Do I need to close IAsyncOperation? - asynchronous

In WinRT, do I need to call Close() on IAsyncOperation<T> ^ objects? Either the the operation initiating code, or in the completion handler? If not, why is it there?

From your syntax, you're using C++/CX.
For C++/CX, IAsyncOperation::Close is automatically called in the destructor, so you do not need to explicitly call it.
For JS, you must explicitly call .close().
For C#, it's projected as IDisposable, so a using() statement is appropriate.

Related

How to implement early stopping as an extension

As this thread, we can stop iteration loop by setting function (f:trainer -> bool) as Trainer's stop_triger.
But in this way, I think we can't use other extension such as LogReport which use stop_trigger=((args.epoch, '10')).
So, my question is how to implement early stopping as the Extension and how to send a signal to stop trainer's iteration from Extension.
thanks.
I implemented the example code on gist,
and updated the answer on the original thread.
I noticed that stop_trigger originally uses tuple notation like (args.epoch, '10'), instead we need to change to pass a callable object (EarlyStoppingTrigger in above example).

Is there an operator which is the opposite of mergeMap()?

We have a case where we only want one HTTP request to go out at a time and only want to allow retry or call again if the pending call either succeeds or fails. We've been doing the classic wrap it in a boolean but were wondering if there was an operator for it. My hunch is no.
I think you would need switchMap. It will only subscribe to the last value of the external observable and it will map it to an internal observable. With switchMap, its possible to cancel previous http requests on the fly.
For more info take a look in the following link

Do we have XQuery abort() function

Do we have an abort() function to terminate a XQuery script from XQuery 3.1? I can write a if-else to return nothing. But I dont want to do that.
You can use the error(...) function for that. The thrown error can then be caught by try/catch if needed. Using errors for control flow is not advisable however, because it is normally not optimized for performance.

Signalr - Serialize callback as event not a function call?

In Signalr, is there any support for having events instead of callbacks.
Let me explain before you grab your pitchforks.
In following with the first example here
Clients.All.addContosoChatMessageToPage(name, message);
Wouldn't call a hub proxy's addContosoChatMessageToPage(name, message), but would dispatch a addContosoChatMessageToPage event with some extra information. (not asking that it be the same api call exactly)
The reason I'm asking all of this is because
This works much better alongside functional reactive programming frameworks like ELM and bacon.js
I don't want to do this myself and essentially create my own sub-framework. Of course I could always do Clients.All.CreateEvent(name,params...) where I'm continually calling back my method to do this event creation
I actually think events work better in some scenarios for separation of concerns.
Am I crazy? does something like this exist?
This is already supported. If you don't want to do the dispatching yourself and you know the name of the "event" or "method" at runtime you can do this:
IClientProxy proxy = Clients.All;
proxy.Invoke(name, args);
This lets you write code where you may not know the name of the event you're trying to callback on the client at compile time.

custom validator against remote object

I have a need to validate a field against our database to verify unique-ness. The problem I seem to be having is that the validators doValidation() exits before we've heard back from database.
How can I have the validator wait to return its payload until after we've heard from the DB?
Or perhaps a better question might be (since I think the first question is impossible), how can I set this up differently, so that I don't need to wait, or so that the wait doesn't cause the validation to automaticallly return valid?
If you're using a remote object, you can specify the method call inside your remote declaration and assign a function to the result call. The result call only runs once the remote server returns something, so it won't be run before your validation.
Do your validation call in said result function call (which you will have to create) and you should be good. Your code should go something like this:
<s:RemoteObject id="employeeService"
destination="ColdFusion"
source="f4iaw100.remoteData.employeeData"
endpoint="http://adobetes.com/flex2gateway/"
result="employeeService_resultHandler(event)"/>
**<s:method name="dataCheckCall" result="dataCheckResult(event)"/>**
<s:RemoteObject />
And in your script:
function protected dataCheckResult(event:ResultEvent):void {
**doValidate();**
}
Edit: As soon as you call "dataCheckCall" the method will start running. If, for whatever reason, you want to call this WITHIN your validator, you can do so, and then dataCheckResult will run whenever it returns with it's payload (pretend doValidate is called elsewhere). I've left a message below as well.
You are trying to fit an asynchronous process (fetching data from a DB) into a synchronous process (checking all the validators in turn).
This won't work...
You'll need to either roll your own validator framework, or use a different method of determining the legality of your controls.
P.S. The MX validators are rubbish anyway!
What I've managed to do, seems to work, mostly. I don't like it, but it at least performs the validation against the remote source.
What I've done, then, is to use an 'keyUp' event handler to spin off the database lookup portion. In the meanwhile, I set up a string variable to act as some kind of a Flag, which'll be marked as 'processing'. When the response event fires, I'll examine its contents, and either clear the flag, or set it to some kind of other error.
Then, I have created a new 'EmptyStringValidator' will check the contents of this flag, and do its job accordingly.
Its indirect, but, so far, seems to work.

Resources