方法调用的ManagedBean和xhtml设计策略

Question

Suppose that I have a class like this:

@ManagedBean
public class A {

@ManagedProperty("#{B}")
private B b;

private Object o;

public void operation() {
    b.methodCall(o);
}

public void setB(B b) {
    this.b = b;
}

public Object getO() {
    return o;
}

public void setO(Object o) {
    this.o = o;
}

}

Now when I make the method call from xhtml I can simply write:

<p:commandButton actionListener="#{a.operation()}" />

However I can also add a getter to my managed bean:

public B getB() {
    return b;
}

and call the same method like this removing the need to write the operation() method in the first version of my managed bean.

<p:commandButton actionListener="#{a.b.methodCall(b.o)}" />

It seems to me that the first approach would be easier while refactoring because of the ide support in refactoring java classes, but at the same time it makes me write a trivial method which only serves as a mean to convey parameters.

Which one is the better approach?

Answer 1

I think there is no straight forward answer for this question. The approach should be based on our requirement. Both has its own advantages and disadvantages.

But one thing i can say is :- your first approach is more feasible. Because in your second approach you should have a reference of b inside a, which is not required until unless it is really required [ A need not know about B , if be is an independent entity]. And you can get the instance of B from spring container (there by you can avoid the duplicate instance of B in application ) provided B is not in view scope.

And second problem is when you use abmethod, we should make sure that b cannot be null at any point of time. When we have a different entity reference in one html, the application will be come tightly coupled. And A html cannot be used with any other entity, if you have any plan to use it as a generic html.

hope i'am clear with my explanation.