Why should hidden_pum1_1 fail?

[tsdh]

This refactoring tries applyPullUpMethod(hidden.program.one.ParentClass, method).

ParentClass has two subclasses:

// ChildClass1.java
public class ChildClass1 extends ParentClass {
    String field = "I am a field";

    public static void main(String[] args) {
        ChildClass1 c = new ChildClass1();
        System.out.println("Field: "+c.field);
        System.out.print("Method: ");
        c.method();
    }

    public void method(){
        System.out.println("Hello World!");
    }
}
// ChildClass2.java
public class ChildClass2 extends ParentClass {
    public void method(){
        System.out.println("Hello World!"+"I am a field");
    }
}

There is no way for a solution to see that these two method() definitions aren't equivalent. They both have the same signature and access only the same method PrintStream.println(). So to my best understanding, method() is pullable.

[SvenPeldszus]

Oh sorry, in my version the program looks a bit different. I've pushed the wrong version.

It should be:

// ChildClass1.java
public class ChildClass1 extends ParentClass {
    String field = "I am a field";

    public static void main(String[] args) {
        ChildClass1 c = new ChildClass1();
        System.out.println("Field: "+c.field);
        System.out.print("Method: ");
        c.method();
    }

    public void method(){
        System.out.println("Hello World!"+c);
    }
}
// ChildClass2.java
public class ChildClass2 extends ParentClass {
    public void method(){
        System.out.println("Hello World!"+"I am a field");
    }
}

The field access got lost.

[tsdh]

Ok, then it makes sense.

[SvenPeldszus]

Does this case now work as expected?

[tsdh]

Yes, so I'm closing this issue.