We have a flex application that connects to a proxy server which handles authentication. If the authentication has timeout out the proxy server returns a json formatted error string. What I would like to do is inspect every URLRequest response and check if there's an error message and display it in the flex client then redirect back to login screen.
So I'm wondering if its possible to create an event listener to all URLRequests in a global fashion. Without having to search through the project and add some method to each URLRequest. Any ideas if this is possible?
Unless you're only using one service, there is no way to set a global URLRequest handler. If I were you, I'd think more about architecting your application properly by using a delegate and always checking the result through a particular service which is used throughout the app.
J_A_X has some good suggestions, but I'd take it a bit farther. Let me make some assumptions based on the limited information you've provided.
The services are scattered all over your application means that they're actually embedded in multiple Views.
If your services can all be handled by the same handler, you notionally have one service, copied many times.
Despite what you see in the Adobe examples showing their new Service generation code, it's incredibly bad practice to call services directly from Views, in part because of the very problem you are seeing--you can wind up with lots of copies of the same service code littered all over your application.
Depending on how tightly interwoven your application is (believe me, I've inherited some pretty nasty stuff, so I know this might be easier said than done), you may find that the easiest thing is to remove all of those various services and replace them by having all your Views dispatch a bubbling event that gets caught at the top level. At the top level, you respond to that event by calling one instance of your service, which is again handled in one place.
You may or may not choose to wrap that single service in a delegate, but once you have your application archtected in a way where the service is decoupled from your Views, you can make that choice at any time.
Would you be able to extend the class and add an event listener in the object's constructor? I don't like this approach but it could work.
You would just have to search/replace the whole project.
Related
I am using Atmosphere Framework 2.0.8.
I have implemented an AtmosphereHandler in my application and have two way communication occurring correctly over WebSockets and Long Polling.
I am trying to add some handling code for when the client disconnects to clean up some resources specific to that client (ie. I have an entry in a table I want to remove).
I have read the following wiki entries:
OnDisconnect Tricks: https://github.com/Atmosphere/atmosphere/wiki/onDisconnect-tricks
Configuring Atmosphere Listener: https://github.com/Atmosphere/atmosphere/wiki/Configuring-Atmosphere-Listener
The thing I am not clear on is where I should add the call to
atmosphereResource.addEventListener( new AtmosphereResourceEventListenerAdapter() {} );
I eventually found some example code in the JavaDoc for the AtmosphereHandler that registers the EventListener in the onRequest() method. http://atmosphere.github.io/atmosphere/apidocs/org/atmosphere/cpr/AtmosphereHandler.html
What I would like to know is if this is the correct way to go about it?
It is my understanding that the AtmosphereResource represents the connection between a client and the server for the life of that connection. The uuid stays consistent for the object on multiple calls through the onRequest() method from the same client. As such, the same AtmosphereResource object will get the EventListener added every time the onRequest method is called.
This seems wrong. Wouldn't this lead to thousands of EventListeners being registered for each AtmosphereResource?
It seems that the EventLister should only be registered once for each AtmosphereResource.
I feel like I am missing something fundamental here. Could someone please explain?
Here's an example using MeteorServlet, so it won't look exactly like what you'll have to do, but it should get you started. I add the listener to a Meteor instance, and you'll add yours to an AtmosphereResource. Each resource gets just one listener.
The overridden onDisconnect() method calls this Grails service method that handles the event. Of course, you'll want to call something that cleans up your database resource.
Note that the servlet is configured using these options. I think you might need the org.atmosphere.interceptor.HeartbeatInterceptor, but it's been so long since I initially set it up, I can't remember if it's necessary.
I am trying to understand the various ways of storing and instantiating Application (i.e. objects available to every user) and Session level (objects created and available to users only for their session) variables. Also, how does OWIN fit into all of this?
Global.asax.cs - This can contain a bunch of different methods. I believe that Application_Start is only called during the first request. However, there are a few candidates here for methods to populate session level variables (e.g. Session_Start and Application_BeginRequest). What is the standard way of doing this?
There is also the Startup class used by OWIN. I get that OWIN lets you store Application level variables, but why wouldn't you just use the HttpApplicationState Application variable accessible from Global.asax.cs to accomplish this? Also - can OWIN handle Session variables?
"I believe that Application_Start is only called during the first
request."
Only for the first request after calling the web application. For instance, this is the case after deploying, ApplicationPool Recycling, restarting or coming out of sleep.
Let's assume 3 users visit your web application. Application_Start will only be called for one of them, specifically the first one that visits it. Therefore it is not suited for populating user-specific session values.
However, there are a few candidates here for methods to populate session level variables (e.g. Session_Start and Application_BeginRequest). What is the standard way of doing this?
In the past I've worked with Session_Start to initialize user-specific session values (like default values) on numerous projects and never had an issue with it.
I'm really not sure what the question is, as I said in the comments. I'm going to ignore the OWIN stuff since I don't know, frankly.
Firstly, try not to store state at all. Design to pass state back and forth between server and client in models, or the URL, even in the HTML on the client such as in the URLs in the <a> tags your rendering, or (rarely) in cookies, rather than keep things in memory. Stateless designs are way more scalable.
Storing state isn't "usually" done in the Global.asax but then what's usual? I store state as and when I need it, load it or otherwise come by that data. For me in MVC, that's usually downstream of a Controller action, maybe while logging someone in, or reading some data received in a model, like a customer clicking 'add to cart'.
Application state I rarely use, though I store long-lived and shared data within normal fields and properties in long-lived static classes. These die when the app is recycled, but I don't usually care since the apps are designed to work without it, stateless; its usually cached bits of data.
Also, Session_Start only fires when a new browser/agent hits the site. You don't know the user at that point.
The methods in the Global.asax were not specifically designed for 'bootstrapping' state-loading, they're just convenient events for doing whatever you want with. You don't have to use them at all, mine usually just contain logging so I know when sessions start etc.!
I don't know if this helps.
Once you have a plan, come back and ask a targeted question about the OWIN stuff.
For manually testing an HTTP client in my application, I'd like to use a tool which starts an HTTP server my application can connect to and that lets me respond to request from my application manually. I'm basically looking for a tool with a GUI that lists all incoming requests and allows me to select a status code and type a response message. I've already tested the functionality with unit tests but I also want to verify it manually with no mocking etc.
Sounds simple but I didn't find such a tool. I've found some that can be scripted but no interactive one. Do you know one?
This can probably be written relatively easily by creating the Swing GUI dialog popup inside the servlet servicing methods. Have never seen Tomcat running this way but probably it would. Only, mind the server time out. It may be not long enough for you to make an input and require to be configured, also on the client side. Also, parallel calls will make multiple popups that may be difficult to respond but probably this is a single client app.
As a simplest solution, server GUI can be simply disposed after call and newly created as the next call arrives. This will make eveything indepenent on how servlet container is managing the servlets (does it stays, is it destroyed, maybe class is unloaded, who knows). Advanced solution could include the "server servlet" that would interact through its own JSP page but then it may be complex to update it when the new call arrives (unless maybe refresh periodically).
I'm currently working with web services that return objects such as a list of files e.g. File array.
I wanted to know whether its best practice to bind this type of object directly to my front end code for example a repeater/listview or whether to first parse it into my own list of "file class" e.g. customFiles[]
If the web service changes then it will break my front end code, however if I create my own CustomFile class, then i would only need to change my code in one place to fix the issue, but it just seems like a lot of extra work to create the same classes from a web service, i wanted to know what is the best practice for this type of work.
There is a delicate balancing act in properly encapsulating implementation details. Too little encapsulation is a maintenance nightmare as small changes in any area break the application. Too many layers is a different kind of maintenance headache altogether.
In this particular case I would create a small layer in your application to encapsulate the web service calls. This will ease your maintenance in both the application and the service as they will be loosely coupled.
It sounds like you have already answered your own problem. Best practice is to create your own custom class for the reasons you point out, but it is significant extra work.
If the webservice isn't likely to change then just use the existing classes, but if you need to cater for change then create your own.
Returning a class is fine as long as your client knows how to deserialize it. If it's truly a web service, where you don't have control over both ends of the conversation, it's more common to start with schemas for XML request and response streams. That decouples the client from the web service a bit more and allows any client that can send XML via HTTP and consume an XML response fair game.
I'm inheriting a legacy project, and there's a page that calls a method that makes a web service call. Load and performance testing has detected that this page sometimes takes a really long time to close, but usually it's fine, and if there is one hanging, all other requests for that page hang until the first one resolves, and then they all resolve.
I think this might have something to do with the web service being instantiated and not disposed, but I don't know how I can be more sure. I tried to add a delegate to the Dispose method but that doesn't seem to ever fire. (I'm not sure it would without Dispose being called explicitly, but that doesn't really matter.)
So what can I look for on the production server or any deployed environment to watch those requests pile up and go out (or get handled in an orderly manner, if they aren't the problem)?
You might consider using a tool like .NET Memory Profiler. You can use it to attach to your running application, and it can find and report all undisposed objects.
I think they have a free two week trial.
In your web app code, you could write to a log (event log, text file) right before you send a request, and again right after you get the response. Include some identifying information about the request. Add timestamps.
If you only want to use this while debugging, wrap it in #debug.
Or if you want to test in release mode, you could put a boolean in appSettings in web.config so you could turn the logging off and on.