I have an application that uses Remote SharedObjects and I am seeing some strange behaviour. I am writing an ActionScript application in AS3 using Flash Builder and connecting to Wowza Media Server 2.
My application is working just fine but I am now trying to write unit tests for it using FlexUnit. My unit tests involve creating multiple connections to the same remote SharedObject and making sure that I am getting updates correctly. Everything seems to be working well except that I am not getting any of the SyncEvent.SYNC events with an info.code of "delete". When I run my applications independently in separate tabs or even separate swfs embedded in the same page it works fine. For some reason though it does not work when inside a unit test. I have also found that if I load the swfs using a Loader inside the same SWF then I get the same behaviour. It seems to me to be something strange about the way multiples of the same SharedObject behave within the same SWF. I have had to work around other strange behaviour in the unit tests such as oldValues not being set properly in the onSync events too.
Anyone have any ideas how I can work around this? Is this a known issue? Am I crazy? :)
Would appreciate any help!
I also faced this problem before when I was working on some Flex application using some coding frameworks like Cairngorm, and connecting to AMFPHP with multi remote objects.
At that time what I come up with to resolve the issue is to make sure that those remote objects won't be fired at the same time. That is trying to make some so called "sequential chain" to fire those remote objects one after another.
In order to achieve this it may be difficult if you do it from scratch, you may consider making use of those modern ActionScript framework 2.0 to help you (e.g. SWIZ or Robotlegs). This may be too complex to handle in the very beginning. I suggest you can just place the remote object parts to be handled by the framework while keeping other things intact.
To get your sharedObject instance, you are using the static method SharedObject.getRemote(). I believe this method will always return the same instance for a given name (and if the persistence parameter has the same value).
You can have the same kind of issue when you remove a SharedObject from your app (mySO = null) and you reinstantiate it before the garbage collector did its job).
This kind of behavior makes sense to me, but it can sometimes be a problem I must admit. Anyway it should be easy to test in a debug session (Have a look at your objects instance number).
Now talking about unit tests, what are you testing? The SharedObjects behavior? If so, I believe there is some misconception here. It you really want to test this kind of behavior (and I would be interested in the reason behind), then I guess you will need some more complex tests that run two separate applications.
Hope it helps!
We had similar behavior with deletes in our project.
When we call so.close(). Then delete some key in shared object. Then connect SO again - it still see deleted key alive.
Workarounds: do not close SO or update deleted keys with some constants values (-2 for example) to mark them deleted.
Wowza 3 was used.
Related
I'm currently restructuring my LibGDX-Projekt to fit an ECS-Structure using Ashley.
At the moment I have the following System-Structure:
InputSystem (handles Player-Input)
PhysicSystem (applies velocity to entities, does collisiondetection and moves them)
CameraSystem (adjusts the camera to the camera-focus)
RenderingSystem (transforms the SpriteBatch with the camera-information and draws all entities)
Now I know that each System is supposed to contains all the logic it needs and is encapsuled to reduce complixity.
But I need the camera in the CameraSystem to adjust it, I need it in the RenderSystem to apply the camera-transformation and I need it in the InputSystem to see where the mouse is pointing at. How do you solve this with the ECS-Approach? Can Systems communicate with one another? Should I just use a Singleton called "SharedData" where I dump all the stuff that multiple systems need? Seems a little ugly to me.
Thanks in advance :)
Yes, using a singleton would defeat the purpose of using an ECS in the first place. In an ECS, the approach used is that all shared data is expressed using entities. It is perfectly fine to have an Entity type for which there will only be once instance of it during the lifetime of the game.
These could then be accessed directly from the Engine when the EntiySystem is added to it. An alternative approach could be to pass commonly accessed entities to the constructor of the systems if it is frequently accessed, in the past I have used this for passing a Box2D World entity.
I believe my code does a better job of explaining my reasoning so you can have a look at one of my previous games using Ashley. https://github.com/basimkhajwal/LSD
P.S - My project also includes an event queue for messaging between different entity systems (without passing data, all the data is still encapsulated in Entity classes) which I found to be extremely useful.
I am working on a custom SSIS component that has 4 asynchronous outputs. It works just fine but now I have a user request for an enhancement and I am not sure how to handle it. They want to use the component in another context where only 2 of the 4 outputs will be well defined. I foolishly said that this would be trivial for me to support, I planned to just look to see if the two "undefined" streams were even connected, if not then I would skip over that part of the processing.
My problem is that I cannot figure out if an output is connected at run time, I had hoped that the output pipeline or output buffer would be missing. It doesn't look like that is the case; even when they are not hooked up the output and buffer are present.
Does anyone know where I should be looking to see if an output has a downstream consumer or not?
Thanks!
Edit: I was never able to figure out how to do this reliably, so I ended up making this behaviour configurable by the user. It is not automatic like I would have hoped but the difference I found between the BIDS environment and the DTExec environment pushed me to the conclusion that a component probably should not be making assumptions about the component graph it is embedded in.
I've been experiencing a strange error while working on my Play Framework project. While my project is running, I will sometimes receive a ClassCastException, but the error is this:
ClassCastException occured : models.Person cannot be cast to models.Person
This occurs usually when I'm calling a find method such as:
Person p = Person.find("name=?","Joe").first();
If I restart the project, the problem goes away, but only temporarily. It makes testing my project a major pain. How do I fix this?
I've experienced this error while in Dev mode in Play, in two scenarios (as far as I can remember):
Modify an entity and try to recover values from cache that are objects of that entity class.
A compilation error while reloading the code of the page/application
In both scenarios fixing compilation errors or cleaning the cache solved the issue.
Not saying that those are the only possibilities, it may be that you are having some other issue.
This most likely occurs because you've somehow loaded the Person class under two different class loaders. When a class is loaded twice in two different class loaders it's effectively two distinct classes.
(Unfortunately, I can't tell you where/how you might have done this.)
(And it is a bit curious to have the problem pop up on the statement you list. Are you certain that's where it's occurring? Perhaps you should show the exception traceback.)
In my case, this is related to applying evolutions from the web interface. Someone raised a bug for this, but so far it hasn't received any attention from the dev team. There is a patch attached to the ticket, but I haven't tried it, so YMMV.
I have created myself a test project with a single test in it. Within the "TestInitialize" method I'm trying to call my StructureMap setup method where I set which concrete instances should be used for which interfaces.
However as soon as the test tries to run I get a {"Operation could destabilize the runtime."} exception. Without using structuremap it works fine (but obiviously I need it in there).
Can anyone point me in the right direction as to what might be causing this exception?
After a lot of messing around, I discovered that all I needed to do was to right-click on the StructureMap dll and click "Unblock" and the permissions were sorted.
I'm trying to implement the advice found in this great blog post for batch processing in grails with MySQL. The problem that I'm having is that inclusion of periodic calls to session.clear() in my loop causes org.hibernate.LazyInitializationException's to be thrown. There's a quote down in the comments section of the page:
You’re second point about potentially
causing LIEs is absolutely true. If
you’re doing other things outside of
importing with the current thread,
you’ll want to make sure to reattach
any objects to the session after
you’re doing your clearing.
But how do I do that? Can anyone help me specifically understand how to "reattach any objects to the session after I'm done clearing?
I'm also interested in parallelizing the database insertion process so that I can take advantage of having a multi core processor. Can anyone provide advice on how to do that in Grails?
Grails has a few methods to help with this (they leverage hibernate under the covers).
If you know an object is detached, you can use the attach method to reconnect it.
If you've made changes to the object while it was detatched, you can use merge.
If for whatever reason, you're not sure if an object is attached to the session, you can use the link text method to find out if it is or isn't.
It might also be worth reviewing the Hibernate documentation on Session.