This code:

class ClassType<U> {
    U fu;
    ClassType() {}
    ClassType(ClassType<U> other) {
        this.fu = other.fu;
    }
    ClassType(U fu) {
        this.fu = fu;
    }
    @Override
    public String toString() {
        return "ClassType: " + String.valueOf(fu);
    }
}

public class Hang {
    static <U> ClassType<U> m(ClassType<U> src) {
        return new ClassType<U>(src);
    }

    public static void main(String argv[]) {
        ClassType<Number> c = m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>(m(new ClassType<>())))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))));
        String result = c.toString();
        if (!result.equals("ClassType: null")) {
            throw new Error();
        }
    }
}

Takes forever to be compiled. Main reason naive implementation of some incorporation steps. A huge speed-up can be obtained if some heavy operations like type substitution are not executed for those cases when they imply a no-op