Understanding Page lifecycle is very crucial in order to develop ASP.NET applications. Most beginners tend to get confused while dealing with dynamic controls and face problems like losing values, state etc on postbacks. Since HTTP is stateless, the nature of web programming is inherently different from windows application development, and the Page lifecycle is one of the primary building blocks while learning ASP.NET. The sequence of events, especially while working with MasterPages in ASP.NET 2.0, has become slightly more complex and this article is aims to shed some light on these events by explaining the order and importance of each event.
Whenever the user requests a particular .aspx page in an application, a lot of interesting things happen on the web server where the application is hosted. Understanding this sequence of events will help us to program and respond to events properly and also clear any confusion which generally arises due to the stateless nature of web programming.
Basics: The New Compilation Model and the Partial Classes
Each web form in an ASP.NET application derives directly or indirectly from a System.Web.UI.Page class. A web form has two components: a code behind file (WebForm.aspx.cs) which contains the code for the events and other methods related to a Page, and the designer ASPX file, which contains HTML control declarations and events (in the Visual Studio 2005 Web Application project model, we have a designer class named WebForm.aspx.designer.cs).
In ASP.NET 2.0, we do not need to define the control variables as well as there event handlers in the code behind, thanks to Partial classes. In ASP.NET 1.x, all this code was auto generated and placed in the code behind file under InitializeComponent() section. But in version 2.0, the runtime will create a partial class dynamically from the ASPX page containing all this info and merge it with the code behind partial class. This will help in making the actual code behind class a lot cleaner and more manageable.
Also, this would eliminate the name change related issues which were common in VS 2003 (if we change any control's ID, it had to be changed everywhere and VS used to modify the code many times). All control related events are defined in the ASPX markup code. So having a single place for controls names and event handlers is cleaner and flexible, whereas the previous VS 2003 model was more "brittle".
Real Thing: The Page life cycle
It is very important to know that for each request, the Page class is instantiated everytime from “scratch”. Which means that any values or whatever state it had previously will get lost unless we use one of the various state maintainance mechanisms provided by ASP.NET like Application, Session, Cache variables or Cookies.
Side Note: View state in ASP.NET 2.0 has changed and now comprises of two parts: Control State and View state. Refer this article for details:
Below is the sequence of events which fire up sequentially with explanation on the relative importance with respect to web programming in code behind:
Important Note: All events except the Init() and Unload() are fired from outermost to the innermost control. For e.g., a user control’s init event would fire before the Page_Init() event of its parent Page class.
In this Page level event, all controls created during design time are initialized with their default values. For e.g., if you have a TextBox control with Text property = “Hello”, it would be set by now. We can create dynamic controls here.
This event occurs only for the Page class and UserControls/MasterPages do not have this method to override. Sample code where you can override this method and add your custom code:
protected override void OnPreInit(EventArgs e)
Note that PreInit() is the only event where we can set themes programmatically.
Special Case with MasterPages
It is important to note that Master Page is treated like a control in the Content Pages.So if a Page has a Master Page associated with it, then the controls on the page will not be initialized and would be null in this stage. Only after the Init() event starts, you can access these controls directly from the page class. Why?
The reason being that all controls placed in the Content Page are within a ContentPlaceholder which is a child control of a MasterPage. Now Master Page is merged and treated like a control in the Content Pages. As I mentioned earlier, all events except the Init() and Unload() are fired from outermost to the innermost control. So PreInit() in the Page is the first event to fire but User Controls or MasterPage (which is itself a Usercontrol) do not have any PreInit event . Therefore in the Page_PreInit() method, neither the MasterPage nor any user control has been initialized and only the controls inside the Page class are set to their default values. Only after the Page_PreInit() event the Init() events of other controls fire up.
See the diagram below showing control hierarchy after the Page_Init() event:
In this event, we can read the controls properties (set at design time). We cannot read control values changed by the user because that changed value will get loaded after LoadPostData() event fires. But we can access control values from the forms POST data as:
string selectedValue = Request.Form[controlID].ToString();
This will only fire if the Page has posted back (IsPostBack == true). Here the runtime de-serializes the view state data from the hidden form element and loads all controls who have view state enabled.
Again, this method will only fire if the Page has posted back.In this event the controls which implement IPostBackDataHandler interface gets loaded by the values from the HTTP POST data. Note that a textbox control does not gets its value from the view state but from the post data in the form in this event. So even if you disable view state for a particular control, it can get its value from the HTTP POST data if it implements IPostBackDataHandler interface.
Also, an important point to note is that if we have a DropDownList control and we have dynamically added some items to it, the runtime cannot load those values unless the view state is enabled (even if the control derives from IPostBackDataHandler). The reason being that HTTP Post data has only one value per control, and the entire value collection is not maintained in the PostData but in view state.
This is the most popular method and the first one for all beginner developers to put their code. Beginners may also think that this is the first method which fires for a Page class. This can lead to a lot of confusion which makes understanding the Page lifecycle all the more important.
Note: If the page has any user control, then it's Load method will fire after the Page class's Load method. The reason as explained earlier is the fact that all method except the Init() are fired from the outermost control to the innermost. So after Page_Load(), load methods of all other controls are fired recursively.
6. Control Event Handlers
These are basically event handlers (like Button1_Click()) which are defined for controls in the ASPX markup. Another source of confusion arises when the developer thinks that an event handler like Button_Click() should fire independently (like in windows apps) as soon as he clicks a Button on the web form, forgetting that Page_Load will fire first before any event handlers.
This event is again recursively fired for each child controls in the Page. If we want to make any changes to control values, this is the last event we have
to peform the same.
Here, the ViewState of the controls gets saved in the form's hidden control.
In this method all controls are rendered recursively (i.e. Render method of each control is called).
Here you can have the page and controls perform clean-up operations. This event has no relevance besides clean up operations because the Page has already rendered.
This Article is taken from the Code Project Articles by Vivek Thakur can read at http://www.codeproject.com/useritems/lifecycle.asp