addAsync chaining help

There is a problem in the current code where the firing of an async will be ignored if it fires before the test or currently executing async has finished executing. This is because AsyncTestHelper.runNext clears the values of func, objToPass, and extraData when it finishes. This has been fixed in a test version of FlexUnit.

I filed bug https://bugs.adobe.com/jira/browse/FXU-19 for this issue.

Ok, if I do that as recommended in your last post, I get a false positive during a failure in the read portion. The error pops up in the app as an exception but the test passes green.

If I have the second addAsync with the callLater call, then it fails red with no pop-up exception

> interesting. Well I added a wrapper call for <br />> Application.application.callLater for the call to readWrittenFile and <br />> now it works.<br /><br /> If you remove the call to callLater you can also remove the second <br />addAsync() call. What that means is that readWrittenFile() isn't <br />performing any asynchronous action which requires the use of addAsync(). <br />Only events that fire asynchronously need to be wrapped with addAsync(). <br />Events that fire immediately can be setup with a normal addEventListener() <br />call.<br /><br />-- Daniel R. <danielr@neophi.com> [http://danielr.neophi.com/]

interesting. Well I added a wrapper call for Application.application.callLater for the call to readWrittenFile and now it works.

Thanks for your help!

Side note: This behavior may need to be reworked as it is not very natural nor obvious to someone trying to create some tests.

> I really am confused here, if there are multiple items on the async<br />> helper stack, why is the timer being stopped on the first call which<br />> subsequently is preventing my second "addAsync" handling function not to<br />> be called?<br /><br /> The timer is restarted when the method called by the first <br />addAsync() finishes executing. The test harness then determines that <br />another addAsync() was registered and it should wait for that instead of <br />running the next test.<br /><br /> If I understand the test flow from your original message <br />"verifyWriteComplete" must exit before the second event fires as noted <br />below.<br /><br />testWrite() {<br /> addAsync("verifyWriteComplete");<br /> // this method must return before the event<br /> // that triggers verifyWriteComplete is called<br />}<br /><br />verifyWriteComplete() {<br /> readWrittenFile();<br /> // this method must return before the event<br /> // that triggers verifyReadComplete is called<br />}<br /><br />readWrittenFile() {<br /> addAsync("verifyReadComplete");<br />}<br /><br />verifyReadComplete() {<br />}<br /><br />-- Daniel R. <danielr@neophi.com> [http://danielr.neophi.com/]

Basically I am doing essentially the same thing as listed here (last example) to no avail.

http://www.adobe.com/cfusion/communityengine/index.cfm?event=showdetails&productId=2&postId=6882

Thanks saw your article earlier. However I am already making multiple calls to addAsync. I have been stepping into the flexunit code:

This function is returned on the first call to addAsync(). The first event that this handles stops the timer and "wasReallyAsync" is true so the code continues. At this point my code calls "addAsync" again. When the event fires for the 2nd event, this code runs again but since the timer is stopped my code is not being called.

I really am confused here, if there are multiple items on the async helper stack, why is the timer being stopped on the first call which subsequently is preventing my second "addAsync" handling function not to be called?

public function handleEvent( event : Object ) : void
{
var wasReallyAsync : Boolean = timer.running;

timer.stop();

if ( shouldFail )
return;

objToPass = event;
if ( wasReallyAsync )
{
testResult.continueRun( testCase );
}
}

> The problem is that the second chained addAsync (C) above in flexunit is <br />> still reporting that it was never called within the timeout, despite the <br />> fact that it obviously is. The test runner is reporting a failure, even <br />> though the verifyReadComplete wrapped in an addAsync was called OK.<br /><br /> This sounds like an issue I ran across where if the function <br />mentioned in addAsync() isn't actually being called in an asynchronous <br />manner, you will get the odd behavior. I wrote up some details on this <br />case (in particular the last example):<br /><br />http://life.neophi.com/danielr/2008/08/odd_addasync_behavior_when_han.html<br /><br />-- Daniel R. <danielr@neophi.com> [http://danielr.neophi.com/]