We Keep Coding

How to test private member objects without making a new object

I'm trying to write unit test against a class. I can't change the class, but I think it's possible to test using reflection. I just don't know how to do it. Here's the class:
public class MyClass extends AnotherClass implements TheInterface
{
private enum SomeTypes
{
SAMPLE01, SAMPLE02, SAMPLE03
}
private CircularList<SomeTypes> someTypesList;
Date date= new Date();
private SomeOtherClassProcessor01 someOtherClassProcessor01;
private SomeOtherClassProcessor02 someOtherClassProcessor02;
private SomeOtherClassProcessor03 someOtherClassProcessor03;
public Properties initialize (Properties properties) throws Exception
{
Properties propertiesToReturn = super.initialize(properties);
someTypesList = new CircularList<SomeTypes> (Arrays.asList(SomeTypes.values()));
someOtherClassProcessor01 = new SomeOtherClassProcessor01();
someOtherClassProcessor02 = new SomeOtherClassProcessor02();
someOtherClassProcessor03 = new SomeOtherClassProcessor03();
return propertiesToReturn;
}
#Override
public void get(ImportedClass someParams) throws Exception
{
SomeTypes types = someTypesList.getFirstAndRotate();
switch(types)
{
case SAMPLE01:
someOtherClassProcessor01.doSomething(someParams, date);
break;
case SAMPLE02:
someOtherClassProcessor02.doSomething(someParams, date);
break;
case SAMPLE03:
someOtherClassProcessor03.doSomething(someParams, date);
break;
default:
throw new IllegalArgumentException("This " + types + " was not implemented.");
}
}
}
For my test this is what I have so far... not sure how to actually do it.
#RunWith(PowerMockRunner.class)
#PrepareForTest(MyClass.class)
public class TestingMyClass
{
MyClass mockMyClass;
SomeOtherClassProcessor01 someOtherClassProcessor01;
SomeOtherClassProcessor02 someOtherClassProcessor02;
SomeOtherClassProcessor03 someOtherClassProcessor03;
Date date;
#Before
public void initialize () throws Exception
{
mockMyClass = spy(new MyClass());
mockSomeOtherClassProcessor01 = mock(SomeOtherClassProcessor01.class);
mockSomeOtherClassProcessor02 = mock(SomeOtherClassProcessor02.class);
mockSomeOtherClassProcessor03 = mock(SomeOtherClassProcessor03.class);
}
#Test
public void testingGet() throws Exception
{
date = new Date();
//this is where I'm stuck
Whitebox.setInternalState(mockMyClass, "someOtherClassProcessor01", mockSomeOtherClassProcessor01);
}
}
Would it be possible to use whitebox for this? I need to make sure that there's a call inside the getter for those objects. Should I try something like when(someOtherClassProcessor01.doSomething(any(), date)).thenReturn(true)? Please let me know if you need more details.
edit: is even possible to mock private enum SomeTypes?

One option is to substitute your own (mocked) implementations of SomeOtherClassProcessor into MyClass using reflection:
MyClass myClass = new MyClass();
SomeOtherProcessor01 mockProcessor01 = mock(SomeOtherProcessor01.class);
// reflection bit: find the field by its name
// handle NoSuchFieldException
Field someProcessorField = MyClass.getDeclaredField("someOtherProcessor01");
// the field is declared as private, so make it accessible in order to work with it
someProcessorField.setAccessible(true);
// now set your mocked processor into the field.
// First argument is the object to change; second argument - new value for the field
someProcessorField.set(myClass, mockProcessor01);
PS. Using PowerMock and/or reflection is surrender to bad design (as per Timothy :). You should not be depending on code you that isn't already well-tested, and if it is, you shouldn't try to test it again. Suppose your testing actually reveals a bug - how would you fix it if you don't control the code? Suppose Java 11 becomes a thing and prohibits your use of reflection. Suppose the code you're testing changes and the fields get renamed - with reflection, you don't have compile-time safety... List of potential issues goes on

NullPointerException when creating new String

SOLUTION
I was adding a new instance of A to the list: aList.add( new A() ), whose name property is of course null, instead of adding the actual initialized instance. Sorry for the dumb question.
I have this A class of ProjectA which overrides its toString method to return a clone of its name property, like this:
public class A {
private String name;
#Override
public String toString() {
return new String(name);
}
}
I then export this library to a jar and import it into ProjectB and when I call a.toString() I get a NullPointerException that says there is an error exactly on the return line: return new String(name);.
However, if I put it like this:
public class A {
private String name;
#Override
public String toString() {
return name;
}
}
I don't get any exception but the String returned is null.
I built the jar (ProjectA) using Eclipse and imported it into ADT (ProjectB - Eclipse too).
NOTE:
I omitted the getters/setters intentionally for the sake of simplicity, but they're in there in the original code and I'm pretty sure I set the name property way before calling the toString() method. In fact, if I call the getName() method, the name is returned perfectly fine, but I'm using lists and I need the toString() method.
This is the part of the code where the List of A objects is created (ProjectA too):
ArrayList<A> aList = new ArrayList<Categoria>();
for (int i = 0; i < random.nextInt(3)+1; i++) {
A a = new A();
a.setId(0);
a.setName("Test name");
a.setDescription("Test desc.");
aList.add(a);
Log.d("app", "Created object A: "+a.getName()); // The name is displayed OK here
}
aList.trimToSize();
And this is the exact part of the code where the toString() method is called (ProjectA):
new ArrayAdapter<A>(
actionBar.getThemedContext(),
android.R.layout.simple_list_item_1,
android.R.id.text1,
DataAccess.getAList() // The ArrayAdapter class calls the toString method to populate a list
)
As you could see, I in fact verify the content of the name property via the getName() method and it is okay. I have to mention that the first approach (the on which uses new String(name), without an finals nor null checks) worked flawlessly on another project.
This is the (relevant part of the) DataAccess class:
public final class DataAccess {
private static final Data data;
public static Arrayist<A> getAList() {
return this.data.getAList();
}
}

When you invoke new String(name); it invokes the overloaded parameterized constructor of String shown below :
public String(String original) {
int size = original.count;
........
.........
}
As you can see the first line in the code tries to compute the length of the String passed to the constructor. Since in your case the String is null invoking the member variable count on that null reference throws NullPointerException.
Note : In the code where you create AList, i dont see you adding the object to the list i.e. AList.add(a); is missing

return new String(name);
Would fail on name being null. It could happen on A a = new A(); System.out.println(a):.
The used constructor String(String) is a relict of the beginning of Java, reminiscent of the C++ copy constructor.
In Java String is immutable, and Java does not need the copy constructor.
Two solutions:
return String.valueOf(name); // Would return "null" for null.
return name == null ? "" : name;

As far as the exception goes you are getting a null pointer as already explained by #Kakarot.
This line blows up
original.count
But if you want to save yourself from null checking etc and at the same time have a efficient toString() method than user some thing like this.
public class A {
private String name = null;
private String address = " Some place";
#Override
public String toString() {
StringBuilder sb = new StringBuilder();
return sb.append(name).append(address).toString();
}
}
Using string builder is efficient and you can see the output for the above toString() even with null values.
null Some place

In java new String(null) results into NullPointerException.
#Override
public String toString() {
if(name!=null){
return name;
}else{
return "";
}
}

Constructor not defined during Test

I am having a pesky error testing this piece of code.Not sure what I am doing wrong. I have tested standard controllers before but not the getSelectedMethod
Error:
Constructor not defined:[myClass].(ApexPages.StandardController)
public class MyClass{
public List<Web__c> postings {get;set;}
public static final String POSTINGSTATUS = 'Yes';
public MyClass (ApexPages.StandardSetController controller) {
List<Web__c> selectedWeb = (List<Web__c>) controller.getSelected();
postings = [Select Id, Name,
FROM Web___c Where Id IN: selectedWeb
AND Date__c != null
AND Date__c > today];
}
}
My Test Class
private class myTestClass {
static testMethod void WebTest() {
List <Web__c> posting = helper.createWeb(4);
insert posting;
Test.startTest();
PageReference pageRef = Page.VFPAGE;
Test.setCurrentPage(pageRef);
ApexPages.StandardController sc = new ApexPages.standardController(posting[0]);
myClass JPC = new myClass(sc);
sc.setSelected(posting);
Test.stopTest();
}

new ApexPages.standardController(posting[0]) looks wrong to me.
It should likely be new ApexPages.StandardSetController(posting[0]) if you want to create a new instance of ApexPages.StandardSetController, or just ApexPages.standardController(posting[0]) (without the new) if it's a static method that returns such an instance.
BTW: you should state what error you get and where in the code it occurs when asking such a question, as figuring it out without that information is pretty hard. I just found this by chance, I'd asked for that information otherwise.

Why cant I say variable.array in this case?

I don't see why I cant do this, without getting an error saying "Songs cannot be resolved or is not a field". I am a noob by the way, trying to learn this stuff :) Thanks in advance for you time, and please tell me if you need more information.
import java.util.ArrayList;
public class Band {
public String bandName;
public ArrayList<String> musician = new ArrayList<String>();
public ArrayList<String> songs = new ArrayList<String>();
// Constructor
public Band(String bandName) {
this.bandName = bandName;
}
public void getBandSongs(String bandName){
for (String s : bandName.songs) { <<<<<<<<<<<ERROR HERE
String rating = s.substring(0,1);
s = s.substring(1);
System.out.println("Rating: " + rating + " - Song name: " + s);
}
}
}

Sometimes the errors that the IDE gives you are misleading, but this one is quite clear.
So, let's analyse it:
"Songs cannot be resolved or is not a field"
So that should ring up an alarm. First thing you should do is look at the type that is your variable bandName.
public void getBandSongs(String bandName){
for (String s : bandName.songs) ...
It's a String! Of course you won't be able to access a field "songs" of a type String.
Maybe in your method signature you meant to have the following:
public void getBandSongs(Band band)
In that case, you should be able to access band.songs just fine.
Or you could have meant the following:
for (String s : this.songs)
That means you would be accessing the "songs" variable of the object instantiation of the class Band.
In summary:
The attributes that you define in your class:
public String bandName;
public ArrayList<String> musician = new ArrayList<String>();
public ArrayList<String> songs = new ArrayList<String>();
can be accessed through a variable that is of type of that class (in this case Band).
Hope this has made it a bit clear. Are you following any book in particular? I recommend the O'Reilly series. Good luck!
ps: I don't want to add too much since you are starting. But I advise you to read up on "encapsulation". That means that, unless strictly necessary, your should, by default, make your class's arguments private and let other classes access them through "getters" and "setters". Such that:
public class Band {
private String bandName;
private ArrayList<String> musician = new ArrayList<String>();
private ArrayList<String> songs = new ArrayList<String>();
public String getBandName(){
return this.bandName;
}
public void setBandName(String bandName){
this.bandName = bandName;
}
//And like that for the other two attributes. That way the classes that need access
// to these will either use a "get" method or a "set" method without directly
// accessing the attributes.
}
The work of creating getters and setters is so redundant that both Eclipse and Netbeans IDEs have a functionality that allows you to do these automatically.

songs is a part of Band, not bandName.
use this.songs instead (or just songs).
It should be noted though that public fields are against encapsulation (and thus OOP). Is this really what you want?
The major benefit of encapsulation (providing getters & setters for instance members) is to have a unified way of accessing your fields in a class. This allows you to, for example, add validation logic to your data.
private List<String> songs = new ArrayList<>();
public List<String> getSongs(){
return songs;
}
public List<String> setSongs(List<String> songlist) {
this.songs = songlist;
}
Now you can add validation to these methods. For example if you want to make sure you can only set the songlist if it has at least 5 songs in it:
public List<String> setSongs(List<String> songlist) {
if(songlist.size() > 5) {
this.songs = songlist;
}
}

Java exception crshes play framework

I have following controller:
public static Result overview() {
class Earning {
public int ammount;
public String description;
}
Earning[] earnings = new Earning[5];
earnings[0].ammount = 5;
return ok(overview.render(earnings));
}
I didn't created corresponding object in array as a result in Java I should get: java.lang.NullPointerException
But instead of showing this error Play framework crashes.
Any ideas how not to crash the framework and see the error in first place?

Add:
earnings[0] = new Earning();
before:
earnings[0].ammount = 5;
This way, earnings[0] will hold an Earning object and you would be able to access its ammount field.

Don't declare the Earning class inside the overview method.
It's creating a visibility issue because this class should only be accessible in the method body, but your "leaking" it by passing it to a view.