I've got an IoC container doing some complicated object construction when resolving some interfaces. I want to use implementations of these interfaces in my unit/integration tests. Is there anything wrong with resolving these interfaces in the tests using the IoC container or is it expected that one should build up instances manually in this situation?
When we are unit testing a class we are concerned with 'does the class do what we want it to do'.
Our starting point is a fully constructed instance; how we got there is not a unit testing question though it may be considered an integration testing question.
Say we have,
A
A(IB b, IC c)
B : IB
B(ID d)
C : IC
D : ID
Where:
IB is an interface for B,
IC is an interface for C, and
ID is an interface for D.
When unit testing A, the fact that B uses ID should be moot (if it is not then we should look at our interfaces. Having A access IB.D.xxx() is not good), all we need to do is provide A with some implementation (mocked/stubbed) of IB and IC.
As far as the unit tests for A are concerned, whether the A instance is created by hand or by a container is not important. We get the same object either way.
As long as we are passing in mocks as the first level dependencies then there is no saving when using a container over creating the object by hand. The saving only happens when we are creating object graphs using the IOC, but if we are doing this then we are into integration testing. This is not necessarily a bad thing but we need to be clear on our goals.
When unit testing the above we create unit testing for
D
C
B (passing in a mocked/stubbed ID)
A (passing in mocked/stubbed IC and IB)
When unit testing A we do not need the correct answer from D to be passed through B up to A.
All we care is that the logic in A works as expected, say, A calls IB.x() with the parameters y and z and returns the result of IB.x(). If we are checking that we get the correct answer (say, one which depends on logic in D) then we are past unit testing and into integration testing.
Bottom Line
It does not matter whether or not the unit under test was created by an IOC container or by hand as long as we are properly isolating the unit under test. If we are using the container to create an object graph then the odds are good that we are into integration testing (and/or have problems with too much coupling between classes - A calling IB.D.xxx())
Mocking for Integration Tests
Caveat: Some of this following is dependent upon the IOC container in use. When using Unity, the last registration 'wins'. I do not know that this holds true for others.
In the minimalist system under question we have
A
A (IB b)
B : IB
B is our 'leaf'. It talks to the outside world (say, reads from a network stream).
When Integration testing, we want to mock this.
For the live system, we set up the ServiceLocator using CreateContainerCore().
This includes the registration of the 'live' implementation of IB.
When executing integration tests that require a mocked version of IB we set up the container using CreateContainerWithMockedExternalDependencies() which wraps CreateContainerCore() and registering a mocked object for IB.
The code below is heavily simplified but the shape extends out to as many classes and dependencies as required. In practice, we have a base test class that aids setting up the service locator, mocking/stubbing classes, accessing the mocks for verification purposes and other house keeping (e.g.so that each test doesn't need to explicitly set the ServiceLocator provider)
[TestClass]
public class IocIntegrationSetupFixture {
[TestMethod]
public void MockedB() {
ServiceLocator.SetLocatorProvider(() => new UnityServiceLocator(CreateContainerWithMockedExternalDependencies()));
var a = ServiceLocator.Current.GetInstance<A>();
Assert.AreEqual("Mocked B", a.GetMessage());
}
[TestMethod]
public void LiveB() {
ServiceLocator.SetLocatorProvider(() => new UnityServiceLocator(CreateContainerCore()));
var a = ServiceLocator.Current.GetInstance<A>();
Assert.AreEqual("Live B", a.GetMessage());
}
private IUnityContainer CreateContainerCore() {
var container = new UnityContainer();
container.RegisterType<IB, B>(new ContainerControlledLifetimeManager());
return container;
}
private IUnityContainer CreateContainerWithMockedExternalDependencies() {
var container = CreateContainerCore();
var mockedB = new Mock<IB>();
mockedB.SetupGet(mk => mk.Message).Returns("Mocked B");
container.RegisterInstance<IB>(mockedB.Object);
return container;
}
public class A {
private IB _b;
public A(IB b) {
_b = b;
}
public string GetMessage() {
return _b.Message;
}
}
public interface IB {
string Message { get; }
}
private class B : IB {
public string Message {
get { return "Live B"; }
}
}
}
Related
Well, I'm looking for the best way to refactor a (huge) legacy code-base and introducing some tests in it..There was no test framework. (yeah, I mean not at all)
It was an JEE 5 application. The goal is to revamp that in JEE7
Let me introduce a quick overview .
The end-users (those of them who are authorized) are free to evolve , configure many aspect of the application behavior by setting in the UI a bunch of preferences.
Theses are stored in an SQL table for the main part (the rest in some xml and properties files).
To fulfill this requirement, there is an #Startup object dedicated to build a sort-of huge map with all key-values.
Then all across the code base when a use case needs to adapt it's processing it checks the current value of the parameter(s) needed to its task.
A real case is that the app has to do a few operations on images;
For instance, Class ImgProcessing has to create thumbnail of a picture via this method :
Optional<Path> generateThumb_fromPath(Path origImg);
for this the method generateThumb_fromPath, calls Thumbnailer,
which uses a generic ImageProcessingHelper,
which holds a few set of generic image related tools and methods,
and specially an static method returning the wished dimensions of the thumbnail to be generated based on the original image constraints and some thumbnail preferences (keys = "THUMBNAIL_WIDTH" and "THUMBNAIL_HEIGHT").
These preferences are the user wishes for what size a thumbnail should have.
So far so good, nothing special.
Now the dark side of this :
The original JEE5 config loader is an bad old fashioned infamous singleton pattern as :
OldBadConfig {
private static OldBadConfig instance ;
public static getInstance(){
if(instance==null){
// create, initialize and populate our big preferences' map
}
return instance;
}
}
Then all across the whole code-base these preferences are used. In my refactoring effort I've already done using #Inject for injecting the singleton object.
But in static utilities ( no injection point available ) you have lots of this nasty calls :
OldBadConfig.getInstance.getPreference(key, defaultValue)
(Briefly I will explain that I use testNg + Mockito, I don't think the tools are relevant here, it seems to be more about an original terrible design,
but if I HAVE to change my toolbox (Junit or whatever) I will. But again I don't think the tooling is the root problem here. )
Trying to refactor the image part and make it test-friendly., I want to do this test with cut = instance of my Class Under Test:
#Test
public void firstTest(){
Optional<Path> op = cut.generateThumb_fromPath(targetPath);
// ..assertThatTheThumbnailWasCreated........
}
So in a few words ,
the execution flow will be like :
class under test --> some business implementation --> someutilities --> static_global_app_preference ---> other_class-othermethod.finalizingProcessing,
then return to the caller.
My testing effort halts here. How to mock the static_global_app_preference ?
How can I refactor the static_global_app_preference part from
*OldBadConfig.getInstance.getPreference(key, defaultValue)*
to something mockable where I could mock like :
Mockito.when(gloablConf.getPreference("THUMBNAIL_WIDTH", anyString)).thenReturn("32");
I've spent quite a time reading boks, blog posts etc all saying
'(these kind of) Singleton is EVIL'. You should NOT do that !
I think we all agree , thanks.
But what about a real word and effective solution to such really trivial, common tasks?
I can not add the singleton instance (or the preferences'map ) as parameters (because as it is already spread all across the code-base it will pollute all and every classes and methods . For instance in the exposed use case, it will pollute 5 methods in 4 classes just for one, poor, miserable, access to a parameter.
It's really not feasible.
So far I tried to refactor OldBadConfig class in two part : one with all initialization/write stuff,
and the other with only the read parts. that way I can at least make this a real JEE #Singleton and benefits from concurrent access once the startup is over and the configuration all loaded.
Then I tried to make this SharedGlobalConf accessible via a factory, called like :
SharedGlobalConf gloablConf= (new SharedGlobalConfFactory()).getShared();
then gloablConf.getPreference(key, defaultValue); is accessible.
It seems to be a little better than the original bottleneck, but didn't help at all for the testing part.
I thought the factory will ease everything but nothing like that comes out.
So there is my question :
For myself, I can split the OldBadConfig to an startup artefact doing the init and refesh, and to an SharedGlobalConf which is a JEE7 pure Singleton,
#Singleton
#ConcurrencyManagement(ConcurrencyManagementType.BEAN)
#Lock(LockType.READ)
Then, as for the legacy use case described here, How Can I make this reasonably mock-able ? Real word solutions are all welcomed.
Thanks sharing your wisdom and skills !
I will like to share my own answer.
Let's say we got these classes after the initial large OldBadConfig class was splitted :
#Startup AppConfigPopulator in charge of loading all information and populating the kind-of internal cache,
which is now a distinct SharedGlobalConf object. The populator is the only one in charge of feeding the SharedGlobalConf via :
#Override
public SharedGlobalConf sharedGlobalConf() {
if (sharedGlobalConf.isDirty()) {
this.refreshSharedGlobalConf();
}
return sharedGlobalConf;
}
private void refreshSharedGlobalConf() {
sharedGlobalConf.setParams(params);
sharedGlobalConf.setvAppTempPath_temp(getAppTempPath_temp());
}
In all components (by that I mean all Classes holding valid injection points) you just do your classic
#Inject private SharedGlobalConf globalConf;
For static utilities that can not do #Inject, we got an SharedGlobalConfFactory which handles the shared data to everything in a one-liner :
SharedGlobalConf gloablConf = (new SharedGlobalConfFactory()).getShared();
That way our old code base can be smoothly upgraded : #Inject in all valid components, And the (too many) old utilities that we can not reasonably rewrite them all in this refactoring step can get these
*OldBadConfig.getInstance.getPreference(key, defaultValue)*
,simply replaced by
(new SharedGlobalConfFactory()).getShared().getPreference(key, defaultValue);
And we are test-compliant and mockable !
Proof of concept :
A really critical Business demands is modeled in this class :
#Named
public class Usage {
static final Logger logger = LoggerFactory.getLogger(Usage.class);
#Inject
private SharedGlobalConf globalConf;#Inject
private BusinessCase bc;public String doSomething(String argument) {
logger.debug(" >>doSomething on {}", argument);
// do something using bc
Object importantBusinessDecision = bc.checks(argument);
logger.debug(" >>importantBusinessDecision :: {}", importantBusinessDecision);
if (globalConf.isParamFlagActive("StackOverflow_Required", "1")) {
logger.debug(" >>StackOverflow_Required :: TRUE");
// Do it !
return "Done_SO";
} else {
logger.debug(" >>StackOverflow_Required :: FALSE -> another");
// Do it another way
String resultStatus = importantBusinessDecision +"-"+ StaticHelper.anotherWay(importantBusinessDecision);
logger.debug(" >> resultStatus " + resultStatus);
return "Done_another_way " + resultStatus;
}
}
public void err() {
xx();
}
private void xx() {
throw new UnsupportedOperationException(" WTF !!!");
}
}
To get it's job done , we need a hand from our old companion StaticHelper :
class StaticHelper {
public static String anotherWay(Object importantBusinessDecision) {// System.out.println("zz #anotherWay on "+importantBusinessDecision);
SharedGlobalConf gloablConf = (new SharedGlobalConfFactory()).getShared();
String avar = gloablConf.getParamValue("deeperParam", "deeperValue");
//compute the importantBusinessDecision based on avar
return avar;
}
}
Usage of this =
#Named public class Usage {
static final Logger logger = LoggerFactory.getLogger(Usage.class);
#Inject
private SharedGlobalConf globalConf;
#Inject
private BusinessCase bc;
public String doSomething(String argument) {
logger.debug(" >>doSomething on {}", argument);
// do something using bc
Object importantBusinessDecision = bc.checks(argument);
logger.debug(" >>importantBusinessDecision :: {}", importantBusinessDecision);
if (globalConf.isParamFlagActive("StackOverflow_Required", "1")) {
logger.debug(" >>StackOverflow_Required :: TRUE");
// Do it !
return "Done_SO";
} else {
logger.debug(" >>StackOverflow_Required :: FALSE -> another");
// Do it another way
String resultStatus = importantBusinessDecision +"-"+ StaticHelper.anotherWay(importantBusinessDecision);
logger.debug(" >> resultStatus " + resultStatus);
return "Done_another_way " + resultStatus;
}
}
public void err() {
xx();
}
private void xx() {
throw new UnsupportedOperationException(" WTF !!!");
}}
As you see the old shared key/value holder is still used every where but this time, we can test
public class TestingAgainstOldBadStaticSingleton {
private final Boolean boolFlagParam;
private final String deepParam;
private final String decisionParam;
private final String argument;
private final String expected;
#Factory(dataProvider = "tdpOne")
public TestingAgainstOldBadStaticSingleton(String argument, Boolean boolFlagParam, String deepParam, String decisionParam, String expected) {
this.argument = argument;
this.boolFlagParam = boolFlagParam;
this.deepParam = deepParam;
this.decisionParam = decisionParam;
this.expected = expected;
}
#Mock
SharedGlobalConf gloablConf = (new SharedGlobalConfFactory()).getShared();
#Mock
BusinessCase bc = (new BusinessCase());
#InjectMocks
Usage cut = new Usage();
#Test
public void testDoSomething() {
String result = cut.doSomething(argument);
assertEquals(result, this.expected);
}
#BeforeMethod
public void setUpMethod() throws Exception {
MockitoAnnotations.initMocks(this);
Mockito.when(gloablConf.isParamFlagActive("StackOverflow_Required", "1")).thenReturn(this.boolFlagParam);
Mockito.when(gloablConf.getParamValue("deeperParam", "deeperValue")).thenReturn(this.deepParam);
SharedGlobalConfFactory.setGloablConf(gloablConf);
Mockito.when(bc.checks(ArgumentMatchers.anyString())).thenReturn(this.decisionParam);
}
#DataProvider(name = "tdpOne")
public static Object[][] testDatasProvider() {
return new Object[][]{
{"**AF-argument1**", false, "AF", "DEC1", "Done_another_way DEC1-AF"},
{"**AT-argument2**", true, "AT", "DEC2", "Done_SO"},
{"**BF-Argument3**", false, "BF", "DEC3", "Done_another_way DEC3-BF"},
{"**BT-Argument4**", true, "BT", "DEC4", "Done_SO"}};
}
The test is with TestNG and Mockito : it shows how we don't need to do the complex stuff (reading the sql table, the xml files etc..) but simply mock different set of values targeting just our sole business case. (if a nice fellow would accept to translate in other frameworks for those interested...)
As for the initial request was about the design allowing to reasonably refactor a -huge- existing code-base away from the 'static singleton anti-pattern' , while introducing tests and mocks I assume this a quite valid answer.
Will like to hear about your opinion and BETTER alternatives
Mock objects introduce a good approach to do deep behavior testing of some program unit.
You just should pass mocked dependency to the tested unit and check if it works with dependency as it should do.
Let you have 2 classes A and B:
public class A
{
private B b;
public A(B b)
{
this.b = b;
}
public void DoSomething()
{
b.PerformSomeAction();
if(b.State == some special value)
{
b.PerformAnotherAction();
}
}
}
public class B
{
public BState State { get; private set; }
public void PerformSomeAction()
{
//some actions
State = some special value;
}
public void PerformAnotherAction()
{
if(State != some special value)
{
fail(); //for example throw new InvalidOperationException();
}
}
}
Imagine class B is being tested with unit test TestB.
To unit test class A we can either pass B to it's constructor (to do state based testing) or pass B's mock to it (to do behavior based testing).
Let say we have chosen the second approach (for example we can't verify A's state directly and can do it indirectly) and created unit test TestA (which doesn't contain any reference to B).
So we will introduce an interface IDependency and classes will look like:
public interface IDependency
{
void PerformSomeAction();
void PerformAnotherAction();
}
public class A
{
private IDependency d;
public A(IDependency d)
{
this.d = d;
}
public void DoSomething()
{
d.PerformSomeAction();
if(d.State == some special value)
{
d.PerformAnotherAction();
}
}
}
public class B : IDependency
{
public BState State { get; private set; }
public void PerformSomeAction()
{
//some actions
State = some special value;
}
public void PerformAnotherAction()
{
if(State != some special value)
{
fail(); //for example throw new InvalidOperationException();
}
}
}
and unit test TestB is something similar to:
[TestClass]
public class TestB
{
[TestMethod]
public void ShouldPerformAnotherActionWhenDependencyReturnsSomeSpecialValue()
{
var d = CreateDependencyMockSuchThatItReturnsSomeSpecialValue();
var a = CreateA(d.Object);
a.DoSomething();
AssertSomeActionWasPerformedForDependency(d);
}
[TestMethod]
public void ShouldNotPerformAnotherActionWhenDependencyReturnsSomeNormalValue()
{
var d = CreateDependencyMockSuchThatItReturnsSomeNormalValue();
var a = CreateA(d.Object);
a.DoSomething();
AssertSomeActionWasNotPerformedForDependency(d);
}
}
Ok. It's a happy moment for developer - everything is tested and all tests are green. Everything is good.
But!
When someone modifies logic of class B (for example modifies if(State != some special value) to if(State != another value) ) only TestB fails.
This guy fixes this test and thinks that everything goes well again.
But if you try to pass B to constructor of A A.DoSomething will fail.
The root cause of it is our mock object. It fixed old behavior of B object. When B changed its behavior mock didn't reflect it.
So, my question is how to make mock of B follow changes of behavior of B?
This is a question of viewpoint. Normally, you mock an interface, not a concrete class. In your example, B's mock is an implementation of IDependency, as is B. B's mock must change whenever the behaviour of IDependency changes, and you can ensure that by looking at all the implementations of IDependency when you change the defined behaviour of IDependency.
So, the enforcement is through 2 simple rules that ought to be followed in the code base:
When a class implements an interface, it must fulfill all defined behaviour of the interface after modification.
When you change an interface, you must adapt all implementers to fulfill the new interface.
Ideally, you have unit tests in place that test against the defined behaviour of an IDependency, which apply to both B and BMock and catch violations of these rules.
I differ from the other answer, which seems to be advocating subjecting both the real implementation and the (hand-made?) mock to a set of contract tests - which specify the behavior of the role/interface. I've never seen tests that exercise mocks - could be done though.
Normally you don't handcraft mocks - rather you use a mocking framework. So w.r.t. your example, my client tests would have inline statements as
new Mock<IDependency>().Setup(d => d.Method(params)
.Returns(expectedValue)
Your question is when the contract changes, how do I guarantee that the inline expectations in the client tests are also updated (or even flagged) with changes to the dependency?
The compiler won't help here. Nor would the tests. What you have is a lack of shared agreement between the client and the dependency. You'd have to manually find-and-replace (or use IDE tooling to locate all references to the interface method) and fix.
The way out is to NOT define a lot of fine-grained IDependency interfaces recklessly. Most problems can be solved with a minimal number of chunky roles (realized as interfaces) with clearly defined non-volatile behavior. You can attempt to minimize role-level changes. Initially this was a sticking point with me too - however discussions with interaction-test experts and practical experience have managed to win me over. If this does happen all too often, a quick retrospective as to the cause of fickle interfaces should yield better results.
I'm enjoying learning JAX-RS and Jersey, but I've hit a roadblock trying to test a simple resource that needs a DAO injected, something like this:
#Stateless
#Path("simple")
public class SimpleResource {
#PersistenceContext
private EntityManager em;
// #GET, etc...
}
(I'll be moving to a more abstracted DAO pattern, but the problem is the same, i.e., how do I inject the #EJB DAO?)
In my unit tests I'm using a embedded Jetty server that configures Jersey in its web.xml, and I'd like to hook into the resource's lifecycle so that I can inject a mock EntityManager, but I've not found a clean answer after a lot of searching. Can you help? Some possible directions I've come across:
1) Use JNDI context lookup in my code to get the DAO bean, and register the mock object in the tests.
Instead of #EJB or #PersistenceContext, use something like this in the resource's constructor:
theDAO = (DAOImpl) new InitialContext().lookup("java:global/EJB/DAOImpl");
However, that means my test environment needs to support JNDI, and doing so in Jetty will probably involve some pain. Plus, it doesn't use the clean annotation approach.
2) Use method injection.
Inject into the method so that I can set the DAO post-instantiation, e.g.,
#PersistenceContext(name = "persistence/pu00")
public void setPersistenceUnit00(final EntityManager em) {
em00 = em;
}
OR
private MyEjbInterface myEjb;
#EJB(mappedName="ejb/MyEjb")
public void setMyEjb(MyEjb myEjb) {
this.myEjb = myEjb;
}
However, to do this I need the Jersey-instantiated instance, e.g., SimpleResource. How do I get that?
3) Use reflection.
A kind of DIY injection, something like:
public static void setPrivateField(Class<? extends Object> instanceFieldClass, Object instance, String fieldName, Object fieldValue) {
Field setId = instanceFieldClass.getDeclaredField(fieldName);
setId.setAccessible(true);
setId.set(instance, fieldValue);
}
Again, I need the Jersey-instantiated instance.
4) Use an Injection Provider.
I'm still sketchy on how this works, but it looks like Jersey provides a means of defining customized injectable annotations, e.g.,
#Provider
public class EJBProvider implements InjectableProvider<EJB, Type> {
public ComponentScope getScope() {
return ComponentScope.Singleton;
}
public Injectable getInjectable(ComponentContext cc, EJB ejb, Type t) {
if (!(t instanceof Class)) {
return null;
}
try {
Class c = (Class) t;
Context ic = new InitialContext();
final Object o = ic.lookup(c.getName());
return new Injectable<Object>() {
public Object getValue() {
return o;
}
};
} catch (Exception e) {
e.printStackTrace();
return null;
}
}
}
A variation using a helper class:
Server server = new Server(8080);
Context root = new Context(server,"/",Context.SESSIONS);
ResourceConfig rc = new PackagesResourceConfig("edu.mit.senseable.livesingapore.platform.restws.representations");
rc.getSingletons().add(new SingletonTypeInjectableProvider<javax.ws.rs.core.Context, Myobj>(Myobj.class, new Myobj(12,13)){});
root.addServlet(new ServletHolder(new ServletContainer(rc)), "/");
server.start();
With this use:
#Path("/helloworld")
public class HelloWorldResource {
#Context Myobj myClass;
....
}
Is this viable for #EJB or #PersistenceContext?
5) Extend javax.ws.rs.core.Application.
Sketchy on this, but:
#javax.ws.rs.ApplicationPath("application")
public class InjectionApplication extends javax.ws.rs.core.Application {
private Set<Object> singletons = new HashSet<Object>();
private Set<Class<?>> classes = new HashSet<Class<?>>();
public InjectionApplication() {
// no instance is created, just class is listed
classes.add(BookResource.class);
}
#Override
public Set<Class<?>> getClasses() {
return classes;
}
#Override
public Set<Object> getSingletons() {
return singletons;
}
}
6) Extend ServletContainer.
An older style of using InjectableProvider? Looks more complex:
public class ServletAdapter extends ServletContainer {
#Override
protected void configure(ServletConfig servletConfig, ResourceConfig rc, WebApplication wa) {
super.configure(servletConfig, rc, wa);
rc.getSingletons().add(new InjectableProvider<Resource, Type>() {
public ComponentScope getScope() {
return ComponentScope.Singleton;
}
public Injectable<Object> getInjectable(ComponentContext ic, Resource r, Type c) {
final Holder value = new Holder();
Context ctx = new InitialContext();
try {
value.value = ctx.lookup(r.name());
} catch (NamingException ex) {
value.value = ctx.lookup("java:comp/env/" + r.name());
}
return new Injectable<Object>() {
public Object getValue() {
return value.value;
}
};
}
});
}
}
7) Use an embedded EJB container.
E.g., http://openejb.apache.org. This is pretty heavy, and I expect it's going to be messy to get working. (In fact, what started me down the "Jetty + Jersey" route was a bug in GlassFish Embedded around security logins. I also looked at other Java EE 6 application containers like JBoss AS, but each had problems in embedded mode, with limited user community support.)
8) Use a third-party IoC library like Spring or Guice.
Spring is apparently commonly used for solving these kinds of problems (injecting mocks when unit testing), but I wanted to avoid having to learn another big set of APIs - pure Java EE has been enough of a challenge! But I'm game if it's the best solution. I haven't yet looked carefully into Spring or Guice.
Have you used any of these successfully? Any other solutions you like? I'm really looking forward to your advice on this. Thanks in advance -- matt
Since you are using Netbeans, give this a try:
Using the Embedded EJB Container to Test Enterprise Applications
The tutorial uses an embedded Glassfish container and injects an EJB that encapsulates the EntityManager (similar to what you described in your first option).
If you only need an EntityManager inside your embedded Jetty container, why use injection in the first place? You can just put one of the JPA implementations (such as eclipselink or hibernate) on your class-path, configure a resource local persistence unit, then obtain it like this:
EntityManagerFactory emf = Persistence.createEntityManagerFactory("your unit name");
EntityManager em = emf.createEntityManager();
It would be enough to have something (maybe a static DAO factory?) that behaves like your #EJB for the purpose of testing your JAX-RS classes.
If you do want your unit tests to be as close to the Java EE environment as possible, look into running them with Arquillian (http://www.jboss.org/arquillian.html). It runs the tests directly on a Java EE container - it's easy, it has great documentation.
I've done writing code on salesforce and in order to release the unit tests have to cover at least 75%.
What I am facing is that the classOne that calls methods from classTwo also have to cover classTwo's unit test within classOne even though it is done in classTwo file already.
File MyClassTwo
public with sharing class ClassTwo {
public String method1() {
return 'one';
}
public String method2() {
return 'two';
}
public static testMethod void testMethod1() {
ClassTwo two = new ClassTwo();
String out = two.method1();
system.assertEquals(out, 'one'); //valid
}
public static testMethod void testMethod2() {
ClassTwo two = new ClassTwo();
String out = two.method2();
system.assertEquals(out, 'two'); // valid
}
}
File MyClassOne
public with sharing class ClassOne {
public String callClassTwo() {
ClassTwo foo = new ClassTwo();
String something = foo.method1();
return something;
}
public static testMethod void testCallClassTwo() {
ClassOne one = new ClassOne();
String out = one.callClassTwo();
system.assertEquals(out, 'one');
}
}
The result of testing MyClassOne would not return 100% test coverage because it says I have not covered MyClassTwo method2() part inside of MyClassOne file.
But I already wrote unit test for MyClassTwo inside of MyClassTwo file as you can see.
So does this mean I have to copy and paste the unit test in MyClassTwo file over to MyClassOne?
Doing so gives me 100% coverage but this seems really annoying and rediculous. Having same test in ClassA and ClassB....? Am I doing wrong or is this the way?
Having said, is it possible to create mock object in salesforce? I haven't figure how yet..
http://sites.force.com/answers/ideaView?c=09a30000000D9xt&id=087300000007m3fAAA&returnUrl=/apex/ideaList%3Fc%3D09a30000000D9xt%26category%3DApex%2B%2526%2BVisualforce%26p%3D19%26sort%3Dpopular
UDPATE
I re-wrote the code and updated above, this time for sure classOne test would not return 100% even though it is not calling classTwo method2()
Comments about Java mock libraries aren't very helpful in Salesforce world ;) At my projects we usually aimed for making our own test data in the test method, calling real functionality, checking the results... and whole test framework on Salesforce side is responsible for transaction rollback (so no test data is saved to DB in the end regardless whether the test failed or passed).
Anyway...
Masato, your classes do not compile (methods outside class scope, public String hello() without any String returned)... After I fixed it I simply right-clicked the MyClassA -> Force.com -> run tests and got full code coverage without any problems so your issue must lie somewhere else...
Here's how it looks: http://dl.dropbox.com/u/709568/stackoverflow/masato_code_coverage.png
I'm trying to think what might have gone wrong... are you sure all classes compile and were saved on server side? Did you put test methods in same classes as functionality or in separate ones (generally I make separate class name with similar name like MyClassATest). If it's a separate class - on which file did you click "run tests"?
Last but not least - if you're facing this issue during deployment from sandbox to production, make sure you selected all classes you need in the deployment wizard?
If you really want to "unit" test, you should test the behavior of your class B AND the behavior of your class A, mocking the call to the class B method.
That's a tough conversation between mock lovers and others (Martin Fowler I think is not a "mocker").
Anyway. You should stop thinking about 100% coverage. You should think about:
Why am i testing?
How am i testing?
Here, i'd definitely go for 2 tests:
One test for the B class into the b class test file to be sure the B method is well implemented, with all the side effects, side values etc.
one test for the A class mocking the class B
What is a mock?
To stay VERY simple: A mock is a portion of code in your test which is gonna say: when the B class method is called, always return this value: "+++" .
By doing this, you allow yourself having a maintanable and modulable test suite.
In java, I love mockito : http://mockito.org/
Although one of my colleagues is lead maintainer for easymock: http://easymock.org/
Hope this helps. Ask me if you need further help.
EDIT SOME EXAMPLE
With Java and mockito:
public class aUTest {
protected A a;
#Mock protected B b;
#Before
public void setUp(){
MockitoAnnotations.initMocks(this);
a = new A();
ReflectionTestUtils.setField(a, "b", b);
}
#Test
public void test_A_method_should_not_throw_exception()
when(b. execute()).thenReturn(true); //just an example of a return value from b. execute()
Boolean result = a.testHello();
// Assert
Assert.assertEquals(true, result);
}
I created an Apex class called TestHelper for all my mock objects. I use constants (static final) for values that I might need elsewhere and public static fields for objects. Works great and since no methods are used, no test coverage is needed.
public without sharing class TestHelper {
public static final string testPRODUCTNAME = 'test Product Name';
public static final string testCOMPANYID = '2508';
public static Account testAccount {
get{
Account tAccount = new Account(
Name = 'Test Account',
BillingStreet = '123 Main St',
BillingCity = 'Dallas',
BillingState = 'TX',
BillingPostalCode = '75234',
Website = 'http://www.google.com',
Phone = '222 345 4567',
Subscription_Start_Date__c = system.today(),
Subscription_End_Date__c = system.today().addDays(30),
Number_Of_Seats__c = 1,
companyId__c = testCOMPANYID,
ZProduct_Name__c = testPRODUCTNAME);
insert tAccount;
return tAccount;
}
}
}
Here is my situation:
I want to test on the "HasSomething()" function, which is in the following class:
public class Something
{
private object _thing;
public virtual bool HasSomething()
{
if (HasSomething(_thing))
return true;
return false;
}
public virtual bool HasSomething(object thing)
{
....some algo here to check on the object...
return true;
}
}
So, i write my test to be like this:
public void HasSomethingTest1()
{
MockRepository mocks = new MockRepository();
Something target = mocks.DynamicMock(typeof(Something)) as Something;
Expect.Call(target.HasSomething(new Object())).IgnoreArguments().Return(true);
bool expected = true;
bool actual;
actual = target.HasSomething();
Assert.AreEqual(expected, actual);
}
Is my test written correctly?
Please help me as i can't even get the result as expected. the "HasSomething(object)" just can't be mock in that way. it did not return me 'true' as being set in expectation.
Thanks.
In response to OP's 'answer': Your main problem is that RhinoMocks does not mock members of classes - instead it creates mock classes and we can then set expectations and canned responses for its members (i.e. Properties and Functions). If you attempt to test a member function of a mock/stub class, you run the risk of testing the mocking framework rather than your implementation.
For the particular scenario of the logical path being dependent on the return value of a local (usually private) function, you really need an external dependency (another object) which would affect the return value that you require from that local function. For your code snippet above, I would write the test as follows:
[Test]
public void TestHasSomething()
{
// here I am assuming that _thing is being injected in via the constructor
// you could also do it via a property setter or a function
var sut = new Something(new object());
Assert.IsTrue(sut.HasSomething);
}
i.e. no mocking required.
This is one point of misunderstanding that I often had in the past with regards to mocking; we mock the behaviour of a dependency of the system under test (SUT). Something like: the SUT calls several methods of the dependency and the mocking process provides canned responses (rather than going to the database, etc) to guide the way the logic flows.
A simple example would be as follows (note that I have used RhinoMocks AAA syntax for this test. As an aside, I notice that the syntax that you are using in your code sample is using the Record-Replay paradigm, except that it isn't using Record and Replay! That would probably cause problems as well):
public class SUT
{
Dependency _depend
public SUT (Dependency depend)
{
_depend = depend;
}
...
public int MethodUnderTest()
{
if (_depend.IsReady)
return 1;
else
return -1;
}
}
...
[Test]
public void TestSUT_MethodUnderTest()
{
var dependency = MockRepository.GenerateMock<Dependency>();
dependency.Stub(d => d.IsReady).Return(true);
var sut = new SUT(dependency);
Assert.AreEqual(1, sut.MethodUnderTest());
}
And so the problem that you have is that you are attempting to test the behaviour of a mocked object. Which means that you aren't actually testing your class at all!
In a case like this, your test double should be a derived version of class Something. Then you override the method HasSomething(object) and ensure that HasSomething() calls your one.
If I understand correctly, you are actually interested in testing the method HasDynamicFlow (not depicted in your example above) without concerning yourself with the algorithm for HasSomething.
Preet is right in that you could simply subclass Something and override the behavior of HasSomething to short-circuit the algorithm, but that would require creating some additional test-dummy code which Rhino is efficient at eliminating.
Consider using a Partial Mock Stub instead of a Dynamic Mock. A stub is less strict and is ideal for working with Properties. Methods however require some extra effort.
[Test]
public void CanStubMethod()
{
Foo foo = MockRepository.GenerateStub<Foo>();
foo.Expect(f => f.HasDynamicFlow()).CallOriginalMethod(OriginalCallOptions.NoExpectation);
foo.Expect(f => f.HasSomething()).CallOriginalMethod(OriginalCallOptions.NoExpectation);
foo.Expect(f => f.HasSomething(null)).IgnoreArguments().Return(true);
Assert.IsTrue(foo.HasDynamicFlow());
}
EDIT: added code example and switched Partial Mock to Stub