Suppose I have two threads. On one thread, I'm performing a sequence of jobs. On the other thread I'm waiting for user input to control those jobs (i.e., skip to the next job). For example (pseudo-C#):
public class Worker {
public List<Job> Jobs { get; set; }
public Worker(Controller anotherThread) {
anotherThread.SkipJobRequested += OnSkipJobRequested;
}
public DoWork() {
foreach (Job job in Jobs) {
// Do various work...
}
}
// Event that fires on Controller thread
public void OnSkipJobRequsted(Object sender, EventArgs args) {
// Somehow skip to the next job
}
}
I'm uncertain how I should handle skipping to the next job. One possibility that occurs to me would be to have an instance variable (like, IsSkipRequested) that I set when a skip is requested, and check it at various junctures within DoWork().
Do any other patterns exist that I could use to handle this event?
Another pattern is the .Net class BackgroundWorker, which seems like it should suit your purpose. You could make Job a subclass of BackgroundWorker, and cycle through. The difference is that BackgroundWorker doesn't know about your UI thread, only whether cancellation has been requested.
In this pattern the UI thread would call CancelAsync, then your DoWork method would check the CancellationPending at convenient intervals to decide whether or not to proceed. You would call RunWorkerAsync for the next job inside the RunWorkerCompleted event handler.
Another suggestion would be (and you have mentioned this above) if you made List<Jobs> a Queue<Jobs>, and rather than DoWork performing a foreach maintained the current executing job in your object, you could simply dequeue each item when ready for processing.
The task parallel library allows you to specify a cancellation token (although you would probably have to pass this to your job execution code, and handle cancellation in there), which you can call RequestCancellation on when skip is pressed, and start the next job from the queue. Additionally, when firing off a new task you can specify an action to perform on completion, this would allow you to chain your tasks together into a sequential order and skip tasks when required. Below is an example without a cancellation token:
_currentJob = Jobs.DeQueue();
Task.Factory.StartNew(() => {_currentJob.execute();},
() =>
{
//On task completion logic
ExecuteNextJobFromQueue();
}
Be aware that this approach would probably work best if Job is performing multiple tasks and not one big blocking task as you would need to check for cancellation during job execution.
Related
In my scenario I am implementing an IObserver interface to poll for updates to a specific variable through a method. I have one thread that is created just to execute this method. This thread for now should never quit looking for updates to some data.
public static void findUpdates()
{
while(true)
{
CheckForUpdates(); //registered to the observer
}
}
I then have multiple threads coming into a class method to receive updates.
TaskCompletionSource<bool>updateHappened = new TaskCompletionSource<bool>();
object updatedValue;
public async Task receiveUpdates(){
while(true)
{
await updateHappened.Task;
//check to see if the thread cares about the updated value or not
//update or do not update and possibly terminate
}
}
and in the OnNext implementation
public void OnNext(Object value){
updatedHappened.SetResult(true);
}
The reason I need to await an event is because when I have threads enter my while() loop is receiveUpdates if they are not awaited then there ends up being a max number of threads that can be scheduled and some threads get blocked from executing until another exits.
This is not behaving properly right now and it I can never get past the await updateHappened.Task; line no matter how many threads I create and send through to receive updates.
I would like the ability to have threads who want to receive updates come through, suspend execution in the threadpool waiting to be executed so that other threads (however many) can also try and receive updates, and then when a update happens resume each thread one at a time and update it or not before suspending it again. The one thread will always poll for updates using the IObserver interface but every other thread trying to receive updates will always continue to look for updates unless it exits under a condition.
A TaskCompletionSource<T> can only be completed once; it's a single-shot signal, and that's it.
Since you're already using Rx, consider having the source thread expose its observable with Publish and Connect/RefCount. This will permit multiple subscribers to receive the same data. Then for the threads that just want to await, use await sequence.FirstAsync().
Consider two classes; Producer and Consumer (the same as classical pattern, each with their own threads). Is it possible for Producer to have an Event which Consumer can register to and when the producer triggers the event, the consumer's event handler is run in its own thread? Here are my assumptions:
Consumer does not know if the Producer's event is triggered
within his own thread or another.
Neither Producer nor Consumer are descendants of Control so they don't have
BeginInvoke method inherited.
PS. I'm not trying to implement Producer - Consumer pattern. These are two simple classes which I'm trying to refactor the producer so it incorporates threads.
[UPDATE]
To further expand my problem, I'm trying to wrap a hardware driver to be worked with in the simplest way possible. For instance my wrapper will have a StateChanged event which the main application will register to so it will be notified when hardware is disconnected. As the actual driver has no means other than polling to check its presence , I will need to start a thread to check it periodically. Once it is not available anymore I will trigger the event which needs to be executed in the same thread as it was added. I know this is a classical Producer-Consumer pattern but since I'm trying to simplify using my driver-wrapper, I don't want the user code to implement consumer.
[UPDATE]
Due to some comments suggesting that there's no solution to this problem, I would like to add few lines which might change their minds. Considering the BeginInvoke can do what I want, so it shouldn't be impossible (at least in theory). Implementing my own BeginInvoke and calling it within the Producer is one way to look at it. It's just that I don't know how BeginInvoke does it!
You want to do inter thread communication. Yes it is possible.
Use System.Windows.Threading.Dispatcher
http://msdn.microsoft.com/en-us/library/system.windows.threading.dispatcher.aspx
The Dispatcher maintains a prioritized queue of work items for a specific thread.
When a Dispatcher is created on a thread, it becomes the only Dispatcher that can be associated with the thread, even if the Dispatcher is shut down.
If you attempt to get the CurrentDispatcher for the current thread and a Dispatcher is not associated with the thread, a Dispatcher will be created. A Dispatcher is also created when you create a DispatcherObject. If you create a Dispatcher on a background thread, be sure to shut down the dispatcher before exiting the thread.
Yes there is a way to do this. It relies on using the SynchronizationContext class (docs). The sync context abstracts the operations of sending messages from one thread to another via the methods Send (synchronous for the calling thread) and Post(async for the calling thread).
Let's take a slightly simpler situation where you only want the capture one sync context, the context of the "creator" thread. You would do something like this:
using System.Threading;
class HardwareEvents
{
private SynchronizationContext context;
private Timer timer;
public HardwareEvents()
{
context = SynchronizationContext.Current ?? new SynchronizationContext();
timer = new Timer(TimerMethod, null, 0, 1000); // start immediately, 1 sec interval.
}
private void TimerMethod(object state)
{
bool hardwareStateChanged = GetHardwareState();
if (hardwareStateChanged)
context.Post(s => StateChanged(this, EventArgs.Empty), null);
}
public event EventHandler StateChanged;
private bool GetHardwareState()
{
// do something to get the state here.
return true;
}
}
Now, the creating thread's sync context will be used when events are invoked. If the creating thread was a UI thread it will have a sync context supplied by the framework. If there is no sync context, then the default implementation is used, which invokes on the thread pool. SynchronizationContext is a class that you can subclass if you want to provide a custom way to send a message from the producer to the consumer thread. Just override Post and Send to send said message.
If you wanted every event subscriber to get called back on their own thread, you would have to capture the sync context in the add method. You then hold on to pairs of sync contexts and delegates. Then when raising the event, you would loop through the sync context / delegate pairs and Post each one in turn.
There are several other ways you could improve this. For example, you may want to suspend polling the hardware if there no subscribers to the event. Or you might want to back off your polling frequency if the hardware does not respond.
First, please note that in .NET / the Base Class Library, it is usually the event subscriber's obligation to ensure that its callback code is executing on the correct thread. That makes it easy for the event producer: it may just trigger its event without having to care about any thread affinities of its various subscribers.
Here's a complete example step-by-step of a possible implementation.
Let's start with something simple: The Producer class and its event, Event. My example won't include how and when this event gets triggered:
class Producer
{
public event EventHandler Event; // raised e.g. with `Event(this, EventArgs.Empty);`
}
Next, we want to be able to subscribe our Consumer instances to this event and be called back on a specific thread (I'll call this kind of thread a "worker thread"):
class Consumer
{
public void SubscribeToEventOf(Producer producer, WorkerThread targetWorkerThread) {…}
}
How do we implement this?
First, we need the means to "send" code to a specific worker thread. Since there is no way to force a thread to execute a particular method whenever you want it to, you must arrange for a worker thread to explicitly wait for work items. One way to do this is via a work item queue. Here's a possible implementation for WorkerThread:
sealed class WorkerThread
{
public WorkerThread()
{
this.workItems = new Queue<Action>();
this.workItemAvailable = new AutoResetEvent(initialState: false);
new Thread(ProcessWorkItems) { IsBackground = true }.Start();
}
readonly Queue<Action> workItems;
readonly AutoResetEvent workItemAvailable;
public void QueueWorkItem(Action workItem)
{
lock (workItems) // this is not extensively tested btw.
{
workItems.Enqueue(workItem);
}
workItemAvailable.Set();
}
void ProcessWorkItems()
{
for (;;)
{
workItemAvailable.WaitOne();
Action workItem;
lock (workItems) // dito, not extensively tested.
{
workItem = workItems.Dequeue();
if (workItems.Count > 0) workItemAvailable.Set();
}
workItem.Invoke();
}
}
}
This class basically starts a thread, and puts it in an infinite loop that falls asleep (WaitOne) until an item arrives in its queue (workItems). Once that happens, the item — an Action — is dequeued and invoked. Then the thread goes to sleep again (WaitOne)) until another item is available in the queue.
Actions are put in the queue via the QueueWorkItem method. So essentially we can now send code to be executed to a specific WorkerThread instance by calling that method. We're now ready to implement Customer.SubscribeToEventOf:
class Consumer
{
public void SubscribeToEventOf(Producer producer, WorkerThread targetWorkerThread)
{
producer.Event += delegate(object sender, EventArgs e)
{
targetWorkerThread.QueueWorkItem(() => OnEvent(sender, e));
};
}
protected virtual void OnEvent(object sender, EventArgs e)
{
// this code is executed on the worker thread(s) passed to `Subscribe…`.
}
}
Voilà!
P.S. (not discussed in detail): As an add-on, you could package the method of sending code to WorkerThread using a standard .NET mechanism called a SynchronizationContext:
sealed class WorkerThreadSynchronizationContext : SynchronizationContext
{
public WorkerThreadSynchronizationContext(WorkerThread workerThread)
{
this.workerThread = workerThread;
}
private readonly WorkerThread workerThread;
public override void Post(SendOrPostCallback d, object state)
{
workerThread.QueueWorkItem(() => d(state));
}
// other overrides for `Send` etc. omitted
}
And at the beginning of WorkerThread.ProcessWorkItems, you'd set the synchronization context for that particular thread as follows:
SynchronizationContext.SetSynchronizationContext(
new WorkerThreadSynchronizationContext(this));
I posted earlier that I've been there, and that there is no nice solution.
However, I just stumbled upon something I have done in another context before: you could instantiate a timer (that is, Windows.Forms.Timer) when you create your wrapper object. This timer will post all Tick events to the ui thread.
Now if you're device polling logic is non-blocking and fast, you could implement it directly inside the timer Tick event, and raise your custom event there.
Otherwise, you could continue to do the polling logic inside a thread, and instead of firing the event inside the thread, you just flip some boolean variable which gets read by the timer every 10 ms, who then fires the event.
Note that this solution still requires that the object is created from the GUI thread, but at least the user of the object will not have to worry about Invoke.
It is possible. One typical approach is to use the BlockingCollection class. This data structure works like a normal queue except that the dequeue operation blocks the calling thread if the queue is empty. The produce will queue items by calling Add and the consumer will dequeue them by calling Take. The consumer typically runs it's own dedicated thread spinning an infinite loop waiting for items to appear in the queue. This is, more or less, how the message loop on the UI thread operates and is the basis for getting the Invoke and BeginInvoke operations to accomplish the marshaling behavior.
public class Consumer
{
private BlockingCollection<Action> queue = new BlockingCollection<Action>();
public Consumer()
{
var thread = new Thread(
() =>
{
while (true)
{
Action method = queue.Take();
method();
}
});
thread.Start();
}
public void BeginInvoke(Action method)
{
queue.Add(item);
}
}
I have a main thread that starts a few worker threads through a class that implements the Event-based Asynchronous Pattern (http://msdn.microsoft.com/en-us/library/ms228969.aspx). Unfortunately, as soon as the main thread is done setting up the worker threads, it exits.
I want the main thread to wait indefinitely until the anEventHandler method is called, then process that method. The asyncevents listen for events indefinitely until canceled, so they might call the anEventHandler method multiple times before completing.
NOTE: There is no related GUI, I'm just using this pattern.
public static void Main(String[] args) {
AsyncEventClass asyncevent = new AsyncEventClass();
// setup event handling
asyncevent.Events += new EventsHandler(anEventHandler);
// start event monitoring threads
asyncevent.monitorAsync(1);
asyncevent.monitorAsync(2);
asyncevent.monitorAsync(3);
System.Console.WriteLine("The main thread now ends. :(");
}
All the examples online use System.Console.ReadLine(), but that seems like a terrible idea in production code.
You can use a simple AutoResetEvent which you wait for in the main method, and set in the event handler.
If you want to wait for a specific amount of times that the eventhandler is called, I guess you can keep a counter, and set the auto-reset-event only when you reach a certain threshold.
I ended up putting all of the IAsyncResutl objects returned by calling WorkerEventHandler.BeginInvoke into a list, and then exposing that list to the Main method. There, I setup all the asyncevents I need to listen for, then use WaitHandle.WaitAll to wait for the events to complete.
I have a class that implements the Begin/End Invocation pattern where I initially used ThreadPool.QueueUserWorkItem() to thread my work. The work done on the thread doesn't loop but does takes a bit of time to process so the work itself is not easily stopped.
I now have the side effect where someone using my class is calling the Begin (with callback) a ton of times to do a lot of processing so ThreadPool.QueueUserWorkItem is creating a ton of threads to do the processing. That in itself isn't bad but there are instances where they want to abandon the processing and start a new process but they are forced to wait for their first request to finish.
Since ThreadPool.QueueUseWorkItem() doesn't allow me to cancel the threads I am trying to come up with a better way to queue up the work and maybe use an explicit FlushQueue() method in my class to allow the caller to abandon work in my queue.
Anyone have any suggestion on a threading pattern that fits my needs?
Edit: I'm currently targeting the 2.0 framework. I'm currently thinking that a Consumer/Producer queue might work. Does anyone have thoughts on the idea of flushing the queue?
Edit 2 Problem Clarification:
Since I'm using the Begin/End pattern in my class every time the caller uses the Begin with callback I create a whole new thread on the thread pool. This call does a very small amount of processing and is not where I want to cancel. It's the uncompleted jobs in the queue I wish to stop.
The fact that the ThreadPool will create 250 threads per processor by default means if you ask the ThreadPool to queue a large amount of items with QueueUserWorkItem() you end up creating a huge amount of concurrent threads that you have no way of stopping.
The caller is able to push the CPU to 100% with not only the work but the creation of the work because of the way I queued the threads.
I was thinking by using the Producer/Consumer pattern I could queue these threads into my own queue that would allow me to moderate how many threads I create to avoid the CPU spike creating all the concurrent threads. And that I might be able to allow the caller of my class to flush all the jobs in the queue when they are abandoning the requests.
I am currently trying to implement this myself but figured SO was a good place to have someone say look at this code or you won't be able to flush because of this or flushing isn't the right term you mean this.
EDIT My answer does not apply since OP is using 2.0. Leaving up and switching to CW for anyone who reads this question and using 4.0
If you are using C# 4.0, or can take a depedency on one of the earlier version of the parallel frameworks, you can use their built-in cancellation support. It's not as easy as cancelling a thread but the framework is much more reliable (cancelling a thread is very attractive but also very dangerous).
Reed did an excellent article on this you should take a look at
http://reedcopsey.com/2010/02/17/parallelism-in-net-part-10-cancellation-in-plinq-and-the-parallel-class/
A method I've used in the past, though it's certainly not a best practice is to dedicate a class instance to each thread, and have an abort flag on the class. Then create a ThrowIfAborting method on the class that is called periodically from the thread (particularly if the thread's running a loop, just call it every iteration). If the flag has been set, ThrowIfAborting will simply throw an exception, which is caught in the main method for the thread. Just make sure to clean up your resources as you're aborting.
You could extend the Begin/End pattern to become the Begin/Cancel/End pattern. The Cancel method could set a cancel flag that the worker thread polls periodically. When the worker thread detects a cancel request, it can stop its work, clean-up resources as needed, and report that the operation was canceled as part of the End arguments.
I've solved what I believe to be your exact problem by using a wrapper class around 1+ BackgroundWorker instances.
Unfortunately, I'm not able to post my entire class, but here's the basic concept along with it's limitations.
Usage:
You simply create an instance and call RunOrReplace(...) when you want to cancel your old worker and start a new one. If the old worker was busy, it is asked to cancel and then another worker is used to immediately execute your request.
public class BackgroundWorkerReplaceable : IDisposable
{
BackgroupWorker activeWorker = null;
object activeWorkerSyncRoot = new object();
List<BackgroupWorker> workerPool = new List<BackgroupWorker>();
DoWorkEventHandler doWork;
RunWorkerCompletedEventHandler runWorkerCompleted;
public bool IsBusy
{
get { return activeWorker != null ? activeWorker.IsBusy; : false }
}
public BackgroundWorkerReplaceable(DoWorkEventHandler doWork, RunWorkerCompletedEventHandler runWorkerCompleted)
{
this.doWork = doWork;
this.runWorkerCompleted = runWorkerCompleted;
ResetActiveWorker();
}
public void RunOrReplace(Object param, ...) // Overloads could include ProgressChangedEventHandler and other stuff
{
try
{
lock(activeWorkerSyncRoot)
{
if(activeWorker.IsBusy)
{
ResetActiveWorker();
}
// This works because if IsBusy was false above, there is no way for it to become true without another thread obtaining a lock
if(!activeWorker.IsBusy)
{
// Optionally handle ProgressChangedEventHandler and other features (under the lock!)
// Work on this new param
activeWorker.RunWorkerAsync(param);
}
else
{ // This should never happen since we create new workers when there's none available!
throw new LogicException(...); // assert or similar
}
}
}
catch(...) // InvalidOperationException and Exception
{ // In my experience, it's safe to just show the user an error and ignore these, but that's going to depend on what you use this for and where you want the exception handling to be
}
}
public void Cancel()
{
ResetActiveWorker();
}
public void Dispose()
{ // You should implement a proper Dispose/Finalizer pattern
if(activeWorker != null)
{
activeWorker.CancelAsync();
}
foreach(BackgroundWorker worker in workerPool)
{
worker.CancelAsync();
worker.Dispose();
// perhaps use a for loop instead so you can set worker to null? This might help the GC, but it's probably not needed
}
}
void ResetActiveWorker()
{
lock(activeWorkerSyncRoot)
{
if(activeWorker == null)
{
activeWorker = GetAvailableWorker();
}
else if(activeWorker.IsBusy)
{ // Current worker is busy - issue a cancel and set another active worker
activeWorker.CancelAsync(); // Make sure WorkerSupportsCancellation must be set to true [Link9372]
// Optionally handle ProgressEventHandler -=
activeWorker = GetAvailableWorker(); // Ensure that the activeWorker is available
}
//else - do nothing, activeWorker is already ready for work!
}
}
BackgroupdWorker GetAvailableWorker()
{
// Loop through workerPool and return a worker if IsBusy is false
// if the loop exits without returning...
if(activeWorker != null)
{
workerPool.Add(activeWorker); // Save the old worker for possible future use
}
return GenerateNewWorker();
}
BackgroundWorker GenerateNewWorker()
{
BackgroundWorker worker = new BackgroundWorker();
worker.WorkerSupportsCancellation = true; // [Link9372]
//worker.WorkerReportsProgress
worker.DoWork += doWork;
worker.RunWorkerCompleted += runWorkerCompleted;
// Other stuff
return worker;
}
} // class
Pro/Con:
This has the benefit of having a very low delay in starting your new execution, since new threads don't have to wait for old ones to finish.
This comes at the cost of a theoretical never-ending growth of BackgroundWorker objects that never get GC'd. However, in practice the code below attempts to recycle old workers so you shouldn't normally encounter a large pool of ideal threads. If you are worried about this because of how you plan to use this class, you could implement a Timer which fires a CleanUpExcessWorkers(...) method, or have ResetActiveWorker() do this cleanup (at the cost of a longer RunOrReplace(...) delay).
The main cost from using this is precisely why it's beneficial - it doesn't wait for the previous thread to exit, so for example, if DoWork is performing a database call and you execute RunOrReplace(...) 10 times in rapid succession, the database call might not be immediately canceled when the thread is - so you'll have 10 queries running, making all of them slow! This generally tends to work fine with Oracle, causing only minor delays, but I do not have experiences with other databases (to speed up the cleanup, I have the canceled worker tell Oracle to cancel the command). Proper use of the EventArgs described below mostly solves this.
Another minor cost is that whatever code this BackgroundWorker is performing must be compatible with this concept - it must be able to safely recover from being canceled. The DoWorkEventArgs and RunWorkerCompletedEventArgs have a Cancel/Cancelled property which you should use. For example, if you do Database calls in the DoWork method (mainly what I use this class for), you need to make sure you periodically check these properties and take perform the appropriate clean-up.
Is there an elegant way to know when a worker thread is done executing so I can access resources it produced?
For example if the worker thread queried a list of SQL Servers using
ServersSqlDataSourceEnumerator.Instance.GetDataSources();
and saved the result in a DataTable variable, what mechanism can I use to know when this DataTable variable has been populated/is available. I don't want to poll ThreadState; it would be ideal to fire an event when it's done so I can perform actions with the result.
Thanks!
You can use a callback mechanism or block on an event to know of completion of an Async operation. See this page for the Asychronous Programming Model in .net - you can call BeginInvoke on any delegate to perform the action in an Async manner.
If you're using the BackgroundWorker type, you can subscribe to the RunWorkerCompleted event.
So fire an event :-P
You could also look at using an AutoResetEvent:
http://msdn.microsoft.com/en-us/library/system.threading.autoresetevent.aspx
What I do in this instance is get the WorkerThread to call a function after it has completed the work, which will invoke the the UI Thread, which can do the work in which you require.
E.g.
private void SetWorkerThreadToDoWork()
{
WorkerThread.Start();
}
private void MyWorkerThreadWork()
{
//This will be on the WorkerThread (called from WorkerThread.Start())
DoWorkFunc();
WorkComplete();
}
private void WorkComplete()
{
if(InvokeRequired == true)
{
//Do the invoke
}
else
{
//Check work done by worker thread
//e.g. ServersSqlDataSourceEnumerator.Instance.GetDataSources();
}
}
If it's a simple process you're using, I'd go for a BackgroundWorkerThread, this comes with it's own events that are fired when work is complete. But if you require to use a Thread, I would either look in to Asynchronous Callbacks or a similar route to that shown above.
You can check my answer on this SO thread
It uses a call back mechanism. When the async operation is done, it will fire the callback method where you can handle the processing that needs to be done post SQL execution.
Use a similar approach to be notified when the asynchronous operation is done.
Hope this helps :)
I don't program in C# but here's what I did with Delphi, maybe you can do it as well with C#.
I have a TThread descendant, and in the "destroy" event I send a message to its creator saying "hey I'm about to die !".
This way its parent (which is the main thread) creates a new one if it needs a new one. To be precise it launches a timer that, when fired, creates a new thread if a new one is needed (sites sucking time (lol) !!).