Question

We usually write a click handler on a button (in GWT or Swing or Android) in the following way (way 1).

class A {

    public void myMethod() {

        int count = 0 ;

        Button myButton = new Button("x");
        myButton.addClickHandler(new ClickHandler() {
          public void onClick(ClickEvent event) {
              count = 1 ;
          }

        });

    }
}

However, we can also write a different class for the click handler and implement it like the following (way 2) :

class A implements ButtonClickInterface {

    int count ;
    public void myMethod() {

        count = 0 ;

        Button myButton = new Button("x");
        myButton.addClickHandler(new MyClickHander(this));

    }

    @Override
    public void buttonClicked() {
        count = 1 ;        
    }
}

interface ButtonClickInterface {
    public void buttonClicked() ;
}

class MyClickHandler implements ClickHandler {


    ButtonClickInterface buttonClickInterface ;

    public MyClickHandler(ButtonClickInterface buttonClickInterface) {
        this.buttonClickInterface = buttonClickInterface ;
    }

    @Override
    public void onClick(ClickEvent event) {
        buttonClickInterface.buttonClicked() ;         
    }

}

Here I have to make count as a class variable. Can this have any drawbacks? Also can the above way of implementation have any other drawbacks? How would the two ways in which I have implemented the ClickHandler face off against each other in terms of complexity, coupling, number of objects created and code readability ?

Answer 1

Your second solution is way too complex.

I often do that on Android (pseudocode):

class A implements OnClickListener {

    onCreate() {
        myButton.setOnClickListener(this);
    }

    @Override
    public void onClick(View v) {
        switch (v.getId()) {
            case R.id.myButtonId:
                // do something
                break;
            // more cases
            default: 
                Log.w(LOG_TAG, "not implemented");
        }
    }
}