Question

Assume this Java code:

public class A {
   public A(String g) {
      x += g.length();
   }

   private int x = 0;
}

If I create an instance of A, like this:

A a = new A("geo");

after this call, the value of x will be 3. What am I doing wrong in my Scala code?

class A(val g:String) {
  x += g.length
  var x:Int = 0
}

object x extends Application {
  val x = new A("geo")
  println(x.x)
}

This prints 0. I assumed that when the compiler reaches the var x:Int = 0, the body of the main constructor has ended. Am I wrong? How else could you declare instance variables in Scala ( assuming I don’t want them in my constructor ) ?

Answer 1

Keep in mind that your code translates into something similar (but not exactly) to this:

public class A {

  private final String g;
  private int x;

  public A(String g) {
    this.g = g;
    x_$eq(x() + g.length());
    x = 0;
  }

  public void x_$eq(int arg0) {
    x = arg0;
  }

  public int x() {
    return x;
  }

  public String g() {
    return g;
  }
}

But vars defined in (non-constructor) methods get translated into actual local variables.

Not sure if this explains the reasoning so much as highlights one of the differences.

---

EDIT – Changed “translation” from scala to java for clarity and ability to more accurately represent what is happening.