We Keep Coding

How do I add and subtract event handlers inside a derived abstract class?

Short version
In my abstract class MyCbo_Abstract (derived from ComboBox class), I want to create a custom property that when set will subtract all the control's event handlers, set the base property value, then re-add all the control's event handlers.
What I have so far
I have a concrete ComboBox class derived from an abstract ComboBox class derived from Microsoft's ComboBox class.
public abstract class MyCbo_Abstract : ComboBox
{
public MyCbo_Abstract() : base()
{
}
}
public partial class MyCboFooList : MyCbo_Abstract
{
public MyCboFooList() : base()
{
}
}
My main Form class subscribes to certain base ComboBox events.
Note: The designer has: this.myCboFooList = new MyCboFooList();
public partial class FormMain : Form
{
public FormMain()
{
myCboFooList.SelectedIndexChanged += myCboFooList_SelectedIndexChanged;
}
private void myCboFooList_SelectedIndexChanged(object sender, EventArgs e)
{
// do stuff
}
}
There are times when I want to suppress the invocation of defined event handlers, e.g., when I programmatically set a ComboBox object's SelectedIndex property.
Instead of having to remember to write the code to subtract and re-add event handlers each time I want to modify the SelectedIndex property and suppress its events, I want to create a custom property SelectedIndex_NoEvents that when set will subtract all the control's event handlers, set the base property value SelectedIndex, then re-add all the control's event handlers.
The problem
My problem is that I don't know how to iterate over a EventHandlerList because it has no GetEnumerator. And, in looking at the list in the debugger, saveEventHandlerList is a weird chained thing that I can't figure out how to otherwise traverse.
public abstract class MyCbo_Abstract : ComboBox
{
int selectedIndex_NoEvents;
public int SelectedIndex_NoEvents
{
get
{
return base.SelectedIndex;
}
set
{
EventHandlerList saveEventHandlerList = new EventHandlerList();
saveEventHandlerList = Events;
//foreach won't work - no GetEnumerator available. Can't use for loop - no Count poprerty
foreach (EventHandler eventHandler in saveEventHandlerList)
{
SelectedIndexChanged -= eventHandler;
}
base.SelectedIndex = value;
//foreach won't work - no GetEnumerator available. Can't use for loop - no Count poprerty
foreach (EventHandler eventHandler in saveEventHandlerList)
{
SelectedIndexChanged += eventHandler;
}
saveEventHandlerList = null;
}
}
//Probably don't need this
public override int SelectedIndex
{
get
{
return base.SelectedIndex;
}
set
{
base.SelectedIndex = value;
}
}
public DRT_ComboBox_Abstract() : base()
{
}
}

Before giving you the solution that I created, let me say that this feels extremely hacky. I urge you to seriously think about another solution. There may be all kinds of crazy edge cases where this code breaks down, I haven't thoroughly tested it beyond the example code shown below.
Add the following utility class:
public class SuspendedEvents
{
private Dictionary<FieldInfo, Delegate> handlers = new Dictionary<System.Reflection.FieldInfo, System.Delegate>();
private object source;
public SuspendedEvents(object obj)
{
source = obj;
var fields = obj.GetType().GetFields(BindingFlags.Static | BindingFlags.Instance | BindingFlags.NonPublic | BindingFlags.Public);
foreach (var fieldInfo in fields.Where(fi => fi.FieldType.IsSubclassOf(typeof(Delegate))))
{
var d = (Delegate)fieldInfo.GetValue(obj);
handlers.Add(fieldInfo, (Delegate)d.Clone());
fieldInfo.SetValue(obj, null);
}
}
public void Restore()
{
foreach (var storedHandler in handlers)
{
storedHandler.Key.SetValue(source, storedHandler.Value);
}
}
}
You can use it like this:
var events = new SuspendedEvents(obj); //all event handlers on obj are now detached
events.Restore(); // event handlers on obj are now restored.
I used the following test setup:
void Main()
{
var obj = new TestObject();
obj.Event1 += (sender, e) => Handler("Event 1");
obj.Event1 += (sender, e) => Handler("Event 1");
obj.Event2 += (sender, e) => Handler("Event 2");
obj.Event2 += (sender, e) => Handler("Event 2");
obj.Event3 += (sender, e) => Handler("Event 3");
obj.Event3 += (sender, e) => Handler("Event 3");
Debug.WriteLine("Prove events are attached");
obj.RaiseEvents();
var events = new SuspendedEvents(obj);
Debug.WriteLine("Prove events are detached");
obj.RaiseEvents();
events.Restore();
Debug.WriteLine("Prove events are reattached");
obj.RaiseEvents();
}
public void Handler(string message)
{
Debug.WriteLine(message);
}
public class TestObject
{
public event EventHandler<EventArgs> Event1;
public event EventHandler<EventArgs> Event2;
public event EventHandler<EventArgs> Event3;
public void RaiseEvents()
{
Event1?.Invoke(this, EventArgs.Empty);
Event2?.Invoke(this, EventArgs.Empty);
Event3?.Invoke(this, EventArgs.Empty);
}
}
It produces the following output:
Prove events are attached
Event 1
Event 1
Event 2
Event 2
Event 3
Event 3
Prove events are detached
Prove events are reattached
Event 1
Event 1
Event 2
Event 2
Event 3
Event 3

There is no way to easily disable event firing of WinForm controls exposed in the .Net framework. However, the Winform controls follow a standard design pattern for events in that all event signatures are based on the EventHandler Delegate and the registered event handlers are stored in an EventHandlerList that is defined in the Control Class. This list is stored in a field (variable) named "events" and is only publicly exposed via the read-only property Events.
The class presented below uses reflection to temporarily assign null to the events field effectively removing all event handlers registered for the Control.
While it may be an abuse of the pattern, the class implements the IDisposable Interface to restore the events field on disposal of the class instance. The reason for this is to facilitate the use of the using block to wrap the class usage.
public class ControlEventSuspender : IDisposable
{
private const string eventsFieldName = "events";
private const string headFieldName = "head";
private static System.Reflection.FieldInfo eventsFieldInfo;
private static System.Reflection.FieldInfo headFieldInfo;
private System.Windows.Forms.Control target;
private object eventHandlerList;
private bool disposedValue;
static ControlEventSuspender()
{
Type compType = typeof(System.ComponentModel.Component);
eventsFieldInfo = compType.GetField(eventsFieldName, System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.NonPublic);
headFieldInfo = typeof(System.ComponentModel.EventHandlerList).GetField(headFieldName, System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.NonPublic);
}
private static bool FieldInfosAquired()
{
if (eventsFieldInfo == null)
{
throw new Exception($"{typeof(ControlEventSuspender).Name} could not find the field '{ControlEventSuspender.eventsFieldName}' on type Component.");
}
if (headFieldInfo == null)
{
throw new Exception($"{typeof(ControlEventSuspender).Name} could not find the field '{ControlEventSuspender.headFieldName}' on type System.ComponentModel.EventHandlerList.");
}
return true;
}
private ControlEventSuspender(System.Windows.Forms.Control target) // Force using the the Suspend method to create an instance
{
this.target = target;
this.eventHandlerList = eventsFieldInfo.GetValue(target); // backup event hander list
eventsFieldInfo.SetValue(target, null); // clear event handler list
}
public static ControlEventSuspender Suspend(System.Windows.Forms.Control target)
{
ControlEventSuspender ret = null;
if (FieldInfosAquired() && target != null)
{
ret = new ControlEventSuspender(target);
}
return ret;
}
protected virtual void Dispose(bool disposing)
{
if (!this.disposedValue)
{
if (disposing)
{
if (this.target != null)
{
RestoreEventList();
}
}
}
this.disposedValue = true;
}
public void Dispose()
{
Dispose(true);
}
private void RestoreEventList()
{
object o = eventsFieldInfo.GetValue(target);
if (o != null && headFieldInfo.GetValue(o) != null)
{
throw new Exception($"Events on {target.GetType().Name} (local name: {target.Name}) added while event handling suspended.");
}
else
{
eventsFieldInfo.SetValue(target, eventHandlerList);
eventHandlerList = null;
target = null;
}
}
}
Example usage in the button1_Click method:
public partial class Form1 : Form
{
public Form1()
{
InitializeComponent();
}
private void button1_Click(object sender, EventArgs e)
{
using (ControlEventSuspender.Suspend(comboBox1))
{
comboBox1.SelectedIndex = 3; // SelectedIndexChanged does not fire
}
}
private void button2_Click(object sender, EventArgs e)
{
comboBox1.SelectedIndex = -1; // clear selection, SelectedIndexChanged fires
}
private void button3_Click(object sender, EventArgs e)
{
comboBox1.SelectedIndex = 3; // SelectedIndexChanged fires
}
private void comboBox1_SelectedIndexChanged(object sender, EventArgs e)
{
Console.WriteLine("index changed fired");
System.Media.SystemSounds.Beep.Play();
}
}
SoapBox Diatribe
Many will say that the use of Reflection to access non-public class members is dirty or some other derogatory term and that it introduces a brittleness to the code as someone may change the underlying code definition such that the code that relies on member names (magic strings) is no longer valid. This is a valid concern, but I view it as no different than code that accesses external databases.
Reflection can be thought of a query of a type (datatable) from an assembly (database) for specific fields (members: fields, properties, events). It is no more brittle than a SQL statement such as Select SomeField From SomeTable Where AnotherField=5. This type of SQL code is prevent in the world and no one thinks twice about writing it, but some external force could easily redefine the database you code relies on an render all the magic string SQL statements invalid as well.
Use of hard coded names is always at risk of being made invalid by change. You have to weigh the risks of moving forward versus the option of being frozen in fear of proceeding because someone wants to sound authoritative (typically a parroting of other such individuals) and criticize you for implementing a solution that solves the current problem.

I was hoping to write code that would programatically locate all event handler method names created using controlObject.Event += EventHandlerMethodName, but as you see in the other answers, code to do this is complicated, limited, and perhaps not able to work in all cases
This is what I came up with. It satisfies my desire to consolidate the code that subtracts and re-adds event handler method names into my abstract class, but at the expense of having to write code to store and manage event handler method names and having to write code for each control property where I want to suppress the event handler, modify the property value, and finally re-add the event handler.
public abstract class MyCbo_Abstract : ComboBox
{
// create an event handler property for each event the app has custom code for
[DesignerSerializationVisibility(DesignerSerializationVisibility.Hidden)]
private EventHandler evSelectedValueChanged;
[DesignerSerializationVisibility(DesignerSerializationVisibility.Hidden)]
public EventHandler EvSelectedValueChanged { get => evSelectedValueChanged; set => evSelectedValueChanged = value; }
public MyCbo_Abstract() : base()
{
}
// Create a property that parallels the one that would normally be set in the main body of the program
public object _DataSource_NoEvents
{
get
{
return base.DataSource;
}
set
{
SelectedValueChanged -= EvSelectedValueChanged;
if (value == null)
{
base.DataSource = null;
SelectedValueChanged += EvSelectedValueChanged;
return;
}
string valueTypeName = value.GetType().Name;
if (valueTypeName == "Int32")
{
base.DataSource = null;
SelectedValueChanged += EvSelectedValueChanged;
return;
}
//assume StringCollection
base.DataSource = value;
SelectedValueChanged += EvSelectedValueChanged;
return;
}
}
}
public partial class MyCboFooList : MyCbo_Abstract
{
public MyCboFooList() : base()
{
}
}
Designer has
this.myCboFooList = new MyCboFooList();
Main form code
public partial class FormMain : Form
{
public FormMain()
{
myCboFooList.SelectedValueChanged += OnMyCboFooList_SelectedValueChanged;
myCboFooList.EvSelectedValueChanged = OnMyCboFooList_SelectedValueChanged;
}
private void OnMyCboFooList_SelectedValueChanged(object sender, EventArgs e)
{
// do stuff
}
}
And now, if I want to set a property and suppress event(s), I can write something like the following and not have to remember to re-add the event handler method name
myCboFooList._DataSource_NoEvents = null;

Defining event with field working like Handled

I my class I have an event with arguments containing boolean field Prevent and a default function doing some stuff if Prevent == false. I want user of my class to be able to prevent default function from doing stuff by setting Prevent = true in function listening to event. But it turned out that default function is invoked first, with Prevent = true, and then the user-defined function changes the value of the field. Is there any way of changing order of function invokation? Or any other approach which can give me the same results? I've done it by defining Func field used to check if default function should do its things and user can modify the field providing own function. It works, but I don't like the fact that 2 separate functions are needed, one to do stuff when event fires and one to prevent default behaviour.
Some of the code:
public Func<int, int, bool> PreventFunc = new Func<int,int,bool>( (x,y) => {return false;});
public struct MyEventArgs
{
int x;
int y;
bool Prevent;
public MyEventArgs (int i, int j, bool prev) { /*assignments*/ }
}
public delegate void MyEventDelegate(object sender, MyEventArgs e);
public event MyEventDelegate MyEvent = DefaultBehaviourFunction;
void DefaultBehaviourFunction (object sender, MyEventArgs e)
{
if (e.Prevent == true) return;
//do stuff
}
And when invoking an event
MyEvent.Invoke(sender, new MyEventArgs(i, j, PreventFunc(i,j)));
I want it to work without the PreventFunc, like that:
MyEvent += Func;
void Func (object sender, MyEventArgs e)
{
if (e.x == 1) e.Prevent = true;
//do stuff
}

You could accomplish what you are asking for literally, by getting the invocation list from the event delegate field, reversing the order, and invoking each target individually. But that would be an abuse of the event pattern, as well as simply being unwieldy.
Instead, you should implement this the way that other classes do: have the method that raises the event be the one that checks the Prevent value, not some other handler.
For example:
private void OnMyEvent(int i, int j, bool prevent)
{
MyEventDelegate handler = MyEvent;
MyEventArgs e = new MyEventArgs(i, j, prevent);
if (handler != null)
{
handler(this, e);
}
if (e.Prevent)
{
return;
}
// Member of same class, and we already know Prevent is false,
// so just need to pass i and j.
DefaultBehaviourFunction(i, j);
}
then instead of invoking the event handlers directly with MyEvent.Invoke() you'd call the OnMyEvent() method instead.

Adding events to a class using +=

Please forgive my little knowledge!
I have the following class in HIDNewDeviceEventMonitor.cs:
public class HIDNewDeviceEventMonitor : IDisposable
{
// used for monitoring plugging and unplugging of USB devices.
private ManagementEventWatcher watcherAttach;
public HIDNewDeviceEventMonitor()
{
// Catch USB HID plugged instance event watching
watcherAttach = new ManagementEventWatcher();
watcherAttach.EventArrived += new EventArrivedEventHandler(watcher_EventArrived);
watcherAttach.Query = new WqlEventQuery(#"SELECT * FROM __InstanceCreationEvent WITHIN 2 WHERE TargetInstance ISA 'Win32_PNPEntity' AND TargetInstance.DeviceID LIKE 'HID\\VID_04D8%'");
watcherAttach.Start();
}
void watcher_EventArrived(object sender, EventArrivedEventArgs e)
{
Debug.WriteLine("my device is inserted..");
}
public void Dispose()
{
watcherAttach.Stop();
watcherAttach.Dispose();
}
~HIDNewDeviceEventMonitor()
{
this.Dispose();
}
}
Now, how can I change this class to be able to add an event handler that the class can call from within watcher_EventArrived where someNewEvent is outside the class file, actually in the form.cs:
// code in the form
HIDNewDeviceEventMonitor ok = new HIDNewDeviceEventMonitor();
ok.Inserted += someNewEvent; // <-- my problem, I don't know how to add an event to the class this way
private void someNewEvent()
{
//Enumerate and add to listbox1
}
private void Form1_FormClosing(object sender, FormClosingEventArgs e)
{
ok.Dispose();
}
I 've seen this thing with other classes, how can I make my class like that?

Your Inserted event should look like this:
public event EventHandler Inserted;
You invoke it like this:
private void OnInserted()
{
if (this.Inserted != null)
{
this.Inserted(this, EventArgs.Empty);
}
}
The signature for the event handler is this:
void someNewEvent(object sender, EventArgs e)
{
//
}
Then you should wrap that code in the constructor of the class:
HIDNewDeviceEventMonitor ok;
public ClassName()
{
ok = new HIDNewDeviceEventMonitor();
ok.Inserted += someNewEvent; // <-- my problem
}
Declare the ok variable outside the constructor, and instantiate it inside. Then add the event handler.
Pro tip: You could use the generic EventHandler<T> if you need to supply a custom implementation of e.

Simply put, you're trying to add events to your HIDNewDeviceMonitor class.
To do this, first you'll need to define a delegate.
public delegate void InsertedHandler;
Next, you'll need to define the event in your HIDNewDeviceMonitor class.
// Notice how the event uses the delegate that's been defined
// v
public event InsertedHandler Inserted;
Now you'll need something that "fires" the event, which could easily be put in your watcher_EventArrived method.
void watcher_EventArrived(object sender, EventArrivedEventArgs e)
{
Debug.WriteLine("my device is inserted..");
// Notice how we check the event handler for null.
// If you don't, it could throw a NullReferenceException.
// Irritating, but easy to debug.. Usually..
if (Inserted != null)
Inserted(); // Or whatever parameters you need.
}
We're all done with the HIDNewDeviceMonitor class.
Now whatever class that uses the HIDNewDeviceMonitor can use the EventHandler code that you provided.
However, it'll have to be the same delegate.
public class MyClass
{
HIDNewDeviceMonitor monitor;
public MyClass()
{
monitor = new HIDNewDeviceMonitor();
monitor.Inserted += DeviceInserted;
}
private void DeviceInserted()
{
// Execute code here
}
}

You need to do following in the HIDNewDeviceEventMonitor class:
1.) First define a public event inside the class like this-
public event EventHandler Inserted;
2.) Then fire this event within the code where you detect the changes in events. Like this-
if(Inserted != null)
Inserted(this,null);
The if condition checks if the event is registered by any listener. It's fired in case it is.
Hope this helps.

Unsubscribing from anonymous event handler inside an static method (extension method)

I have an extension method to subscribe a PropertyChanged event of an object that implements INotifyPropertyChanged.
I would like that the event fires just once. Not more.
This is my method.
public static void OnPropertyChanged<T>(this INotifyPropertyChanged target, string propertyName, Action action)
{
if (target == null)
{
return;
}
PropertyChangedEventHandler handler = (obj, e) =>
{
if (propertyName == e.PropertyName)
{
action();
}
};
target.PropertyChanged -= handler;
target.PropertyChanged += handler;
}
But it does not work. I cannnot remove the event handler so the event fires every time I call this method.
I have try a different approach. Instead of using annonymous methods, something more traditional, like this:
public static void OnPropertyChanged<T>(this INotifyPropertyChanged target, string propertyName, Action action)
{
if (target == null)
{
return;
}
target.PropertyChanged -= target_PropertyChanged;
target.PropertyChanged += target_PropertyChanged;
}
static void target_PropertyChanged(object sender, PropertyChangedEventArgs e)
{
//do stuff here
}
And it just works fine. The event fires just once, but I also need the Action parameter. I cannot use it with this approach.
Any workaround or different aproach to solve this issue?Is there something strange with anonymous methods inside static methods?
Thanks in advance.

That is a limitation of using anonymous methods as event handlers. They cannot be removed as you would a normal method (which is technically a delegate instance automatically create via a method group conversion) because anonymous methods get compiled into a compiler-generated container class and a new instance of the class is created each time.
In order to preserve the action parameter you could create a container class which would have the delegate for your event handler inside. The class can be declared private inside the of the other class you're working with - or made internal, maybe in a "Helpers" namespace. It would look something like this:
class DelegateContainer
{
public DelegateContainer(Action theAction, string propName)
{
TheAction = theAction;
PopertyName = propName;
}
public Action TheAction { get; private set; }
public string PropertyName { get; private set; }
public void PropertyChangedHandler(object sender, PropertyChangedEventArgs e)
{
if(PropertyName == e.PropertyName)
TheAction();
}
}
Then, create and store the reference to the container in your class. You might create a static member currentContainer and then set the handler like this:
private static DelegateContainer currentContainer;
public static void OnPropertyChanged<T>(this INotifyPropertyChanged target, string propertyName, Action action)
{
if (target == null)
{
return;
}
if(currentContainer != null)
target.PropertyChanged -= currentContainer.PropertyChangedHandler;
currentContainer = new DelegateContainer(action, propertyName);
target.PropertyChanged += currentContainer.PropertyChangedHandler;
}

You can get your first example to work if you unsubscribe from within the event handler itself.
public static void OnPropertyChanged<T>(this INotifyPropertyChanged target, string propertyName, Action action)
{
if (target == null)
{
return;
}
// Declare the handler first, in order to create
// a concrete reference that you can use from within
// the delegate
PropertyChangedEventHandler handler = null;
handler = (obj, e) =>
{
if (propertyName == e.PropertyName)
{
obj.PropertyChanged -= handler; //un-register yourself
action();
}
};
target.PropertyChanged += handler;
}
The above code serves as a "one and done" event handler. You can register an unlimited number of these, and each one will only be executed once before unregistering itself.
Keep in mind that it's possible to have one of these handlers execute multiple times, if you raise the event across multiple threads in short succession. To prevent this, you might need to create a static Dictionary(T,T) mapping object instances to "lock objects," and add some sentry code to ensure that a handler is only executed once. Those implementation specifics seem to be a bit outside the scope of your question as currently written, however.

Technically, it's not the same anonymous method you are trying to unsubscribe. .NET creates new instance of that method every time your OnPropertyChanged called. That's why unsubscription will not work.

Return a value from an Event -- is there a Good Practice for this?

I'm doing a small multi-threaded app that uses asynchronous TCP sockets, but I will get to the point: I'm using a custom event to read a value from a form and the delegate used by the event returns a string when finished.
My question here is: is that correct? is it OK to return values from the events? or is there a better way to do this? (like using a simple delegate to the form to read the values)

It's often awkward to return values from events. In practice, I've found it much easier to include a writable property on a set of custom EventArgs that is passed to the event, and then checked after the event fires -- similar to Cancel property of the WinForms FormClosing event.

I don't think it's a good idea... events are basically multicast delegates, so there can be multiple handlers. Which return value will you take in that case ?

I know this is ages after the post but thought of adding comment with code to explain Dustin Campbell answer for if someone else comes across this thread. I came across this post while trying to decide what would be best practice and this is what is meant by the answer.
Create your own custom event handler class
public class myCustomeEventArgs:EventArgs
{
public bool DoOverride { get; set; }
public string Variable1 { get; private set; }
public string Variable2{ get; private set; }
public myCustomeEventArgs(string variable1 , string variable2 )
{
DoOverride = false;
Variable1 = variable1 ;
Variables = variable2 ;
}
}
So when you create your event delegate you use your created event args like this.
public delegate void myCustomeEventHandler(object sender, myCustomeEventArgs e);
And in the class raising the event you declare the event.
public event myCustomeEventHandler myCustomeEvent;
So when you trigger the event in your class the class that listens for the event you can just in the body of the event set e.DoOverride = true; as it will be declared in the class firing the event.
Fire event for example:
if(myCustomeEvent != null)
{
var eventArgs = new myCustomeEventArgs("Some Variable", "Another Varaible");
myCustomeEvent(this, eventArgs);
//Here you can now with the return of the event work with the event args
if(eventArgs.DoOverride)
{
//Do Something
}
}

The closest example I can think of is the FormClosing event in WinForms. It lets the form cancel the event by setting the eventArgs.Cancel property to true. For you to do something similar, you would define your own event args class with the return value as a property on that class. Then pass an event args object whenever you raise the event. Whoever raised the event can inspect the event args object for the return value. Others who are receiving the event can also inspect or change the event args object.
Update: I just ran across the AppDomain.AssemblyResolve event, and it appears to be an event that returns a value. It seems you just need to declare a delegate type that returns a value, and then define your event with that delegate type. I haven't tried creating my own event like this, though. One advantage to using a property on the event argument is that all subscribers to the event can see what previous subscribers have returned.

Note: only the last event returns the result.
class Program
{
static event Func<string, bool> TheEvent;
static void Main(string[] args)
{
TheEvent += new Func<string, bool>(Program_TheEvent);
TheEvent +=new Func<string,bool>(Program_TheEvent2);
TheEvent += new Func<string, bool>(Program_TheEvent3);
var r = TheEvent("s"); //r == flase (Program_TheEvent3)
}
static bool Program_TheEvent(string arg)
{
return true;
}
static bool Program_TheEvent2(string arg)
{
return true;
}
static bool Program_TheEvent3(string arg)
{
return false;
}
}

I don't know if this is best practice but i did it this way.
Func<DataRow, bool> IsDataValid;
// some other code ....
isValid = true;
if (IsDataValid != null)
{
foreach (Func<DataRow, bool> func in IsDataValid.GetInvocationList())
{
isValid &= func(Row);
}
}

If event returns a value and there are multiple handlers registered the event returns the result value of the last called handler.
Look for an example at http://blogs.msdn.com/b/deviations/archive/2008/11/27/event-handlers-returning-values.aspx

I looped over the properties of the EventArgs like this and pulled out its X and Y values.
private void navBarControl1_Click(object sender, EventArgs e)
{
int _x = 0;
int _y = 0;
Type t = e.GetType();
IList<PropertyInfo> props = new List<PropertyInfo>(t.GetProperties());
foreach (PropertyInfo prop in props)
{
if (prop.Name == "X")
{
object propValue = prop.GetValue(e, null);
_x = Convert.ToInt32(propValue);
}
if (prop.Name == "Y")
{
object propValue = prop.GetValue(e, null);
_y = Convert.ToInt32(propValue);
}
}

void method()
{
list<string> strings = new list<string>();
dostuff += stuff;
dostuff += stuff;
dostuff(this, new EventHandlerArgs(){ Parameter = strings })
foreach(string currString in strings)
{
//....
}
}
void stuff(object sender, EventHandlerArgs e)
{
list<string> strings = e.Parameter as list<string>;
if (strings != null)
{
strings.Add(MyString)
}
}