RE: JSF Lifecycle and Design Question

[CONNER, BRENDAN \(SBCSI\)]

Title: Message

Also, you shouldn't have to worry about populating the bean's values again after the save (unless you're changing them or re-querying the database), since the beans already have their values.  From what I recall (although it's been awhile since I've worked with Struts), this is one of the areas in which JSF makes our life a lot easier than with Struts.

 

You may want to use <t:saveState>, though, to make sure your bean values get passed from one request to another.

 

- Brendan

-----Original Message-----
From: CONNER, BRENDAN (SBCSI)
Sent: Wednesday, March 01, 2006 12:59 PM
To: MyFaces Discussion
Subject: RE: JSF Lifecycle and Design Question

I would advise against doing any kind of action as part of a managed bean's constructor.  In my experience, you want to separate out Action methods from Managed Bean data.  Your action methods should be the only things accessing the EJBs, etc.  So you should have an action method that gets invoked before the screen is displayed.  It goes out and gets the data and populates the data fields, which then get displayed.  Your Save then is very simple, since JSF automatically populates your bean values (at least on simple screens).  Your "save" method then just has to refer to the bean values when calling the back end to save the data.

 

In our applications, we make the distinction between Actions and Data very concrete in that we have Action managed beans that are separate from Data managed beans.  This helps keep the developers on the right track, although the same thing can be accomplished by having Action methods and data accessors within the same bean.

 

- Brendan

 

-----Original Message-----
From: j2ee Developer [mailto:[EMAIL PROTECTED]
Sent: Wednesday, March 01, 2006 12:31 PM
To: users@myfaces.apache.org
Subject: JSF Lifecycle and Design Question

I have JSF lifecycle and design question. I am building a prototype, porting an existing Struts based application to JSF. The issue I am facing is I believe related to the JSF lifecycle, I need input from you folks as to whether what I am doing is the right way.

Here is the scenario of my application -

I have a page with a form with values populated from the database. The end user makes changes to the fields, and presses the commandbutton save. The values are saved to the database and the same page presented again with the new saved values.

Here is how I have implemented it -

The constructor of the backing bean makes the call to the model layer ( eventually EJB's), gets the initial data and displays it to the page. After user makes the changes, on pressing the commandbutton, the event handler, makes a call to the modellayer for saving the new data. After the save I populate the backing bean attributes with the new values, and display the same page again. All my backing beans are request scope ( and that's how I want them). I am using Myfaces implementation, and myfaces components.

Here is the problem I face -

The issue I am facing is that on calling save, or any other event, the constructor is called again . So eventually it ends up that there are double calls to the modellayer/database than what is required.

Is the way I am populating my backing beans wrong? Or is this the only way JSF handles things. Any input would be appreciated.