Despite my best efforts, I am unable to produce the kind of synchronization effects I would like to in ActionScript. The issue is that I need to make several external calls to get various pieces of outside information in response to a user request, and the way items will be laid out on the page is dependent on what each of these external calls returns. So, I don't care that all of these calls return asynchronously. However, is there any way to force some amount of synchronization on ActionScript, so that at least calling the method for doing the final layout and placement of items on the page is dependent on all of my calls finishing?
If I understand the question right, event listeners are probably your best bet. Most loader classes throw an Event.COMPLETE message when they finish doing everything behind the scenes. If you wrote those external calls, it would be easy to dispatch a complete event at the end.
So when you make all these external calls, have a function that listens to when those calls complete. This function would keep track of how many calls have been made, and when there's none left to run, continue building your layout.
Rough Sketch to explain below:
var numProcesses:int = 0;
slowthing.addEventListener(Event.COMPLETE,waitForSlowest);
numProcesses++;
slowthing.load();
quickThing.addEventListener(Event.COMPLETE,waitForSlowest);
numProcesses++;
quickthing.load();
function waitForSlowest(e:Event)
{
numProcesses--;
if(numProcesses == 0)
finalizeLayout();
}
Related
Lets say you have a complex Lua application, and there is some base function that different parts of your code call repeatedly. It's a stateless function with little to no side effects, and fairly simple.
How does the virtual machine handle this? Does it queue up all the calls, and let them run one by one, waiting for the function to to return before calling it again? Or does it do some trickery to avoid this sort of situation? What if the function had some big side effects, like print()?
Lua is single threaded so every function call must return before the next one is called. If a function is blocked then so is the VM. The only way around that is coroutines or Lua lanes or C threads.
Examples:
Asynchronous method with its own dispatching:
// Library
func asyncAPI(callback: Result -> Void) {
dispatch_async(self.queue) {
...
callback(result)
}
}
// Caller
asyncAPI() { result in
...
}
Synchronous method with exposed dispatch queue:
// Library
func syncAPI() -> Result {
assert(isRunningOnCorrectQueue())
...
return result
}
// Caller
dispatch_async(api.queue) {
let result = api.syncAPI()
...
}
These two examples behave the same but I am looking to learn whether one of these ends up complicating a larget codebase more than the other, especially when there is a lot of asynchrony.
I would argue against both of the patterns you propose.
For the first pattern (where the API manages it's own backgrounding) I see little or no benefit to doing it this way, as opposed to leaving it to the caller. If you want to use a private, serial queue to protect data (or any other sort of critical section) internal to your API, that's fine, but that queue should be private, and it should specifically not target any public, non-global-concurrent queue (Note: it should especially not target the main queue). Ideally, the primary implementation of your API would also take a second parameter, so callers can specify on which queue to invoke the callback. (People can work around the lack of such a parameter by passing a callback block that re-dispatches to their desired queue, but I think that's clunkier than having an extra, optional parameter.) This puts the API consumer in complete control of the concurrency, while preserving your freedom to use queues internally to protect state.
As to the second approach, it's my opinion that we all should avoid creating new synchronous, blocking API. When you provide a synchronous, blocking API and don't provide a callback-based version, that means that you have denied consumers of your API any opportunity to avoid blocking. When you only provide synchronous, blocking API, then if someone wants to call your API in the background, at least one thread (in addition to any additional threads that your API consumes behind the scenes) will be consumed from the finite number of threads available to each process. (In the worst case this can lead to starvation conditions that are effectively deadlocks.)
Another red flag with this second example is that it vends a queue; Any time an API vends a queue, something is amiss. As mentioned, if you want to use a private serial queue to protect state or other critical sections internal to your API, go for it, but don't expose that queue to the outside world. If nothing else, it unnecessarily exposes details of your implementation. In looking at the system framework headers, I couldn't find a single case where a dispatch_queue_t was vended where it wasn't immediately obvious that the intent was for the API consumer to push in the queue, and not read it out.
It's also worth mentioning that these patterns are problematic regardless of whether your workload is CPU-bound or IO-bound. If it's CPU-bound, then not managing your own dispatch gives consumers of the API explicit control over how this CPU work is executed. If your workload is IO-bound, then you should use the OS- and libdispatch-provided asynchronous IO mechanisms (dispatch_io, dispatch_sources, kevent, etc) to avoid consuming a thread (or more than one) for the duration of your work.
Another answer here implied that forcing consumers to manage their own concurrency leads to "boilerplate" code. If you feel that the burden of API consumers potentially having to wrap calls to your API with dispatch_async is too great, then feel free to provide a convenience overload that dispatches to the default global concurrent queue, but please always leave the version that allows API consumers the ability to explicitly manage their own concurrency.
If, on the other hand, all this is internal to the implementation, and not part of the public API, then do whatever is most expedient, knowing that you can refactor the implementation behind the public API any time in the future.
As you said, the 2 generally accomplish the same thing but the first is more preferable in most scenarios. There are several benefits to using the first method.
The API is simpler. You simply call the method and provide code for the callback block.
Less boilerplate code, No typing dispatch_async every time you want to call it as it is just included in the method itself.
Less room for bugs/errors. By wrapping the asynchronous logic inside the method itself, you ensure that it is called on the right queue internally without the caller having to worry about any of that.
Touching on the last point, you also have finer control over the queue itself. Let's say you are trying to perform certain tasks on a particular queue. It is way simpler to simply wrap the code in a GCD call on that queue a single time rather than having to remember to reuse that same queue every time you want to call the method.
If I write the following Dart code, how do I know which click handler happens first?
main() {
var button = new ButtonElement();
var stream = button.onClick.asBroadcastStream();
stream.listen(clickHandler1);
stream.listen(clickHandler2);
}
Let's say I'm in other code that doesn't know anything about the first two click handlers, but I register another one.
Can I know that the stream has two listeners?
Can I pause or cancel all other subscribers?
If I write button.onClick.asBroadcastStream() again elsewhere, does it point to the same stream as was used in main?
Can I say in one of the handlers to not pass event on to the other broadcast listener? Is that a consumer?
Let's say I'm in other code that doesn't know anything about the first
two click handlers, but I register another one.
Can I know that the stream has two listeners?
No, you can't. You could extend the stream class or wrap it and provide this functionality yourself, but it does not feel like a good design choice, because I don't think a listener should know about other listeners. What are you trying to do exactly? Perhaps there's a better way than letting listeners know about each other.
Can I pause or cancel all other subscribers?
You can cancel/pause/resume only the subscriber you are dealing with. Again, you probably shouldn't touch other listeners, but I guess you could wrap/extend the Stream class to have this behavior.
If I write button.onClick.asBroadcastStream() again elsewhere, does it point to the same stream as was used in main?
No, at least not at the current version of SDK. So, unfortunately, you need to store a reference to this broadcast stream somewhere, and refer to it, because calling asBroadcastStream() multiple times will not yield in the result you might expect. (Note: at least based on quick testing: http://d.pr/i/Ip0K although the documentation seems to indicate different, I have yet to test a bit more when I find the time).
Can I say in one of the handlers to not pass event on to the other broadcast listener?
Well, there's stopPropagation() in the HTML land which means that the event won't propagate to other elements, but it's probably not what you were looking for.
For being able to stop an event firing in other listeners, there needs to be an order of which the listeners are getting called. I believe the order is the order of registration of those listeners. From the design perspective, I don't think it would be a good idea to allow a listener to cancel/pause others.
Event propagation in HTML makes sense since it's about hierarchy, but here we don't have that (and even in case of events in HTML there can be multiple listeners for the single element).
There's no way to assign weight to listeners or define the order of importance, therefore it's not surprising that there isn't a way to stop the event.
Instead of letting listeners know about each other and manipulate each other, maybe you should try to think of another way to approach your problem (whatever that is).
Is that a consumer?
The StreamConsumer is just a class that you can implement if you want to allow other streams to be piped into your class.
Can I know that the stream has two listeners?
No, you have a ´Stream´ that wraps the DOM event handling. There is no such functionality.
Can I pause or cancel all other subscribers?
Look at Event.stopPropagation() and Event.stopImmediatePropagation(), and possibly Event.preventDefault().
If I write button.onClick.asBroadcastStream() again elsewhere, does it point to the same stream as was used in main?
[Updated] No, the current implementation doesn't gives you the same Stream back since the onClick getter returns a new stream every time it is invoked. However, the returned stream is already a broadcast stream so you shouldn't invoke asBroadcastStream() on it. If you do you will hower just get a reference to the same object back.
Stream<T> asBroadcastStream() => this;
Can I say in one of the handlers to not pass event on to the other broadcast listener? Is that a consumer?
Again, take a look at Event.stopPropagation() and Event.stopImmediatePropagation(), and possibly Event.preventDefault().
Using Silverlight 3 with RIA: What's the difference between the LoadOperation.Completed event and using a callback through the DomainContext.Load method? Both fire asynchronously and both provide access to the LoadOperation. When/why would I use one over the other?
Thanks :-)
There's no difference; the 2 options are offered for flexibility. Many times, the callback will suffice, but if you return the LoadOperation from a method, the caller could then choose to subscribe.
Note that even if the Load completes before you subscribe to the Completed event, your handler will still get called. We guarantee every subscriber to the event will be called.
Agreed that there is not difference in functionality. It's about coding style. If the work I have to do following completion of the query is simple, like binding data to a grid, I like the use the following syntax to inline the completed code, rather than defining a separate method.
context.Load<EntityType>(query).Completed += (lo, args) =>
{
myGrid.ItemsSource = ((LoadOperation) lo).Entities;
}
This has the cleanliness of synchronous code, but the code inside the braces will in fact be executed asynchronously.
Good luck!
I am creating an action script library.I am calling some APIs which parses some xml and gets me the result. It dispatches an Event.COMPLETE when the parsing is done. I want to monitor whether this event is dispatched in some while loop like "while(eventnotdispatched)"
is it possible? I know the other way would be to addeventlistener. But please let me know if the other thing is possible.
Thanks
NO, it is not possible. Actionscript is single threaded. Thus while you are waiting in your while loop, that is the only thread running, and the process you are waiting for can never complete. This is why everything is done with events, so that's what you should use. If you need to update your display periodically while you are waiting for something to complete...again, use events. Create a Timer object which generates a TIMER event every so often, and use that to make your updates.
EDIT: Davr is right, you would not be able to use the while loop like this. You would need a timer.
Yes, it is possible to poll for it. BUT you will still need to create an event listener. It will work something like this:
private var loadCompleted = false;
private var timer:Timer= new Timer(1);
private function onInitCompleted(event:Event):void
{
timer.addEventListener(TimerEvent.TIMER, timerHandler);
timer.start();
}
private function loadCompleteEventHandler(event:Event):void
{
loadCompleted = true;
...
}
private function timerHandler()
{
if(!loadCompleted)
{
... // stop the timer or something.
timer.stop();
}
}
Please note, this is VERY BAD code. I would NEVER use it in production because Actionscript is a event driven language. There should be absolutely NO REASON for you to need to do this. Whatever you are trying to do could be accomplished using another method much simpler. Tell me what you are trying to accomplish with this and I will present a better solution.
Sorry for yelling, it's late and I am sleepy.
Doing that means forcing a synchronous model of execution on the underlying asynchronous model (that works with callbacks).
What are you trying to achieve exactly, and why not use a callback?
I agree with the statements about it probably being a bad idea and a while loop will certainly not work this way in ActionScript. However, there may be legitimate reasons for doing what you are attempting to do. Only you can prevent bad code. Instead of judging, I'll just get to an answer for your question.
First I'm going to make an assumption, that what you really want to do is monitor a property and for some reason the API for this object does not dispatch an event when this property changes. I'm making this assumption because if you have the event available, I assume you would just use the event.
So... you have an object weirdXmlObj with a property loaded that defaults to false but goes to true when the XML is loaded.
In this case with slight modifications the code posted by CookieOfFortune would in fact work. You wouldn't need the loadCompleteEventHandler function (which was never attached anyway) and in the timer handler you would simply check if( weirdXmlObj.loaded ) and then branch however you wanted to.
Ah but there may be a simpler way, depending on what you are doing.
If you have a display object handy. (i.e. something that makes sense, not just some random object.) You can attach your code to the stage's EnterFrame event instead of using a timer.
myDisplayObject.stage.addEventListner(Event.ENTER_FRAME,frameEnterHandler);
A couple of things to be aware of:
You don't really even need to go to the stage level, all display objects support the EnterFrame event, but it's a nice place to attach the event listener.
You really should keep whatever the function calls to a minimum. In particular the actual frameEnterHandler function should do nothing more than do the if( weirdXmlObj.loaded ) check.
You are attempting to circumvent event-driven programming, which is not a good idea. This is often the case when someone approaches from an older model and does not yet have a good frame of reference to appreciate the elegance of event-driven programming.
Events are your friends. They work very well. Your loadCompleteHandler is all that is required. Want to do something else in response? Add the call in that handler:
private function loadCompletedHandler(event:Event):void
{
waitingObject.fileWasLoadedSoGoDoThatThing();
}
There is no need to make it any more complicated than that. No need for a semaphore or a loop to check the semaphore. Unnecessary environmental semaphores can break the encapsulation that could shield you from unwanted side effects.