from("direct:A")
.process(//processing here)
.recipientList(//expression that return two recipients [direct:B, direct:C] )
from("direct:B")
.process(//processing here)...
from("direct:C")
.process(//processing here)...
.from("direct:A") behaves like a java method i.e the thread that calls it will continue to process().
So what will happen in above case ?
Let say Thread t1 calls from("direct:A") then
t1 will continue to process()
and then t1 will enter into recipientList()
Now from here on-wards will t1 call from("direct:B") and then call from("direct:C") synchronously
or
direct:b and direct:c will be called in two new thread asynchronously.
Read the recipient list documentation for a lot more detail. By default it processes messages synchronously. You can use the parallel processing feature of the recipient list to run this concurrently. You can also define your own thread pools.
Go read the documentation it is in there.
Related
I found this article explain about non-blocking IO Play framework: https://engineering.linkedin.com/play/play-framework-async-io-without-thread-pool-and-callback-hell
Code example in that article
object ProxyController extends Controller {
def proxy = Action {
val responseFuture: Future[Response] = WS.url("http://example.com").get()
Logger.info("Before map")
val resultFuture: Future[Result] = responseFuture.map { resp =>
Logger.info("Within map")
// Create a Result that uses the http status, body, and content-type
// from the example.com Response
Status(resp.status)(resp.body).as(resp.ahcResponse.getContentType)
}
Logger.info("After map")
Async(resultFuture)
}
The explaination from that article:
Under the hood, Play uses a thread pool sized to one thread per CPU
core. One of these scarce threads, T1, executes the proxy action,
running through the code from top to bottom, except the contents of
the function passed to the map method, since that depends on a
non-blocking I/O call that has not yet completed. Once T1 returns the
AsyncResult, it moves on to process other requests. Later on, when the
response from example.com is finally available, another thread T2
(which may or may not be the same as T1) executes the function passed
to the map method. At no point were either of the threads blocked
waiting on the response from example.com.
I don't understand the highlighted in that paragraph. T1 has returned to thread pool and how the application keep track and receive the response from example.com then submit thread T2 to execute map function after that.
Please someone explain me this.
One of these scarce threads, T1, executes the proxy action, running through the code from top to bottom, except the contents of the function passed to the map method, since that depends on a non-blocking I/O call that has not yet completed.
At this point an object called a future has been created.
Once T1 returns the AsyncResult, it moves on to process other requests.
The future sits in memory waiting to be fulfilled when some IO comes in. Meanwhile the thread T1 can process other requests.
Later on, when the response from example.com is finally available,
When the response is available this will call some internal Play code which reads the response and changes the state of the future to give it a value. The moment the future gets a value it schedules its map code to run. This will run in some thread on thread pool, e.g. T2.
another thread T2 (which may or may not be the same as T1) executes the function passed to the map method.
At no point were either of the threads blocked waiting on the response from example.com.
Instead of waiting for IO by blocking a thread, Play secretly registered a callback so that when the IO is available it will fulfill a future. When the future is fulfilled it automatically calls the next part of the application code. This means the application code can wait (without using a thread) until the IO is available.
I have a working thread running all along the runtime, who generates events.
I can handle those events inside the UI thread by using disp = Windows::UI::Core::CoreWindow::GetForCurrentThread()->Dispatcher.
more precisely, I do the modifications to the UI by using disp->RunAsync(...) anywhere inside the working thread.
but I don't know how to do the inverted operation. I want to have some Async function inside the UI thread to perform operation (on some std::unique_ptr) in the working thread when I click on some button.
If I understand correctly you want to be able to run an async operation when a button is clicked, but on a specific thread to which you refer as your worker thread.
First - Since you want to use a resource in 2 threads you should not use unique_ptr and use shared_ptr since you share this resource between the two threads.
Second - if you don't necessarily have to run the action on a specific thread then you can simply use Windows::System::Threading::ThreadPool::RunAsync and capture the shared_ptr by value.
e.g:
namespace WST = Windows::System::Threading;
WST::ThreadPool::RunAsync(
ref new WST::WorkItemHandler(
[mySharedPtr](Windows::Foundation::IAsyncAction^ operation)
{
mySharedPtr->Foo();
}));
In case you have to run the operation on a specific thread then I assume you want to be able to append operations to an already running thread, otherwise you are creating a thread and you can use the above example.
So in order to append operations to an already running thread, that thread must have the functionality of getting a new operations and then running those operations in a synchronous order. This functionality is basically what the Dispatcher provides. This is what an Event Loop is, also called: message dispatcher, message loop, message pump, or run loop. Also you can find information by reading on the Recator\Proactor design pattern.
This CodeProject page shows one way of implementing the pattern, and you can use Winrt component to make it better \ more conveniant \ more familiar
Is there ever any reason to add blocks to a serial dispatch queue asynchronously as opposed to synchronously?
As I understand it a serial dispatch queue only starts executing the next task in the queue once the preceding task has completed executing. If this is the case, I can't see what you would you gain by submitting some blocks asynchronously - the act of submission may not block the thread (since it returns straight-away), but the task won't be executed until the last task finishes, so it seems to me that you don't really gain anything.
This question has been prompted by the following code - taken from a book chapter on design patterns. To prevent the underlying data array from being modified simultaneously by two separate threads, all modification tasks are added to a serial dispatch queue. But note that returnToPool adds tasks to this queue asynchronously, whereas getFromPool adds its tasks synchronously.
class Pool<T> {
private var data = [T]();
// Create a serial dispath queue
private let arrayQ = dispatch_queue_create("arrayQ", DISPATCH_QUEUE_SERIAL);
private let semaphore:dispatch_semaphore_t;
init(items:[T]) {
data.reserveCapacity(data.count);
for item in items {
data.append(item);
}
semaphore = dispatch_semaphore_create(items.count);
}
func getFromPool() -> T? {
var result:T?;
if (dispatch_semaphore_wait(semaphore, DISPATCH_TIME_FOREVER) == 0) {
dispatch_sync(arrayQ, {() in
result = self.data.removeAtIndex(0);
})
}
return result;
}
func returnToPool(item:T) {
dispatch_async(arrayQ, {() in
self.data.append(item);
dispatch_semaphore_signal(self.semaphore);
});
}
}
Because there's no need to make the caller of returnToPool() block. It could perhaps continue on doing other useful work.
The thread which called returnToPool() is presumably not just working with this pool. It presumably has other stuff it could be doing. That stuff could be done simultaneously with the work in the asynchronously-submitted task.
Typical modern computers have multiple CPU cores, so a design like this improves the chances that CPU cores are utilized efficiently and useful work is completed sooner. The question isn't whether tasks submitted to the serial queue operate simultaneously — they can't because of the nature of serial queues — it's whether other work can be done simultaneously.
Yes, there are reasons why you'd add tasks to serial queue asynchronously. It's actually extremely common.
The most common example would be when you're doing something in the background and want to update the UI. You'll often dispatch that UI update asynchronously back to the main queue (which is a serial queue). That way the background thread doesn't have to wait for the main thread to perform its UI update, but rather it can carry on processing in the background.
Another common example is as you've demonstrated, when using a GCD queue to synchronize interaction with some object. If you're dealing with immutable objects, you can dispatch these updates asynchronously to this synchronization queue (i.e. why have the current thread wait, but rather instead let it carry on). You'll do reads synchronously (because you're obviously going to wait until you get the synchronized value back), but writes can be done asynchronously.
(You actually see this latter example frequently implemented with the "reader-writer" pattern and a custom concurrent queue, where reads are performed synchronously on concurrent queue with dispatch_sync, but writes are performed asynchronously with barrier with dispatch_barrier_async. But the idea is equally applicable to serial queues, too.)
The choice of synchronous v asynchronous dispatch has nothing to do with whether the destination queue is serial or concurrent. It's simply a question of whether you have to block the current queue until that other one finishes its task or not.
Regarding your code sample code, that is correct. The getFromPool should dispatch synchronously (because you have to wait for the synchronization queue to actually return the value), but returnToPool can safely dispatch asynchronously. Obviously, I'm wary of seeing code waiting for semaphores if that might be called from the main thread (so make sure you don't call getFromPool from the main thread!), but with that one caveat, this code should achieve the desired purpose, offering reasonably efficient synchronization of this pool object, but with a getFromPool that will block if the pool is empty until something is added to the pool.
I am using NSURLSession dataTaskWithURL:completionHandler. It looks like completionHandler is executed in a thread which is different than the thread(in my case, it's the main thread) which calls dataTaskWithURL. So my question is, since it is asynchronized, is it possible that the main thread exit, but the completionHandler thread is still running since the response has not come back, which is the case I am trying to avoid. If this could happen, how should I solve the problem? BTW, I am building this as a framework, not an application.Thanks.
In the first part of your question you seem un-sure that the completion handler is running on a different thread. To confirm this let's look at the NSURLSession Class Reference. If we look at the "Creating a Session" section we can see in the description for the following method the answer.
+ sessionWithConfiguration:delegate:delegateQueue:
Swift
init(configuration configuration: NSURLSessionConfiguration,
delegate delegate: NSURLSessionDelegate?,
delegateQueue queue: NSOperationQueue?)
Objective-C
+ (NSURLSession *)sessionWithConfiguration:(NSURLSessionConfiguration *)configuration
delegate:(id<NSURLSessionDelegate>)delegate
delegateQueue:(NSOperationQueue *)queue
In the parameters table for the NSOperationQueue queue parameter is the following quote.
An operation queue for scheduling the delegate calls and completion handlers. The queue need not be a serial queue. If nil, the session creates a serial operation queue for performing all delegate method calls and completion handler calls.
So we can see the default behavior is to provide a queue whether from the developer or as the default class behavior. Again we can see this in the comments for the method + sessionWithConfiguration:
Discussion
Calling this method is equivalent to calling
sessionWithConfiguration:delegate:delegateQueue: with a nil delegate
and queue.
If you would like a more information you should read Apple's Concurrency Programming Guide. This is also useful in understanding Apple's approach to threading in general.
So the completion handler from - dataTaskWithURL:completionHandler: is running on a different queue, with queues normally providing their own thread(s). This leads the main component of your question. Can the main thread exit, while the completion handler is still running?
The concise answer is no, but why?
To answer this answer this we again turn to Apple's documentation, to a document that everyone should read early in their app developer career!
The App Programming Guide
The Main Run Loop
An app’s main run loop processes all user-related events. The
UIApplication object sets up the main run loop at launch time and uses
it to process events and handle updates to view-based interfaces. As
the name suggests, the main run loop executes on the app’s main
thread. This behavior ensures that user-related events are processed
serially in the order in which they were received.
All of the user interact happens on the main thread - no main thread, no main run loop, no app! So the possible condition you question mentions should never exist!
Apple seems more concerned with you doing background work on the main thread. Checkout the section "Move Work off the Main Thread"...
Be sure to limit the type of work you do on the main thread of your
app. The main thread is where your app handles touch events and other
user input. To ensure that your app is always responsive to the user,
you should never use the main thread to perform long-running or
potentially unbounded tasks, such as tasks that access the network.
Instead, you should always move those tasks onto background threads.
The preferred way to do so is to use Grand Central Dispatch (GCD) or
NSOperation objects to perform tasks asynchronously.
I know this answer is long winded, but I felt the need to offer insight and detail in answering your question - "the why" is just as important and it was good review :)
NSURLSessionTasks always run in background by default that's why we have completion handler which can be used when we get response from Web service.
If you don't get any response explore your request URL and whether HTTPHeaderFields are set properly.
Paste your code so that we can help it
I just asked the same question. Then figured out the answer. The thread of the completion handler is setup in the init of the NSURLSession.
From the documentation:
init(configuration configuration: NSURLSessionConfiguration,
delegate delegate: NSURLSessionDelegate?,
delegateQueue queue: NSOperationQueue?)`
queue - A queue for scheduling the delegate calls and completion handlers. If nil, the session creates a serial operation queue for performing all delegate method calls and completion handler calls.*
My code that sets up for completion on main thread:
var session = NSURLSession(configuration: configuration, delegate:nil, delegateQueue:NSOperationQueue.mainQueue())
(Shown in Swift, Objective-C the same) Maybe post more code if this does not solve.
It's known that, in cases when one needs comunicate between UI thread and working thread, an hidden window must be created because of thread safety(handle reconstruction).
For exemplify:
Form1 has N dynamicaly created TProgressBar instances with the same name of a background running .
Is always garanteed that WM_REFRESH will only be called inside Task Thread.
Form1 has H : THandle property that allocates the following procedure:
procedure RefreshStat(var Message: TMessage); message WM_REFRESH;
Inside RefreshStat, in cases when there is only 1 background thread I could easily use L and W parameter to map Task Id and position.
I don't know if the title says what I want to know, but let's imagine if we have an application that has multiple background tasks running.
In my case I use TProgressBar to report progress the done.
Does AllocateHwnd garantee that all messages arrives with no race condition the hidden window?
What happens if two or more tasks post the message at the same time?
If this needs to be controled manually, I wonder if there is something else to do besides creating another message loop system in the custom message.
I hope the question is clear enough.
The message queue associated with a thread is a threadsafe queue. Both synchronous and asynchronous messages from multiple other thread are delivered safely no harmful date races. There is no need for any external synchronization when calling the Windows message API functions like SendMessage and PostMessage.
If two threads post or send messages to the same window at the same time, then there is no guarantee as to which message will be processed first. This is what is known as a benign race condition. If you want one message to be processed before the other then you must impose an ordering.