for each dataset.
Question,
How should I show/hide the respective view according to the selected dataset using VSM?
Question 2:
Should I be able to define another nested state inside the loaded view to enable the scenario of edit/add details for each view?
If someone have any idea to do this, will be of great help to have a starting code.
Best regards
May be there's other schemes to access VSM but I prefer to create AttachedProperty for it. Let me explain.
Here is VisualState manager
/// <summary>
/// Class will allow to change VisualSate on ViewModel via attached properties
/// </summary>
public static class VisualStateManagerEx
{
private static PropertyChangedCallback callback = new PropertyChangedCallback(VisualStateChanged);
/// <summary>
/// Gets the state of the visual.
/// </summary>
/// <param name="obj">The obj.</param>
/// <returns></returns>
public static string GetVisualState(DependencyObject obj)
{
return (string)obj.GetValue(VisualStateProperty);
}
/// <summary>
/// Sets the state of the visual.
/// </summary>
/// <param name="obj">The obj.</param>
/// <param name="value">The value.</param>
public static void SetVisualState(DependencyObject obj, string value)
{
obj.SetValue(VisualStateProperty, value);
}
/// <summary>
/// DP for 'VisualState'
/// </summary>
public static readonly DependencyProperty VisualStateProperty =
DependencyProperty.RegisterAttached(
"VisualState",
typeof(string),
typeof(VisualStateManagerEx),
new PropertyMetadata(null, VisualStateManagerEx.callback)
);
/// <summary>
/// Visuals the state changed.
/// </summary>
/// <param name="d">The d.</param>
/// <param name="e">The <see cref="System.Windows.DependencyPropertyChangedEventArgs"/> instance containing the event data.</param>
public static void VisualStateChanged(DependencyObject d, DependencyPropertyChangedEventArgs e)
{
//Control changeStateControl = d as Control;
FrameworkElement changeStateControl = d as FrameworkElement;
if (changeStateControl == null)
throw (new Exception("VisualState works only on Controls type"));
if (Application.Current.Dispatcher.CheckAccess() == false)
{
// Wrong thread
System.Diagnostics.Debug.WriteLine("[VisualStateManagerEx] 'VisualStateChanged' event received on wrong thread -> re-route via Dispatcher");
Application.Current.Dispatcher.BeginInvoke(
//() => { VisualStateChanged(d, e); }
VisualStateManagerEx.callback
, new object[] { d, e }); //recursive
}
else
{
if (string.IsNullOrEmpty(e.NewValue.ToString()) == false)
{
//VisualStateManager.GoToState(changeStateControl, e.NewValue.ToString(), true);
VisualStateManager.GoToElementState(changeStateControl, e.NewValue.ToString(), true);
System.Diagnostics.Debug.WriteLine("[VisualStateManagerEx] Visual state changed to " + e.NewValue.ToString());
}
}
}
}
now - in XAML you attach it to your ViewModel like this:
<UserControl
xmlns:VSManagerEx=clr-namespace:Namespace.namespace;assembly=Assembly01"
VSManagerEx:VisualStateManagerEx.VisualState="{Binding Path=ViewModelVisualState, Mode=TwoWay, UpdateSourceTrigger=PropertyChanged}"
...
...
Now your VSM in XAML is bound to ViewModelVisualState property in ViewModelBase (or whatever will be bound to DataContext of this UserControl. So Actually in your ViewModelBase you using is like this:
/// <summary>
/// Base class for all 'view-models'
/// </summary>
[Export(typeof(ViewModelBase))]
public abstract class ViewModelBase : INavigationAware, INotifyPropertyChanged
{
private SynchronizationContext parentSyncContent;
#region VisualState
private string viewModelVisualState = string.Empty;
/// <summary>
/// Gets or sets the state of the view model visual.
/// </summary>
/// <value>
/// The state of the view model visual.
/// </value>
public virtual string ViewModelVisualState
{
get { return viewModelVisualState; }
set
{
viewModelVisualState = value;
RaisePropertyChanged(this, "ViewModelVisualState");
}
}
#endregion
/// <summary>
/// Raises the property changed.
/// </summary>
/// <param name="Sender">The sender.</param>
/// <param name="PropertyName">Name of the property.</param>
public void RaisePropertyChanged(object Sender, string PropertyName)
{
parentSyncContent.Post((state) =>
{
if (PropertyChanged != null)
PropertyChanged(Sender, new PropertyChangedEventArgs(PropertyName));
}, null);
}
...
...
So - in any ViewModel that inherit from this ViewModelBase could declare it own VMS states and manage them like this:
[Export(typeof(IViewModel1))
public ViewModel1 : ViewModelBase, IViewModel1
{
private const string VM_STATE_WORKING = "WorkingState";
internal void StartWorking()
{
this.ViewModelVisualState = VM_STATE_WORKING;
...
...
Regards question 2: No - you don't need to declare any additional Views inside anything. Read PRISM documentation about Navigation. There's great examples on how to create View/ViewModel that support various presentation logic.
Is this helpful to you ?
Related
This question already has answers here:
What's the best way to pass event to ViewModel?
(3 answers)
Closed 8 years ago.
I am trying to find a simple example on how to bind some TextBox events (PreviewTextInput and PreviewKeyDown) to a Commands, however I can't find any clear example and all the exmaples I found so far enforce me to use some MVVM framework (Light toolkit, Prism, etc.), however currently I don't want to use a framework because I want to understand more deeply how the business works.
Can anyone please supply a simple example on how this can be achieved?
Is it absolutely necessary to use MVVM framework?
Thanks in advance.
The easy way is to attach a common event handler to the event in XAML, and invoke the Command in code-behind. Such an event handler could look like the following:
private void TextBox_OnTextChanged(object sender, EventArgs e)
{
var viewmodel = this.DataContext as MyViewmodel;
if (viewmodel != null)
{
viewmodel.SomeCommand.Execute();
}
}
An alternative that works without any code in the code-behind (but is a bit tricky to implement, and works only on .NET 4.5) is to implement your own MarkupExtension, such that you can code something like
<TextBox TextChanged="{myMarkupExtension:CommandBinding SomeCommand]">...</TextBox>
There are a few articles out there describing this approach, for example this one
You can inherit TextBox and implement ICommandSource. I have done the same, my implementation looked like this. You should be able to extend this to work on PreviewTextInput.
public class CommandTextBox : TextBox, ICommandSource
{
private bool _canExecute;
private EventHandler _canExecuteChanged;
/// <summary>
/// DependencyProperty for Command property.
/// </summary>
public static readonly DependencyProperty CommandProperty = DependencyProperty.Register("Command", typeof(ICommand), typeof(CommandTextBox), new PropertyMetadata(OnCommandChanged));
/// <summary>
/// Gets or sets the command to invoke when the enter key is pressed.
/// </summary>
public ICommand Command
{
get { return (ICommand)GetValue(CommandProperty); }
set { SetValue(CommandProperty, value); }
}
/// <summary>
/// DependencyProperty for CommandParameter property.
/// </summary>
public static readonly DependencyProperty CommandParameterProperty = DependencyProperty.Register("CommandParameter", typeof(object), typeof(CommandTextBox));
/// <summary>
/// Gets or sets the parameter to pass to the Command property.
/// </summary>
public object CommandParameter
{
get { return GetValue(CommandParameterProperty); }
set { SetValue(CommandParameterProperty, value); }
}
/// <summary>
/// Gets or sets a value that indicates whether the command resets the text property.
/// </summary>
public bool CommandResetsText { get; set; }
/// <summary>
/// DependencyProperty for CommandTarget property.
/// </summary>
public static readonly DependencyProperty CommandTargetProperty = DependencyProperty.Register("CommandTarget", typeof(IInputElement), typeof(CommandTextBox));
/// <summary>
/// Gets or sets the element on which to raise the specified command.
/// </summary>
public IInputElement CommandTarget
{
get { return (IInputElement)GetValue(CommandTargetProperty); }
set { SetValue(CommandTargetProperty, value); }
}
/// <summary>
/// Gets a value that becomes the return value of
/// System.Windows.UIElement.IsEnabled in derived classes.
/// </summary>
protected override bool IsEnabledCore
{
get { return base.IsEnabledCore && _canExecute; }
}
/// <summary>
/// Command dependency property change callback.
/// </summary>
/// <param name="d">Dependency Object</param>
/// <param name="e">Event Args</param>
private static void OnCommandChanged(DependencyObject d, DependencyPropertyChangedEventArgs e)
{
CommandTextBox tb = (CommandTextBox)d;
tb.HookUpCommand((ICommand)e.OldValue, (ICommand)e.NewValue);
}
/// <summary>
/// If Command is defined, pressing the enter key will invoke the command;
/// Otherwise, the textbox will behave normally.
/// </summary>
/// <param name="e">Provides data about the event.</param>
protected override void OnKeyDown(KeyEventArgs e)
{
base.OnKeyDown(e);
if (e.Key == Key.Enter && Command != null)
{
RoutedCommand command = Command as RoutedCommand;
if (command != null)
command.Execute(CommandParameter, CommandTarget);
else
Command.Execute(CommandParameter);
if (CommandResetsText)
this.Text = String.Empty;
}
}
/// <summary>
/// Add a command to the Command Property.
/// </summary>
/// <param name="command">Command</param>
private void AddCommand(ICommand command)
{
var handler = new EventHandler(CanExecuteChanged);
_canExecuteChanged = handler;
if (command != null)
command.CanExecuteChanged += _canExecuteChanged;
}
private void CanExecuteChanged(object sender, EventArgs e)
{
if (Command != null)
{
RoutedCommand command = Command as RoutedCommand;
// If a RoutedCommand.
if (command != null)
_canExecute = command.CanExecute(CommandParameter, CommandTarget);
else
_canExecute = Command.CanExecute(CommandParameter);
}
CoerceValue(UIElement.IsEnabledProperty);
}
/// <summary>
/// Add a new command to the Command Property.
/// </summary>
/// <param name="oldCommand">Old Command</param>
/// <param name="newCommand">New Command</param>
private void HookUpCommand(ICommand oldCommand, ICommand newCommand)
{
// If oldCommand is not null, then we need to remove the handlers.
if (oldCommand != null)
RemoveCommand(oldCommand);
AddCommand(newCommand);
}
/// <summary>
/// Remove a command from the Command Property.
/// </summary>
/// <param name="command">Command</param>
private void RemoveCommand(ICommand command)
{
EventHandler handler = CanExecuteChanged;
command.CanExecuteChanged -= handler;
}
}
I have lots of properties in my viewmodels that have very little logic in them but they have the RaisePropertyChanged() method in them to refresh the GUI. i.e.
private bool _StatesIsSelected;
public bool StatesIsSelected
{
get { return _StatesIsSelected; }
set
{
_StatesIsSelected = value;
RaisePropertyChanged("StatesIsSelected");
}
}
I am starting to wonder if I should have unit tests confirming that the RaisePropertyChanged() method has been called. If I forgot to put it in the property the GUI wouldn't get refreshed and the application would have a bug...so it should have a unit test. But how do you test that?
So to sum it up....Am I being to militant about having unit tests for this logic? And if I am not being to militant...what is there a good way to test for this?
Are you being militant? That is not really an easy question to answer. We do test the majority of our property changed events, of which there are a lot, and I'm not really sure how much value there is in those tests. By that I mean if we removed them and stopped writing them in the future would we start seeing more bugs, or even any that wouldn't be pretty obvious as soon as you used the client? To be honest the answer is probably no. Conversely they are easy tests to write and certainly don't hurt.
Anyway, yes there is a very nice way to do this (had to make a few minor tweeks, so can't guarantee the code will compile, but should make the concepts clear):
public static class PropertyChangedTestHelperFactory
{
/// <summary>
/// Factory method for creating <see cref="PropertyChangedTestHelper{TTarget}"/> instances.
/// </summary>
/// <param name="target">
/// The target.
/// </param>
/// <typeparam name="TTarget">
/// The target type.
/// </typeparam>
/// <returns>
/// The <see cref="PropertyChangedTestHelper{TTarget}"/>
/// </returns>
public static PropertyChangedTestHelper<TTarget> CreatePropertyChangedHelper<TTarget>(
this TTarget target)
where TTarget : INotifyPropertyChanged
{
return new PropertyChangedTestHelper<TTarget>(target);
}
}
public sealed class PropertyChangedTestHelper<TTarget> : IDisposable
where TTarget : INotifyPropertyChanged
{
/// <summary>
/// This list contains the expected property names that should occur in property change notifications
/// </summary>
private readonly Queue<string> propertyNames = new Queue<string>();
/// <summary>
/// The target of the helper
/// </summary>
private readonly TTarget target;
/// <summary>
/// Initialises a new instance of the <see cref="StrictPropertyChangedTestHelper{TTarget}"/> class.
/// </summary>
/// <param name="target">The target.</param>
public PropertyChangedTestHelper(TTarget target)
{
this.target = target;
this.target.PropertyChanged += this.HandleTargetPropertyChanged;
}
/// <summary>
/// Performs application-defined tasks associated with freeing, releasing, or resetting unmanaged resources.
/// </summary>
public void Dispose()
{
this.target.PropertyChanged -= this.HandleTargetPropertyChanged;
if (this.propertyNames.Count != 0)
{
Assert.Fail("Property change notification {0} was not raised", this.propertyNames.Peek());
}
}
/// <summary>
/// Sets an expectation that a refresh change notification will be raised.
/// </summary>
public void ExpectRefresh()
{
this.propertyNames.Enqueue(string.Empty);
}
/// <summary>
/// Sets an expectation that a property change notification will be raised.
/// </summary>
/// <typeparam name="TProperty">The type of the property.</typeparam>
/// <param name="propertyExpression">The property expression.</param>
public void Expect<TProperty>(Expression<Func<TTarget, TProperty>> propertyExpression)
{
this.propertyNames.Enqueue(((MemberExpression)propertyExpression.Body).Member.Name);
}
/// <summary>
/// Handles the target property changed event.
/// </summary>
/// <param name="sender">The sender.</param>
/// <param name="e">The <see cref="PropertyChangedEventArgs"/> instance containing the event data.</param>
private void HandleTargetPropertyChanged(object sender, PropertyChangedEventArgs e)
{
if (this.propertyNames.Count == 0)
{
Assert.Fail("Unexpected property change notification {0}", e.PropertyName);
}
var expected = this.propertyNames.Dequeue();
var propertyName = (e.PropertyName ?? string.Empty).Trim();
if (propertyName != expected)
{
Assert.Fail("Out of order property change notification, expected '{0}', actual '{1}'", expected, propertyName);
}
}
}
Usage:
[TestMethod]
public void StatesIsSelected_RaisesIsValidChangeNotification()
{
// Arrange
var target = new SomeViewModel();
using (var helper = target.CreatePropertyChangedHelper())
{
helper.Expect(item => item.StatesIsSelected);
// Act
target.StatesIsSelected = true;
// Assert
}
}
When the helper is disposed the expectations are interrogated and the test will fail if they are not all met in the order they were defined.
We also have a Weak version that only requires that the expectations are met, not that they are met exactly (i.e. other property change events could be raised) and that is not order dependent.
FYI - if I were you I'd think about ditching MVVMLight and moving to Caliburn.Micro, its in a different league.
you can test it easily:
void TestMethod()
{
Container container = new Container();
bool isRaised = false;
container.PropertyChanged += (o,e) => {
if(e.PropertyName == "StatesIsSelected")
isRaised = true;
};
container.StatesIsSelected = true;
Assert.True(isRaised);
}
It think it would be useful to write the test for the ViewModel base class but not for all properties which raise a change, thats just too extreme
If you're targeting .NET Framework >= 4.5 you can inherit from ViewModelBase and write helper method:
public class ViewModelBaseExtended : ViewModelBase
{
protected void TryRaisePropertyChanged<T>(ref T oldValue, T newValue,
[CallerMemberName] string propertyName = "")
{
if (oldValue == null || !oldValue.Equals(newValue))
{
oldValue = newValue;
RaisePropertyChanged(propertyName);
}
}
}
and your property code would look like that:
private bool _StatesIsSelected;
public bool StatesIsSelected
{
get { return _StatesIsSelected; }
set
{
TryRaisePropertyChanged(ref _StatesIsSelected, value);
}
}
Now you would only have to assert property value in your unit tests.
When navigating between Views/ViewModels using RequestNavigate (i.e. programmatically), the IConfirmNavigationRequest methods on the appropriate ViewModels are called as expected. However, if you switch views in a TabControl region by clicking on the tab, it does not call those methods.
Is this the expected and accepted behaviour? Would I be able to implement a prism behavior to make this work?
Any advice would be appreciated.
UPDATE
I've decided to explain the problem more thoroughly based on Viktor's feedback. I want to prevent navigation if the user has unsaved edits on the screen. Switching tabs IMHO is just another way to navigate. I expect the Prism implementation to be consistent: navigating programmatically or otherwise should have the same behaviour.
If I were to create an ItemsControl with buttons that when clicked navigates by using RequestNavigate (to effectively switch tabs) it would work, but that isn't the point of the question.
I suppose I can see your point, and I understand why you would like it to call the RequestNavigate method.
To answer your question, yes this is by design and it is not supposed to call RequestNavigate while switching tabs. However, you can modify this behavior to do what you want. Prism is open source. You should have the source code, you can add the project to your project and easily step through the code for the following:
TabControlRegionAdapter - Adapts the region to the tab control
public class TabControlRegionAdapter : RegionAdapterBase<TabControl>
{
/// <summary>
/// <see cref="Style"/> to set to the created <see cref="TabItem"/>.
/// </summary>
public static readonly DependencyProperty ItemContainerStyleProperty =
DependencyProperty.RegisterAttached("ItemContainerStyle", typeof(Style), typeof(TabControlRegionAdapter), null);
/// <summary>
/// Initializes a new instance of the <see cref="TabControlRegionAdapter"/> class.
/// </summary>
/// <param name="regionBehaviorFactory">The factory used to create the region behaviors to attach to the created regions.</param>
public TabControlRegionAdapter(IRegionBehaviorFactory regionBehaviorFactory)
: base(regionBehaviorFactory)
{
}
/// <summary>
/// Gets the <see cref="ItemContainerStyleProperty"/> property value.
/// </summary>
/// <param name="target">Target object of the attached property.</param>
/// <returns>Value of the <see cref="ItemContainerStyleProperty"/> property.</returns>
public static Style GetItemContainerStyle(DependencyObject target)
{
if (target == null) throw new ArgumentNullException("target");
return (Style)target.GetValue(ItemContainerStyleProperty);
}
/// <summary>
/// Sets the <see cref="ItemContainerStyleProperty"/> property value.
/// </summary>
/// <param name="target">Target object of the attached property.</param>
/// <param name="value">Value to be set on the <see cref="ItemContainerStyleProperty"/> property.</param>
public static void SetItemContainerStyle(DependencyObject target, Style value)
{
if (target == null) throw new ArgumentNullException("target");
target.SetValue(ItemContainerStyleProperty, value);
}
/// <summary>
/// Adapts a <see cref="TabControl"/> to an <see cref="IRegion"/>.
/// </summary>
/// <param name="region">The new region being used.</param>
/// <param name="regionTarget">The object to adapt.</param>
protected override void Adapt(IRegion region, TabControl regionTarget)
{
if (regionTarget == null) throw new ArgumentNullException("regionTarget");
bool itemsSourceIsSet = regionTarget.ItemsSource != null;
if (itemsSourceIsSet)
{
throw new InvalidOperationException(Resources.ItemsControlHasItemsSourceException);
}
}
/// <summary>
/// Attach new behaviors.
/// </summary>
/// <param name="region">The region being used.</param>
/// <param name="regionTarget">The object to adapt.</param>
/// <remarks>
/// This class attaches the base behaviors and also keeps the <see cref="TabControl.SelectedItem"/>
/// and the <see cref="IRegion.ActiveViews"/> in sync.
/// </remarks>
protected override void AttachBehaviors(IRegion region, TabControl regionTarget)
{
if (region == null) throw new ArgumentNullException("region");
base.AttachBehaviors(region, regionTarget);
if (!region.Behaviors.ContainsKey(TabControlRegionSyncBehavior.BehaviorKey))
{
region.Behaviors.Add(TabControlRegionSyncBehavior.BehaviorKey, new TabControlRegionSyncBehavior { HostControl = regionTarget });
}
}
/// <summary>
/// Creates a new instance of <see cref="Region"/>.
/// </summary>
/// <returns>A new instance of <see cref="Region"/>.</returns>
protected override IRegion CreateRegion()
{
return new SingleActiveRegion();
}
}
And also, TabControlRegionSyncBehavior. This is the one which you could call RequestNavigate.
public class TabControlRegionSyncBehavior : RegionBehavior, IHostAwareRegionBehavior
{
///<summary>
/// The behavior key for this region sync behavior.
///</summary>
public const string BehaviorKey = "TabControlRegionSyncBehavior";
private static readonly DependencyProperty IsGeneratedProperty =
DependencyProperty.RegisterAttached("IsGenerated", typeof(bool), typeof(TabControlRegionSyncBehavior), null);
private TabControl hostControl;
/// <summary>
/// Gets or sets the <see cref="DependencyObject"/> that the <see cref="IRegion"/> is attached to.
/// </summary>
/// <value>A <see cref="DependencyObject"/> that the <see cref="IRegion"/> is attached to.
/// This is usually a <see cref="FrameworkElement"/> that is part of the tree.</value>
public DependencyObject HostControl
{
get
{
return this.hostControl;
}
set
{
TabControl newValue = value as TabControl;
if (newValue == null)
{
throw new InvalidOperationException(Resources.HostControlMustBeATabControl);
}
if (IsAttached)
{
throw new InvalidOperationException(Resources.HostControlCannotBeSetAfterAttach);
}
this.hostControl = newValue;
}
}
/// <summary>
/// Override this method to perform the logic after the behavior has been attached.
/// </summary>
protected override void OnAttach()
{
if (this.hostControl == null)
{
throw new InvalidOperationException(Resources.HostControlCannotBeNull);
}
this.SynchronizeItems();
this.hostControl.SelectionChanged += this.OnSelectionChanged;
this.Region.ActiveViews.CollectionChanged += this.OnActiveViewsChanged;
this.Region.Views.CollectionChanged += this.OnViewsChanged;
}
/// <summary>
/// Gets the item contained in the <see cref="TabItem"/>.
/// </summary>
/// <param name="tabItem">The container item.</param>
/// <returns>The item contained in the <paramref name="tabItem"/> if it was generated automatically by the behavior; otherwise <paramref name="tabItem"/>.</returns>
protected virtual object GetContainedItem(TabItem tabItem)
{
if (tabItem == null) throw new ArgumentNullException("tabItem");
if ((bool)tabItem.GetValue(IsGeneratedProperty))
{
return tabItem.Content;
}
return tabItem;
}
/// <summary>
/// Override to change how TabItem's are prepared for items.
/// </summary>
/// <param name="item">The item to wrap in a TabItem</param>
/// <param name="parent">The parent <see cref="DependencyObject"/></param>
/// <returns>A tab item that wraps the supplied <paramref name="item"/></returns>
protected virtual TabItem PrepareContainerForItem(object item, DependencyObject parent)
{
TabItem container = item as TabItem;
if (container == null)
{
object dataContext = GetDataContext(item);
container = new TabItem();
container.Content = item;
container.Style = TabControlRegionAdapter.GetItemContainerStyle(parent);
container.DataContext = dataContext; // To run with SL 2
container.Header = dataContext; // To run with SL 3
container.SetValue(IsGeneratedProperty, true);
}
return container;
}
/// <summary>
/// Undoes the effects of the <see cref="PrepareContainerForItem"/> method.
/// </summary>
/// <param name="tabItem">The container element for the item.</param>
protected virtual void ClearContainerForItem(TabItem tabItem)
{
if (tabItem == null) throw new ArgumentNullException("tabItem");
if ((bool)tabItem.GetValue(IsGeneratedProperty))
{
tabItem.Content = null;
}
}
/// <summary>
/// Creates or identifies the element that is used to display the given item.
/// </summary>
/// <param name="item">The item to get the container for.</param>
/// <param name="itemCollection">The parent's <see cref="ItemCollection"/>.</param>
/// <returns>The element that is used to display the given item.</returns>
protected virtual TabItem GetContainerForItem(object item, ItemCollection itemCollection)
{
if (itemCollection == null) throw new ArgumentNullException("itemCollection");
TabItem container = item as TabItem;
if (container != null && ((bool)container.GetValue(IsGeneratedProperty)) == false)
{
return container;
}
foreach (TabItem tabItem in itemCollection)
{
if ((bool)tabItem.GetValue(IsGeneratedProperty))
{
if (tabItem.Content == item)
{
return tabItem;
}
}
}
return null;
}
/// <summary>
/// Return the appropriate data context. If the item is a FrameworkElement it cannot be a data context in Silverlight, so we use its data context.
/// Otherwise, we just us the item as the data context.
/// </summary>
private static object GetDataContext(object item)
{
FrameworkElement frameworkElement = item as FrameworkElement;
return frameworkElement == null ? item : frameworkElement.DataContext;
}
private void SynchronizeItems()
{
List<object> existingItems = new List<object>();
if (this.hostControl.Items.Count > 0)
{
// Control must be empty before "Binding" to a region
foreach (object childItem in this.hostControl.Items)
{
existingItems.Add(childItem);
}
}
foreach (object view in this.Region.Views)
{
TabItem tabItem = this.PrepareContainerForItem(view, this.hostControl);
this.hostControl.Items.Add(tabItem);
}
foreach (object existingItem in existingItems)
{
this.Region.Add(existingItem);
}
}
private void OnSelectionChanged(object sender, SelectionChangedEventArgs e)
{
// e.OriginalSource == null, that's why we use sender.
if (this.hostControl == sender)
{
foreach (TabItem tabItem in e.RemovedItems)
{
object item = this.GetContainedItem(tabItem);
// check if the view is in both Views and ActiveViews collections (there may be out of sync)
if (this.Region.Views.Contains(item) && this.Region.ActiveViews.Contains(item))
{
this.Region.Deactivate(item);
}
}
foreach (TabItem tabItem in e.AddedItems)
{
object item = this.GetContainedItem(tabItem);
if (!this.Region.ActiveViews.Contains(item))
{
this.Region.Activate(item);
}
}
}
}
private void OnActiveViewsChanged(object sender, NotifyCollectionChangedEventArgs e)
{
if (e.Action == NotifyCollectionChangedAction.Add)
{
this.hostControl.SelectedItem = this.GetContainerForItem(e.NewItems[0], this.hostControl.Items);
}
else if (e.Action == NotifyCollectionChangedAction.Remove
&& this.hostControl.SelectedItem != null
&& e.OldItems.Contains(this.GetContainedItem((TabItem)this.hostControl.SelectedItem)))
{
this.hostControl.SelectedItem = null;
}
}
private void OnViewsChanged(object sender, NotifyCollectionChangedEventArgs e)
{
if (e.Action == NotifyCollectionChangedAction.Add)
{
int startingIndex = e.NewStartingIndex;
foreach (object newItem in e.NewItems)
{
TabItem tabItem = this.PrepareContainerForItem(newItem, this.hostControl);
this.hostControl.Items.Insert(startingIndex, tabItem);
}
}
else if (e.Action == NotifyCollectionChangedAction.Remove)
{
foreach (object oldItem in e.OldItems)
{
TabItem tabItem = this.GetContainerForItem(oldItem, this.hostControl.Items);
this.hostControl.Items.Remove(tabItem);
this.ClearContainerForItem(tabItem);
}
}
}
}
Of course, you'll have to figure out where to call RequestNavigate, such that you can actually cancel the TabSelectionChanging. Unfortunately, this event doesn't exist in WPF. I would resort to the trick recommended by Josh Smith How to Prevent a TabItem from changing
What I understood from your question is that you expect that switching tabs calls IConfirmNavigationRequest. Method from this interface is called when you navigating from view/viewModel implementing this interface.
But, what you experiencing when you switch tabs in TabControl is not Navigation request. All views in TabControl already handled Navigation operation and all views are already in TabControl(Your Region). So what you do when you switch tabs? You only Activating view within your region. Previously active view gets deactivated.
I really don't know what are you trying to accomplish. I cannot imagine why whould I prevent somebody from switching tabs. But you could try that by using IActiveAware interface. You can get the idea from this blog
EDIT
Implement OnDeactivate to ask user whether or not he wants to save changes before deactivating view
Implement OnActivate to call RequestNavigate to already existing View. U can read about Navigating to Existing Views in Prism documentation.
Disable all other tabItems and enable them again after saving changes(bad approach)
I am really not an expert, but I don't think you have more options left
I'm trying to convert an event to a command on a devexpress wpf grid context menu item which is derived from FrameworkContentElement instead of FrameworkElement. This causes a runtime error :
{"Cannot attach type \"EventToCommand\" to type \"BarButtonItem\". Instances of type \"EventToCommand\" can only be attached to objects of type \"FrameworkElement\"."}
Is there any workaround?
<dxg:TableView.RowCellMenuCustomizations>
<dxb:BarButtonItem Name="deleteRowItem" Content="Delete" >
<i:Interaction.Triggers>
<i:EventTrigger EventName="ItemClick">
<cmd:EventToCommand Command="{Binding FooChangeCommand}"
PassEventArgsToCommand="True" />
</i:EventTrigger>
</i:Interaction.Triggers>
</dxb:BarButtonItem>
<!--ItemClick="deleteRowItem_ItemClick"/>-->
</dxg:TableView.RowCellMenuCustomizations>
Unfortunately devexpress have run into problems changing the base class to FrameworkElement having intended to make that change...
The FrameworkConentElement is a class that is only available in WPF and not in Silverlight. As MVVM Light is intended to provide a common functionality for all WPF dialects (WPF 3.5, WPF 4, Silverlight 3, Silverlight 4, Sivlverlight 5, WP 7, WP 7.1) it cannot include an implementation that only works in one of the frameworks.
For a discussion about the differences between FrameworkElement and FrameworkContentElement see here.
However, you can just easily implement your own EventToCommand class supporting ContentElement (from which FrameworkContentElement inherits). The class was copied from BL0015 of the MVVM Light source code and modified:
using System;
using System.Windows;
using System.Windows.Input;
using System.Windows.Interactivity;
namespace GalaSoft.MvvmLight.Command
{
/// <summary>
/// This <see cref="System.Windows.Interactivity.TriggerAction" /> can be
/// used to bind any event on any FrameworkElement to an <see cref="ICommand" />.
/// Typically, this element is used in XAML to connect the attached element
/// to a command located in a ViewModel. This trigger can only be attached
/// to a FrameworkElement or a class deriving from FrameworkElement.
/// <para>To access the EventArgs of the fired event, use a RelayCommand<EventArgs>
/// and leave the CommandParameter and CommandParameterValue empty!</para>
/// </summary>
////[ClassInfo(typeof(EventToCommand),
//// VersionString = "3.0.0.0",
//// DateString = "201003041420",
//// Description = "A Trigger used to bind any event to an ICommand.",
//// UrlContacts = "http://stackoverflow.com/q/6955785/266919",
//// Email = "")]
public partial class EventToCommandWpf : TriggerAction<DependencyObject>
{
/// <summary>
/// Gets or sets a value indicating whether the EventArgs passed to the
/// event handler will be forwarded to the ICommand's Execute method
/// when the event is fired (if the bound ICommand accepts an argument
/// of type EventArgs).
/// <para>For example, use a RelayCommand<MouseEventArgs> to get
/// the arguments of a MouseMove event.</para>
/// </summary>
public bool PassEventArgsToCommand
{
get;
set;
}
/// <summary>
/// Provides a simple way to invoke this trigger programatically
/// without any EventArgs.
/// </summary>
public void Invoke()
{
Invoke(null);
}
/// <summary>
/// Executes the trigger.
/// <para>To access the EventArgs of the fired event, use a RelayCommand<EventArgs>
/// and leave the CommandParameter and CommandParameterValue empty!</para>
/// </summary>
/// <param name="parameter">The EventArgs of the fired event.</param>
protected override void Invoke(object parameter)
{
if (AssociatedElementIsDisabled())
{
return;
}
var command = GetCommand();
var commandParameter = CommandParameterValue;
if (commandParameter == null
&& PassEventArgsToCommand)
{
commandParameter = parameter;
}
if (command != null
&& command.CanExecute(commandParameter))
{
command.Execute(commandParameter);
}
}
private static void OnCommandChanged(
EventToCommandWpf element,
DependencyPropertyChangedEventArgs e)
{
if (element == null)
{
return;
}
if (e.OldValue != null)
{
((ICommand)e.OldValue).CanExecuteChanged -= element.OnCommandCanExecuteChanged;
}
var command = (ICommand)e.NewValue;
if (command != null)
{
command.CanExecuteChanged += element.OnCommandCanExecuteChanged;
}
element.EnableDisableElement();
}
private bool AssociatedElementIsDisabled()
{
var element = GetAssociatedObject();
return AssociatedObject == null
|| (element != null
&& !element.IsEnabled);
}
private void EnableDisableElement()
{
var element = GetAssociatedObject();
if (element == null)
{
return;
}
var command = this.GetCommand();
if (this.MustToggleIsEnabledValue
&& command != null)
{
SetIsEnabled(element, command.CanExecute(this.CommandParameterValue));
}
}
private void OnCommandCanExecuteChanged(object sender, EventArgs e)
{
EnableDisableElement();
}
/// <summary>
/// Identifies the <see cref="CommandParameter" /> dependency property
/// </summary>
public static readonly DependencyProperty CommandParameterProperty = DependencyProperty.Register(
"CommandParameter",
typeof(object),
typeof(EventToCommandWpf),
new PropertyMetadata(
null,
(s, e) => {
var sender = s as EventToCommandWpf;
if (sender == null)
{
return;
}
if (sender.AssociatedObject == null)
{
return;
}
sender.EnableDisableElement();
}));
/// <summary>
/// Identifies the <see cref="Command" /> dependency property
/// </summary>
public static readonly DependencyProperty CommandProperty = DependencyProperty.Register(
"Command",
typeof(ICommand),
typeof(EventToCommandWpf),
new PropertyMetadata(
null,
(s, e) => OnCommandChanged(s as EventToCommandWpf, e)));
/// <summary>
/// Identifies the <see cref="MustToggleIsEnabled" /> dependency property
/// </summary>
public static readonly DependencyProperty MustToggleIsEnabledProperty = DependencyProperty.Register(
"MustToggleIsEnabled",
typeof(bool),
typeof(EventToCommandWpf),
new PropertyMetadata(
false,
(s, e) => {
var sender = s as EventToCommandWpf;
if (sender == null)
{
return;
}
if (sender.AssociatedObject == null)
{
return;
}
sender.EnableDisableElement();
}));
private object _commandParameterValue;
private bool? _mustToggleValue;
/// <summary>
/// Gets or sets the ICommand that this trigger is bound to. This
/// is a DependencyProperty.
/// </summary>
public ICommand Command
{
get
{
return (ICommand)GetValue(CommandProperty);
}
set
{
SetValue(CommandProperty, value);
}
}
/// <summary>
/// Gets or sets an object that will be passed to the <see cref="Command" />
/// attached to this trigger. This is a DependencyProperty.
/// </summary>
public object CommandParameter
{
get
{
return this.GetValue(CommandParameterProperty);
}
set
{
SetValue(CommandParameterProperty, value);
}
}
/// <summary>
/// Gets or sets an object that will be passed to the <see cref="Command" />
/// attached to this trigger. This property is here for compatibility
/// with the Silverlight version. This is NOT a DependencyProperty.
/// For databinding, use the <see cref="CommandParameter" /> property.
/// </summary>
public object CommandParameterValue
{
get
{
return this._commandParameterValue ?? this.CommandParameter;
}
set
{
_commandParameterValue = value;
EnableDisableElement();
}
}
/// <summary>
/// Gets or sets a value indicating whether the attached element must be
/// disabled when the <see cref="Command" /> property's CanExecuteChanged
/// event fires. If this property is true, and the command's CanExecute
/// method returns false, the element will be disabled. If this property
/// is false, the element will not be disabled when the command's
/// CanExecute method changes. This is a DependencyProperty.
/// </summary>
public bool MustToggleIsEnabled
{
get
{
return (bool)this.GetValue(MustToggleIsEnabledProperty);
}
set
{
SetValue(MustToggleIsEnabledProperty, value);
}
}
/// <summary>
/// Gets or sets a value indicating whether the attached element must be
/// disabled when the <see cref="Command" /> property's CanExecuteChanged
/// event fires. If this property is true, and the command's CanExecute
/// method returns false, the element will be disabled. This property is here for
/// compatibility with the Silverlight version. This is NOT a DependencyProperty.
/// For databinding, use the <see cref="MustToggleIsEnabled" /> property.
/// </summary>
public bool MustToggleIsEnabledValue
{
get
{
return this._mustToggleValue == null
? this.MustToggleIsEnabled
: this._mustToggleValue.Value;
}
set
{
_mustToggleValue = value;
EnableDisableElement();
}
}
/// <summary>
/// Called when this trigger is attached to a DependencyObject.
/// </summary>
protected override void OnAttached()
{
base.OnAttached();
EnableDisableElement();
}
/// <summary>
/// This method is here for compatibility
/// with the Silverlight version.
/// </summary>
/// <returns>The object to which this trigger
/// is attached casted as a FrameworkElement.</returns>
private IInputElement GetAssociatedObject()
{
return AssociatedObject as IInputElement;
}
private void SetIsEnabled(IInputElement element, bool value)
{
if (element is UIElement)
{
((UIElement)element).IsEnabled = value;
}
else if (element is ContentElement)
{
((ContentElement)element).IsEnabled = value;
}
else
{
throw new InvalidOperationException("Cannot set IsEnabled. Element is neither ContentElemen, nor UIElement.");
}
}
/// <summary>
/// This method is here for compatibility
/// with the Silverlight version.
/// </summary>
/// <returns>The command that must be executed when
/// this trigger is invoked.</returns>
private ICommand GetCommand()
{
return Command;
}
}
}
To inlcude it into your code you have to define a xml namespace pointing to the correct dll and then use it just like the normal EventToCommand class.
NOTE: This class does not work in Silverlight!
For those trying to solve this specific issue using dev express, this will do the trick!
<dxg:TableView.RowCellMenuCustomizations>
<dxb:BarButtonItem Name="deleteRowItem" Content="Delete" Command="{Binding View.DataContext.DeleteSelectionCommand}" />
</dxg:TableView.RowCellMenuCustomizations>
I was following their example which had an event on the button, little realising there was also a command I could use. Then the challenge was working out the binding as the menu item is not on the main visual tree. However the above solves that.
I found this to work with DEV express
<dxb:BarButtonItem Content="123" Name="item1">
<dxmvvm:Interaction.Triggers>
<dxmvvm:EventToCommand EventName="ItemClick" Command="{Binding SomeCommand}" CommandParameter="{Binding ElementName=item1, Path=Content}"/>
</dxmvvm:Interaction.Triggers>
</dxb:BarButtonItem>
I want to be able to route the double-click of a grid to a Command. I'm using Rx to simulate the double click but I can't figure out what control to attach the mouse handler to (the mouse event on the e.Row object in DataGrid.RowLoading event doesn't seem to work).
Anyone got any ideas?
Rx code for handling the Double click is as follows:
Observable.FromEvent<MouseButtonEventArgs>(e.Row, "MouseLeftButtonDown").TimeInterval().Subscribe(evt =>
{
if (evt.Interval.Milliseconds <= 300)
{
// Execute command on double click
}
});
I changed this code from handling MouseLeftButtonDown to MouseLeftButtonUp and it works now. The row must have something else handling the button down events.
Observable.FromEvent<MouseButtonEventArgs>(e.Row, "MouseLeftButtonUp").TimeInterval().Subscribe(evt =>
{
if (evt.Interval != TimeSpan.Zero && evt.Interval.TotalMilliseconds <= 300)
{
// Execute command on double click
}
});
Full source code is included below:
CommandBehaviorBase.cs
using System;
using System.Windows;
using System.Windows.Input;
using System.Windows.Interactivity;
/// <summary>
/// Provides the base implementation of all Behaviors that can be attached to a <see cref="FrameworkElement"/> which trigger a command.
/// </summary>
/// <typeparam name="T">The type of control this behavior can be attached to, must derive from <see cref="FrameworkElement"/>.</typeparam>
public abstract class CommandBehaviorBase<T> : Behavior<T> where T : FrameworkElement
{
#region Constants and Fields
/// <summary>The DependencyProperty backing store for CommandParameter. This enables animation, styling, binding, etc...</summary>
public static readonly DependencyProperty CommandParameterProperty =
DependencyProperty.Register(
"CommandParameter",
typeof(object),
typeof(CommandBehaviorBase<T>),
new PropertyMetadata(null, OnCommandParameterPropertyChanged));
/// <summary>The DependencyProperty backing store for Command. This enables animation, styling, binding, etc...</summary>
public static readonly DependencyProperty CommandProperty = DependencyProperty.Register(
"Command", typeof(ICommand), typeof(CommandBehaviorBase<T>), new PropertyMetadata(null));
#endregion
/// <summary>
/// Gets or sets the command to execute
/// </summary>
public ICommand Command
{
get
{
return (ICommand)this.GetValue(CommandProperty);
}
set
{
this.SetValue(CommandProperty, value);
}
}
/// <summary>
/// Gets or sets the command parameter to execute with.
/// </summary>
public object CommandParameter
{
get
{
return this.GetValue(CommandParameterProperty);
}
set
{
this.SetValue(CommandParameterProperty, value);
}
}
/// <summary>
/// Gets or sets the command binding path (Hack for SL3).
/// </summary>
/// <remarks>This is a hack to overcome the fact that we cannot
/// bind to the Command dependency property due to a limitation in Silverlight 3.0
/// This shouldn't be necessary as in Silverlight 4.0 <see cref="DependencyObject"/> supports data binding hooray!</remarks>
public string CommandPath { get; set; }
/// <summary>
/// Gets or sets a value indicating whether this mapping is currently enabled.
/// </summary>
public bool IsEnabled { get; set; }
/// <summary>
/// Implements the logic that disables the key mapping based on whether the command can be executed.
/// </summary>
/// <summary>
/// Updates the target object's IsEnabled property based on the commands ability to execute.
/// </summary>
public virtual void UpdateEnabledState()
{
if (this.Command == null && !string.IsNullOrEmpty(this.CommandPath))
{
this.Command = this.AssociatedObject.DataContext.GetPropertyPathValue<ICommand>(this.CommandPath, null);
}
if (this.AssociatedObject == null)
{
this.Command = null;
this.CommandParameter = null;
}
else if (this.Command != null)
{
this.IsEnabled = this.Command.CanExecute(this.CommandParameter);
}
}
/// <summary>
/// Executes the command, if it's set, providing the <see cref="CommandParameter"/>
/// </summary>
protected virtual void ExecuteCommand()
{
if (this.Command != null)
{
this.Command.Execute(this.CommandParameter);
}
}
/// <summary>
/// Attaches to the target <see cref="FrameworkElement"/> and sets up the command.
/// </summary>
protected override void OnAttached()
{
base.OnAttached();
this.UpdateEnabledState();
}
/// <summary>
/// Raised when the command parameter changes, re-evaluates whether the Command can execute
/// </summary>
/// <param name="sender">The KeyCommandBehavior that command parameter changed for.</param>
/// <param name="args">The parameter is not used.</param>
private static void OnCommandParameterPropertyChanged(
DependencyObject sender, DependencyPropertyChangedEventArgs args)
{
((CommandBehaviorBase<T>)sender).UpdateEnabledState();
}
}
DoubleClickCommandBehavior.cs
using System;
using System.Linq;
using System.Windows;
using System.Windows.Input;
/// <summary>
/// Provides behavior for any clickable control and will execute a command when the control is double clicked.
/// Does not disable the control if the command cannot be executed.
/// </summary>
public class DoubleClickCommandBehavior : CommandBehaviorBase<FrameworkElement>
{
#region Constants and Fields
/// <summary>
/// Stores the observable that subscribes to click events.
/// </summary>
private IDisposable observable;
#endregion
#region Constructors and Destructors
/// <summary>
/// Initializes a new instance of the DoubleClickCommandBehavior class.
/// </summary>
public DoubleClickCommandBehavior()
{
// Default double-click interval is 220 milliseconds.
this.Interval = 220;
}
#endregion
/// <summary>
/// Gets or sets the double click interval in milliseconds.
/// </summary>
public int Interval
{
get;
set;
}
/// <summary>
/// Subscribes to the MouseLeftButtonUp of the data grid and times the intervals between the events,
/// if the time between clicks is less than the configured interval the command is executed.
/// </summary>
/// <remarks>Originally attached to MouseLeftButtonDown but the events were not firing.</remarks>
protected override void OnAttached()
{
base.OnAttached();
this.observable =
Observable.FromEvent<MouseButtonEventArgs>(this.AssociatedObject, "MouseLeftButtonUp").TimeInterval().
Subscribe(
evt =>
{
if (evt.Interval != TimeSpan.Zero && evt.Interval.TotalMilliseconds <= this.Interval)
{
this.UpdateEnabledState();
this.ExecuteCommand();
}
});
}
/// <summary>
/// Disposes of the observable
/// </summary>
protected override void OnDetaching()
{
if (this.observable != null)
{
this.observable.Dispose();
this.observable = null;
}
base.OnDetaching();
}
}
I'm having similar problems (though not using Rx to handle the double click, instead using a generic DoubleClickTrigger). My specific problem is more related to the fact that I'm not sure how or where to hook up my trigger.
I've tried something like the following:
<data:DataGrid.Resources>
<ControlTemplate x:Key="rowTemplate" TargetType="data:DataGridRow">
<data:DataGridRow>
<fxui:Interaction.Triggers>
<fxui:DoubleClickTrigger>
<Interactivity:InvokeCommandAction Command="{Binding Source={StaticResource selectCommand}}" CommandParameter="{Binding}"/>
</fxui:DoubleClickTrigger>
</fxui:Interaction.Triggers>
</data:DataGridRow>
</ControlTemplate>
</data:DataGrid.Resources>
<data:DataGrid.RowStyle>
<Style TargetType="data:DataGridRow">
<Setter Property="Template" Value="{StaticResource rowTemplate}"/>
</Style>
</data:DataGrid.RowStyle>
With no luck.