explain why an object has lifetime outside scope - asp.net

I've written a class as follows. I've written it taking into consideration that its mainly for use on the web (i.e. it will be used from aspx pages).
public class TestHelper
{
public TestHelper()
{
HttpContext ctxt = HttpContext.Current;
IHttpHandler RequestHandler = ctxt.Handler;
Page CurrentPage;
CurrentPage = (Page)RequestHandler;
CurrentPage.Unload += new EventHandler(CurrentPage_Unload);
Debug.Print("Open all connection here...");
}
void CurrentPage_Unload(object sender, EventArgs e)
{
Debug.Print("Close all connection here...");
}
}
And I've written my aspx page's code behind like this:
public partial class _Default : System.Web.UI.Page
{
protected void Page_Load(object sender, EventArgs e)
{
TestHelper helper = new TestHelper();
helper = null;
}
}
In spite of assigning helper to 'null' I find that CurrentPage_Unload() gets executed. Why is this behaviour so? What is this behaviour generally called?
The reason why I've written the class in that style was because I thought I could best manage my db connections in a central fashion in the class. Usually people would call methods on the object like helper.IsValid() followed by helper.ProfileExists() in the aspx code behind. Each of those methods would have their own db connection objects(IDbConnection), and their corresponding Open() & Close() calls to open/close db connection. I just felt that we should only do this only once in code. Hence I used the constructor to open db connections, and the Page object's unload event to close the connection objects. Are there any pitfalls to writing classes this way?

What is happening is you're attaching a delegate to the Unload event of the page. Even after your variable has been set to null, your page still exists and still has an Unload event which still holds a reference to the delegate you added to it.
To remove a delegate you need to use -= syntax.
CurrentPage.Unload -= new EventHandler(CurrentPage_Unload);

Assigning null to a variable does not end its actual lifetime. Because .NET uses a non-deterministic garbage collection system (where objects are periodically purged based on several criteria, rather than as soon as they fall out of scope), you cannot rely on an object ever being collected before the process that created it ends.
Furthermore, because you're attaching an event handler (meaning a delegate, which includes a reference to your newly-constructed instance) to another object, you're extending the viable lifetime of your object as well. As long as that event handler is attached, your object cannot be collected until the object it's attached to is eligible.

You have to think about what happens when you execute the line
helper = null;
You simply have a reference to an object. When you assign that object to null, you're simply setting your reference to null. Nothing has happened to the object. Consider the following code:
var first = new object();
second = first;
first = null;
Would second now be null? You can think of a reference as simply a number - it's simply the address of the object in memory.
The .NET garbage collector simply looks at objects and checks if there are any references to that object. If not, it will remove that object.

Related

How does the logic for "the reference to the object" work with ViewState?

I store a list to the ViewState. Whenever I do a change to the list, the ViewState is automatically updated. When I checked couple of related posts in StackOverflow I learned that because ViewState stores the reference to the List object.
However, I want to learn more about the details. I though the information stored in ViewState travels through PostBacks. Then, the whole list is supposed to be serialized and stored in the hidden field, is that right? If it stores the reference to the List object, then what is the lifetime of the object in the Server memory? Is it re-created each time server receives a request?
Can anyone explain the mechanism for me?
Let me give you a quick example:
public partial class WebForm1 : System.Web.UI.Page
{
protected void Page_Load(object sender, EventArgs e)
{
List<string> myList = new List<string> { "a", "b", "c" };
Response.Write(string.Join(";", myList.ToArray()) + "<br>");
ViewState["myList"] = myList;
myList.Add("d");
List<string> myListSavedInViewstate = (List<string>)ViewState["myList"];
Response.Write(string.Join(";", myListSavedInViewstate.ToArray()) + "<br>");
myListSavedInViewstate.Add("e");
Response.Write(string.Join(";", myList.ToArray()) + "<br>");
Response.Write(myList.GetHashCode() + "<br>");
Response.Write(myListSavedInViewstate.GetHashCode() + "<br>");
}
}
The output is:
a;b;c
a;b;c;d
a;b;c;d;e
34781257
34781257
So basically yes, Viewstate keeps a reference not a copy of the object (of reference types) throughout the page life cycle. However as soon as the Request ends the Viewstate is serialized in the hidden field and returned to the client and the Viewstate instance is dropped. When postback occurs, viewstate is deserialized and recreated. The recreated instance of viewstate is however only a shallow copy of all the objects it used to hold references to during page life cycle execution. That is why all classes which are permitted in the viewstate have to implement ISerializable interface (or have to have a Serializable attribute).
This is however only the tip of an iceberg ;) Here is a very nice write up about viewstate if you'd like to read more on the subject.
http://weblogs.asp.net/infinitiesloop/Truly-Understanding-Viewstate

Silverlight Custom Control and Databinding doesn't work properly in WP7

I'm trying to create my calendar control with databinding.
public partial class Calendar : UserControl
{
public static readonly DependencyProperty DateProperty =
DependencyProperty.Register("Date", typeof(DateTime),
typeof(Calendar), null);
public object Date
{
get { return GetValue(DateProperty); }
set
{
SetValue(DateProperty, value);
OnPropertyChanged("Date");
}
}
public Calendar()
{
// Required to initialize variables
InitializeComponent();
DayText.Text = ((DateTime)Date).ToString("dd");
MonthText.Text = ((DateTime)Date).ToString("MMM");
this.Loaded += new RoutedEventHandler(Calendar_Loaded);
this.GotFocus += new RoutedEventHandler(Calendar_Loaded);
}
void Calendar_Loaded(object sender, RoutedEventArgs e)
{
DayText.Text = ((DateTime)Date).ToString("dd");
MonthText.Text = ((DateTime)Date).ToString("MMM");
}
}
But When I create the listbox with this control, same calndar have the wrong date. I'm sure that the Date passed thorough databinding is correct but I don't understand why same calender show a different day (I'm noticed that is the day of a previous calendar control intance)
Thank you for supporting!
Hmm ... where do we start? Here's a few things I've noticed:
If you're using a dependency property, there's no need to call OnPropertyChanged from the Date property setter.
The dependency property declares the type as DateTime, but your public exposed property is of type object, which then requires you to cast it elsewhere.
If Calendar_Loaded is to be called in more situations than in response to the Loaded event (such as the GotFocus event, then I'd recommend that you call it something else, or create a method with a relevant name (e.g. UpdateDateParts) and call it from properly named separate event handlers.
Using fixed format specifiers when processing date strings does not localize well.
In addition to that, I'd suggest that you could implement the user interface in a manner that supports databinding (and re-templating) by using bindings and exposing the date parts of the Date dependency property instead of manually updating the Text property of some text blocks/boxes in event handlers. In fact, if you derive from Control instead of UserControl then you can create and actuall lookless control that has it's user interface defined by a style in themes\generic.xaml that can be re-defined by users of your control.
As for why the date is incorrect in different instances of your calendar control, we'd need to see some of your XAML/code to see how the control is being used and initialized to be able to provide a better answer. However, I thought the above was worth putting in an Answer, instead of trying to say it in a Comment.

Abstracting HttpContext Request and Session - thread safety

I have the following assemblies in my ASP.NET app:
Website - this is an ASP.NET website
ClassLib - this is just a class lib that contains all the business logic
Class Lib needs to interact with the HttpContext Session and Request objects. This is a code upgrade from an old ASP app, where I've hoovered all the VBScript that contained the logic and put it into VB.NET. We simply didn't have time to rewrite.
Instead of ClassLib interacting directly with HttpContext, which I thought was BAD and also prevented us from unit testing, I introduced the following abstraction layer:
Public Class Request
Private Shared _requestWrapper as IRequestWrapper
Public Shared ReadOnly Property RequestWrapper()
Get
If _requestWrapper Is Nothing Then
Throw New Exception("_requestWrapper is null. Make sure InitRequest() is called with valid parameters")
End If
Return _requestWrapper
End Get
End Property
Public Shared Sub InitRequest(ByRef requestWrapper As IRequestWrapper)
_requestWrapper = requestWrapper
End Sub
Public Shared Function GetVal(ByVal key As String) As Object
Return RequestWrapper.GetVal(key)
End Function
etc.
This means in the unit tests I can supply my own MockRequest object into this Request class, which is just a simple NameValue collection. The code in the ClassLib and the Website then simply use the Request class and are none the wiser that it isn't coming from the HttpContext, but rather this mock class.
When it comes to the real deal, I simply have the following (C#) class:
public class RealRequest : IRequestWrapper
{
public void Initialize(HttpContext context)
{
}
#region Implementation of IRequestWrapper
public object GetVal(string index)
{
return HttpContext.Current.Request[index];
}
etc.
This is initialised in Session_Start of global.asax in the Website, as follows:
protected void Session_Start(object sender, EventArgs e)
{
IRequestWrapper requestWrapper = new RealRequest();
WebSession.Request.InitRequest(ref requestWrapper);
}
I think this is similar to the Static Gateway pattern.
Now, I am aware of singletons and static vars in a multi threaded environment such as ASP.NET, but this is slightly different. When it gets down to the RequestWrapper.GetVal(), its actually going to the HttpContext for that running thread - and pulling the value from that.
Certainly, any concurrent tests that we do with multiple users hitting the same server have never shown up any strange behaviour.
I'm just looking for re-assurance that this is a sound design, and if not why not?
Thanks
Duncan
This is fine. We have a very similar case in our applications that either uses HttpContext if it exists or fake implementations otherwise.
The one thing to watch out for is that there is a very specific instance where HttpContext.Current will return a value but HttpContext.Current.Request will throw an exception when triggered by the Application_Start event. In framework code, you really don't know (or want to know) what triggered the call though.
Workaround for HttpContext.HideRequestResponse being internal? Detect if HttpContext.Request is really available?

Why does my ASP.Net static function's "context" crossover between user sessions?

I think I need some help understanding how static objects persist in an ASP.Net application. I have this scenario:
someFile.cs in a class library:
public delegate void CustomFunction();
public static class A {
public static CustomFunction Func = null;
}
someOtherFile.cs in a class library:
public class Q {
public Q() {
if (A.Func != null) {
A.Func();
}
}
}
Some ASP.Net page:
Page_Init {
A.Func = MyFunc;
}
public void MyFunc() {
System.IO.File.AppendAllText(
"mydebug.txt", DateTime.Now.ToString("hh/mm/ss.fff", Session.SessionID));
}
Page_Load {
Q myQ = new Q();
System.Threading.Thread.Sleep(20000);
mQ = new Q();
}
The idea is that I have a business object which does some operation based on a callback function at the UI level. I set the callback function to a static variable on Page_Init (in the real code version, in the Master page, if that makes a difference). I thought that every execution of the page, no matter what user session it came from, would go through that function's logic but operate on its own set of data. What seems to be happening instead is a concurrency issue.
If I run one user session, then while it is sleeping between calls to that callback function, start another user session, when the first session comes back from sleeping it picks up the session ID from the second user session. How can this be possible?
Output of mydebug.txt:
01/01/01.000 abababababab (session #1, first call)
01/01/05.000 cdcdcdcdcdcd (session #2, first call - started 5 seconds after session #1)
01/01/21.000 cdcdcdcdcdcd (session #1 returns after the wait but has assumed the function context from session #2!!!!!)
01/01/25.000 cdcdcdcdcdcd (session #2 returns with its own context)
Why is the function's context (meaning, its local data, etc.) being overwritten from one user session to another?
Each request to an asp.net site comes in and is processed on it's own thread. But each of those threads belong to the same application. That means anything you mark as static is shared across all requests, and therefore also all sessions and users.
In this case, the MyFunc function that's part of your page class is copied over top of the static Func member in A with every page_init, and so every time any user does a page_init, he's replacing the A.Func used by all requests.
Static data is shared among the entire application domain of your webapp.
In short, it's shared among all the threads serving requests in your webapp, it's not bound to a session/thread/user in any way but to the webapp as a whole.(unlike e.g. php where each request lives in its own isolated environment bar a few knobs provided - such as the session variable.)
I won't try to improve on the other answers' explanations of static members, but do want to point out another way to code around your immediate problem.
As a solution, you could make an instance-oriented version of your class A, store it in a page-level variable, and pass it to Q's constructor on page load:
public class MyPage: Page {
private A2 _a2;
// I've modified A2's constructor here to accept the function
protected Page_Init() { this._a2 = new A2(MyFunc); }
protected Page_Load() {
Q myQ = new Q(this._a2);
// etc..
}
}
In fact, if there's no pressing need to declare A2 earlier, you could just instantiate it when you create your instance of Q in Page_Load.
Edit: to answer the question you raised in other comments, the reason the variables are being shared is that the requests are sharing the same delegate, which has only a single copy of its variables. See Jon Skeet's The Beauty of Closures for more details.
One solution you might consider is using [ThreadStatic].
http://msdn.microsoft.com/en-us/library/system.threadstaticattribute(VS.71).aspx
It will make your statics per thread. There are cavaets however so you should test.
If you want the data to persist only for the current request, use HttpContext.Items:
http://msdn.microsoft.com/en-us/library/system.web.httpcontext.items.aspx
If you want the data to persist for the current user's session (assuming you have session state enabled), use HttpContext.Session:
http://msdn.microsoft.com/en-us/library/system.web.httpcontext.session.aspx

What's the best way to load highly re-used data in a .net web application

Let's say I have a list of categories for navigation on a web app. Rather than selecting from the database for every user, should I add a function call in the application_onStart of the global.asax to fetch that data into an array or collection that is re-used over and over. If my data does not change at all - (Edit - very often), would this be the best way?
You can store the list items in the Application object. You are right about the application_onStart(), simply call a method that will read your database and load the data to the Application object.
In Global.asax
public class Global : System.Web.HttpApplication
{
// The key to use in the rest of the web site to retrieve the list
public const string ListItemKey = "MyListItemKey";
// a class to hold your actual values. This can be use with databinding
public class NameValuePair
{
public string Name{get;set;}
public string Value{get;set;}
public NameValuePair(string Name, string Value)
{
this.Name = Name;
this.Value = Value;
}
}
protected void Application_Start(object sender, EventArgs e)
{
InitializeApplicationVariables();
}
protected void InitializeApplicationVariables()
{
List<NameValuePair> listItems = new List<NameValuePair>();
// replace the following code with your data access code and fill in the collection
listItems.Add( new NameValuePair("Item1", "1"));
listItems.Add( new NameValuePair("Item2", "2"));
listItems.Add( new NameValuePair("Item3", "3"));
// load it in the application object
Application[ListItemKey] = listItems;
}
}
Now you can access your list in the rest of the project. For example, in default.aspx to load the values in a DropDownList:
<asp:DropDownList runat="server" ID="ddList" DataTextField="Name" DataValueField="Value"></asp:DropDownList>
And in the code-behind file:
protected override void OnPreInit(EventArgs e)
{
ddList.DataSource = Application[Global.ListItemKey];
ddList.DataBind();
base.OnPreInit(e);
}
Premature optimization is evil. That being a given, if you are having performance problems in your application and you have "static" information that you want to display to your users you can definitely load that data once into an array and store it in the Application Object. You want to be careful and balance memory usage with optimization.
The problem you run into then is changing the database stored info and not having it update the cached version. You would probably want to have some kind of last changed date in the database that you store in the state along with the cached data. That way you can query for the greatest changed time and compare it. If it's newer than your cached date then you dump it and reload.
If it never changes, it probably doesn't need to be in the database.
If there isn't much data, you might put it in the web.config, or as en Enum in your code.
Fetching all may be expensive. Try lazy init, fetch only request data and then store it in the cache variable.
In an application variable.
Remember that an application variable can contain an object in .Net, so you can instantiate the object in the global.asax and then use it directly in the code.
Since application variables are in-memory they are very quick (vs having to call a database)
For example:
// Create and load the profile object
x_siteprofile thisprofile = new x_siteprofile(Server.MapPath(String.Concat(config.Path, "templates/")));
Application.Add("SiteProfileX", thisprofile);
I would store the data in the Application Cache (Cache object). And I wouldn't preload it, I would load it the first time it is requested. What is nice about the Cache is that ASP.NET will manage it including giving you options for expiring the cache entry after file changes, a time period, etc. And since the items are kept in memory, the objects don't get serialized/deserialized so usage is very fast.
Usage is straightforward. There are Get and Add methods on the Cache object to retrieve and add items to the cache respectively.
I use a static collection as a private with a public static property that either loads or gets it from the database.
Additionally you can add a static datetime that gets set when it gets loaded and if you call for it, past a certain amount of time, clear the static collection and requery it.
Caching is the way to go. And if your into design patterns, take a look at the singleton.
Overall however I'm not sure I'd be worried about it until you notice performance degradation.

Resources