2

If in spring portlet controller, I have one methof annotated with @ActionMapping and another method annotated with @RenderMapping(params = "action=detail"), I see that to come from method #1 to method #2, we do not call the method#2 directly, instead we set a Actionresponse.setRenderParameter("action", "detail") in method #1 and method#2 will pick it up from here.

My first question is what advantages are we getting from separating the render phase from the Action Phase (or should I say Request Phase)? And also why we do not do a direct method call and instead call the method by setting a render parameter?

Victor
  • 16,609
  • 71
  • 229
  • 409

2 Answers2

3

To answer your questions

  1. Portlets can co-exist on a page. Irrespective of what action you performed on a portlet, the Portal container will invoke render() method on all the portlets to collect the html fragments and aggregate them, in an effort to prepare the whole page view and serve to the browser. Having 2 methods is not only an advantage but an enforcement to implement the aforementioned requirement which separates the business logic - processAction() that may change the state of your system as @Mark Chorley said, which you are aware of the execution "AND" render logic - render() which decides the view to render, which sometimes you are not aware of the execution.

  2. This is simple, why do you call start() rather then run() on a Thread though you can invoke run method directly. There are many reasons why you shouldn't call the method directly

    1. You are not responsible for the aggregation of html fragments from all the Portlets residing on the page.
    2. You are not the one who make the Public Render parameters available in the render() method of the interested Portlets.
    3. Definitely you dont want to take the headache of delegating the Events to the subscribed portlets
Niranjan
  • 1,776
  • 1
  • 13
  • 21
1

This is what portlets do basically. As to what advantage you get - well it allows you to separate your action (typically modifying data) from preparing data for your view.

Both Action and Render are different stages of a single request. Action requests precede Render requests as you say. "Request phase" sounds a confusing term to me. Action Phase and Render Phase are more helpful terms.

Why you don't call it directly - well the portlet container will do it for you. Also you must remember that the render methods will be called on all the visible portlets, but the action request will target only your portlet.

So given that the portlet container is guaranteed to call all the render methods on all the visible portlets, it doesn't make sense to call the render method manually on one portlet just because the action request targeted it.

If you only have one portlet visible this is perhaps less clear. It makes more sense when you have multiple portlets on a single page.

Mark Chorley
  • 2,087
  • 2
  • 22
  • 29
  • Thank you. But can you elaboratea bit more?"if issuing an action request meant that the render request was not called on that, which is how it would be if you manually called the render method" – Victor Apr 16 '13 at 20:36