Category Archives: Unit Testing

Unit Testing With Entity Framework and Entity Framework Core 2.1

Entity Framework Core 2.1 comes with a nifty little feature: an In Memory Database setting. What this means, is that with a single option setting, your tests can interact directly with the database (or at least EF’s impression of the database) but not actually touch any physical database. In other words, you can write unit tests for data access; an example:

// Arrange
DbContextOptions<ApplicationDbContext> options = new DbContextOptionsBuilder<ApplicationDbContext>()
    .UseInMemoryDatabase(Guid.NewGuid().ToString())
    .EnableSensitiveDataLogging()                
    .Options;

using (var context = new ApplicationDbContext(options))
{
    context.Database.EnsureDeleted();
	ResourceCategory resourceCategory = new ResourceCategory()
    {
        Name = "TestCategory"
    }
};
 
// Act
_applicationDbContext.ResourceCategories.Add(resourceCategory);
_applicationDbContext.SaveChanges();
	
 
// Assert                
Assert.Equal("TestCategory", context.ResourceCategories.First().Name);               

To just quickly explain what this is doing: we have a DbContext called ApplicationDbContext and we’re building a set of options on top of that context. We’re then instantiating the context and cleaning the in memory database. Finally, we’re adding a new piece of data to the context and then asserting that it has been added.

Told you it was nifty.

But what about if you’re still using Entity Framework 6?

Glad you asked.

Out of the box, EF does not come with this kind of functionality; however, I recently came across (and contributed) to a NuGet library that provides just such a facility. It provides a wrapper for both Moq and Nsubstitute. The GitHub Repo is here.

Short Walks – NSubstitute extension methods like .Received() can only be called on objects created using Substitute.For() and related methods

Not a huge post, but this has tripped me up several times, and sent me down some quite deep rabbit holes.

On to the story…

I once had a check to received calls like this:

var myData = Substitute.For<IData>();

. . .

myData
    .Add(Arg.Is<MyEntity>(a =>
        a.Active == true
        && a.Field1 == 1
        && a.Field2 == 42))
    .Received(1);

And I was getting the error:

NSubstitute extension methods like .Received() can only be called on objects created using Substitute.For() and related methods

I was confident that I had declared it correctly, a few lines above… I could, in fact see that I had declared it correctly. As my brain slowly dribbled from my nose, I did a quick search on the web; and found a solitary article that suggested I might have confused the Received and method calls; the correct syntax being:

myData
    .Received(1)
    .Add(Arg.Is<MyEntity>(a =>
        a.Active == true
        && a.Field1 == 1
        && a.Field2 == 42));

Hopefully, now that I’ve doubled the available resources available to people finding this error, it will plague me less in compensation.

Short Walks – XUnit Warning

As with many of these posts – this is more of a “note to self”.

Say you have an assertion that looks something like this in your Xunit test:

Assert.True(myEnumerable.Any(a => a.MyValue == "1234"));

In later versions (not sure exactly which one this was introduced it), you’ll get the following warning:

warning xUnit2012: Do not use Enumerable.Any() to check if a value exists in a collection.

So, Xunit has a nice little feature where you can use the following syntax instead:

Assert.Contains(myEnumerable, a => a.MyValue == "1234");

Using NSubstitute for partial mocks

I have previously written about how to, effectively, subclass using Nsubstitute; in this post, I’ll cover how to partially mock out that class.

Before I get into the solution; what follows is a workaround to allow badly written, or legacy code to be tested without refactoring. If you’re reading this and thinking you need this solution then my suggestion would be to refactor and use some form of dependency injection. However, for various reasons, that’s not always possible (hence this post).

Here’s our class to test:

public class MyFunkyClass
{
    public virtual void MethodOne()
    {        
        throw new Exception("I do some direct DB access");
    }
 
    public virtual int MethodTwo()
    {
        throw new Exception("I do some direct DB access and return a number");

        return new Random().Next(5);
    }
 
    public virtual int MethodThree()
    {
        MethodOne();
        if (MethodTwo() <= 3)
        {
            return 1;
        }
 
        return 2;
    }
}

The problem

Okay, so let’s write our first test:

[Fact]
public void Test1()
{
    // Arrange
    MyFunkyClass myFunkyClass = new MyFunkyClass();
 
    // Act
    int result = myFunkyClass.MethodThree();
 
    // Assert
    Assert.Equal(2, result);
}

So, what’s wrong with that?

Well, we have some (simulated) DB access, so the code will error.

Not the but a solution

The first thing to do here is to mock out MethodOne(), as it has (pseudo) DB access:

[Fact]
public void Test1()
{
    // Arrange
    MyFunkyClass myFunkyClass = Substitute.ForPartsOf<MyFunkyClass>();
    myFunkyClass.When(a => a.MethodOne()).DoNotCallBase();
 
    // Act
    int result = myFunkyClass.MethodThree();
 
    // Assert
    Assert.Equal(2, result);
}

Running this test now will fail with:

Message: System.Exception : I do some direct DB access and return a number

We’re past the first hurdle. We can presumably do the same thing for MethodTwo:

[Fact]
public void Test1()
{
    // Arrange
    MyFunkyClass myFunkyClass = Substitute.ForPartsOf<MyFunkyClass>();
    myFunkyClass.When(a => a.MethodOne()).DoNotCallBase();
    myFunkyClass.When(a => a.MethodTwo()).DoNotCallBase();
 
    // Act
    int result = myFunkyClass.MethodThree();
 
    // Assert
    Assert.Equal(2, result);
}

Now when we run the code, the test still fails, but it no longer accesses the DB:

Message: Assert.Equal() Failure
Expected: 2
Actual: 1

The problem here is that, even though we don’t want MethodTwo to execute, we do want it to return a predefined result. Once we’ve told it not to call the base method, you can then tell it to return whatever we choose (there are separate events – see the bottom of this post for a more detailed explanation of why); for example:

[Fact]
public void Test1()
{
    // Arrange
    MyFunkyClass myFunkyClass = Substitute.ForPartsOf<MyFunkyClass>();
    myFunkyClass.When(a => a.MethodOne()).DoNotCallBase();
    myFunkyClass.When(a => a.MethodTwo()).DoNotCallBase();
    myFunkyClass.MethodTwo().Returns(5);
 
    // Act
    int result = myFunkyClass.MethodThree();
 
    // Assert
    Assert.Equal(2, result);
}

And now the test passes.

TLDR – What is this actually doing?

To understand this better; we could do this entire process manually. Only when you’ve felt the pain of a manual mock, can you really see what mocking frameworks such as NSubtitute are doing for us.

Let’s assume that we don’t have a mocking framework at all, but that we still want to test MethodThree() above. One approach that we could take is to subclass MyFunkyClass, and then test that subclass:

Here’s what that might look like:

class MyFunkyClassTest : MyFunkyClass
{
    public override void MethodOne()
    {
        //base.MethodOne();
    }
 
    public override int MethodTwo()
    {
        //return base.MethodTwo();
        return 5;
    }
}

As you can see, now that we’ve subclassed MyFunkyClass, we can override the behaviour of the relevant virtual methods.

In the case of MethodOne, we’ve effectively issued a DoNotCallBase(), (by not calling base!).

For MethodTwo, we’ve issued a DoNotCallBase, and then a Returns statement.

Let’s add a new test to use this new, manual method:

[Fact]
public void Test2()
{
    // Arrange 
    MyFunkyClassTest myFunkyClassTest = new MyFunkyClassTest();
 
    // Act
    int result = myFunkyClassTest.MethodThree();
 
    // Assert
    Assert.Equal(2, result);
}

That’s much cleaner – why not always use manual mocks?

It is much cleaner if you always want MethodThree to return 5. Once you need it to return 2 then you have two choices, either you create a new mock class, or you start putting logic into your mock. The latter, if done wrongly can end up with code that is unreadable and difficult to maintain; and if done correctly will end up in a mini version of NSubstitute.

Finally, however well you write the mocks, as soon as you have more than one for a single class then every change to the class (for example, changing a method’s parameters or return type) results in a change to more than one test class.

It’s also worth mentioning again that this problem is one that has already been solved, cleanly, by dependency injection.

Using Entity Framework with IoC

One thing to bear in mind about using entity framework is that the DbContext object is not thread safe. This threw me when I first discovered it; it also confused why I was getting this error, as I was running in an environment that I thought was pretty much single threaded (in fact, it was an Azure function). I was using IoC and so the DbContext is shared across instances.

Because the error is based on, effectively, a race condition, your mileage may vary, but you’ll typically get one of the following errors:

System.InvalidOperationException: ‘The context cannot be used while the model is being created. This exception may be thrown if the context is used inside the OnModelCreating method or if the same context instance is accessed by multiple threads concurrently. Note that instance members of DbContext and related classes are not guaranteed to be thread safe.’

System.InvalidOperationException: ‘A second operation started on this context before a previous operation completed. Any instance members are not guaranteed to be thread safe.’

I had a good idea what the cause of this might be, but in order to investigate, I set-up a console app accessing Entity Framework; in this case, EF Core.

If you’re here to find a solution, then you can probably scroll down to the section labelled To Fix.

Reproducing the error

To reproduce the error, we need to introduce Unity (I imagine the same would be true of any IoC provider, as the problem is more with the concept than the implementation):

Install-Package Unity

The next step is to abstract away the data access layer, in order to provide a base for our dependency injection:

As you can see, we’re introducing a data access layer – and we’re creating an interface for our DbContext. The idea being that this can subsequently be resolved by Unity. Here are our interfaces:

public interface IDataAccess
{
    List<string> GetData();
 
    void AddData(List<string> newData);
}

public interface IMyDbContext
{
    DbSet<MyData> MyData { get; set; }
 
    void SaveChanges();
}

public interface IDoStuff
{
    void DoStuffQuickly();
}

Finally, we can implement the interfaces:

public class DataAccess : IDataAccess
{
    private readonly IMyDbContext _myDbContext;
 
    public DataAccess(IMyDbContext myDbContext)
    {
        _myDbContext = myDbContext;
    }
    public void AddData(List<string> newData)
    {
        foreach (var data in newData)
        {
            MyData myData = new MyData()
            {
                FieldOne = data
            };
            _myDbContext.MyData.Add(myData);
        }
        _myDbContext.SaveChanges();
    }
 
    public List<string> GetData()
    {
        List<string> data = 
            _myDbContext.MyData.Select(a => a.FieldOne).ToList();
 
        return data;
    }
}

The DbContext:

public class MyDbContext : DbContext, IMyDbContext
{
    public DbSet<MyData> MyData { get; set; }
 
    protected override void OnConfiguring(DbContextOptionsBuilder optionsBuilder)
    {
        string cn = @"Server=.\SQLEXPRESS;Database=test-db . . .";
        optionsBuilder.UseSqlServer(cn);
 
        base.OnConfiguring(optionsBuilder);
    }
 
    void IMyDbContext.SaveChanges()
    {
        base.SaveChanges();
    }
}

Let’s register a class to actually start the process:

public class DoStuff : IDoStuff
{
    private readonly IDataAccess _dataAccess;
    private readonly ILogger _logger;
 
    public DoStuff(IDataAccess dataAccess, ILogger logger)
    {
        _dataAccess = dataAccess;
        _logger = logger;
    }
 
    public void DoStuffQuickly()
    {
        _dataAccess.AddData(new List<string>()
        {
            "testing new data",
            "second test",
            "last test"
        });
 
        Parallel.For(1, 100, (a) =>
        {
            List<string> data = _dataAccess.GetData();
            foreach(string text in data)
            {
                _logger.Log(text);
            }
        });
    }
}

Now let’s register the types in out IoC container:

static void Main(string[] args)
{
    Console.WriteLine("Hello World!");
 
    UnityContainer container = new UnityContainer();
    container.RegisterType<IDataAccess, DataAccess.DataAccess>();
    container.RegisterType<IMyDbContext, MyDbContext>();
    container.RegisterType<IDoStuff, DoStuff>();
    container.RegisterType<ILogger, ConsoleLogger>();
 
    container.Resolve<IDoStuff>().DoStuffQuickly();
}

(Console logger is just a class that calls Console.WriteLine()).

Now, when you run it, you’ll get the same error as above (or something to the same effect).

To fix

One possible fix is to simply instantiate a DbContext as you need it; for example:

public List<string> GetData()
{
    using (var myDbContext = new MyDbContext())
    {
        List<string> data =
            myDbContext.MyData.Select(a => a.FieldOne).ToList();
        return data;
    }
    
}

However, the glaring problem that this creates is unit testing. The solution is a simple factory class:

public interface IGenerateDbContext
{
    IMyDbContext GenerateNewContext();
}

public class GenerateDbContext : IGenerateDbContext
{
    public IMyDbContext GenerateNewContext()
    {
        IMyDbContext myDbContext = new MyDbContext();
 
        return myDbContext;
    }
}

We’ll need to make the DbContext implementation disposable:

public interface IMyDbContext : IDisposable
{
    DbSet<MyData> MyData { get; set; }
 
    void SaveChanges();
}

And, finally, we can change the data access code:

public void AddData(List<string> newData)
{
    using (IMyDbContext myDbContext = _generateDbContext.GenerateNewContext())
    {
        foreach (string data in newData)
        {
            MyData myData = new MyData()
            {
                FieldOne = data
            };
            myDbContext.MyData.Add(myData);
        }
        myDbContext.SaveChanges();
    }
}
 
public List<string> GetData()
{
    using (IMyDbContext myDbContext = _generateDbContext.GenerateNewContext())
    {
        List<string> data =
            myDbContext.MyData.Select(a => a.FieldOne).ToList();
        return data;
    }
    
}

Now we can use IoC to generate the DBContext, and therefore it is testable, but we don’t pass the DbContext itself around, and therefore there are no threading issues.

Short Walks – XUnit Tests Not Appearing in Test Explorer

On occasion, there may be a case where you go into Test Explorer, knowing that you have XUnit tests within the solution; the Xunit tests are in a public class, they are public, and they are decorated correctly (for example, [Fact]). However, they do not appear in the Text Explorer.

If you have MS Test tests, you may find that they do appear in the Test Explorer – only the XUnit tests do not.

Why?

To run Xunit tests from the command line, you’ll need this package.

To run Xunit tests from Visual Studio, you’ll need this package.

References

https://xunit.github.io/docs/nuget-packages.html

Short Walks – NSubstitute – Subclassing and Partial Substitutions

I’ve had this issue a few times recently. Each time I have it, after I’ve worked out what it was, it makes sense, but I keep running into it. The resulting frustration is this post – that way, it’ll come up the next time I search for it on t’internet.

The Error

The error is as follows:

“NSubstitute.Exceptions.CouldNotSetReturnDueToNoLastCallException: ‘Could not find a call to return from.”

Typically, it seems to occur in one of two circumstances: substituting a concrete class and partially substituting a concrete class; that is:

var myMock = Substitute.ForPartsOf<MyClass>();

Or:

var myMock = Substitute.For<MyClass>();

Why?

If you were to manually mock out an interface, how would you do that? Well, say you had IMyClass, you’d just do something like this:

public class MyClassMock : IMyClass 
{
	// New and imaginative behaviour goes here
}

All’s good – you get a brand new implementation of the interface, and you can do anything with it. But what would you do if you were trying to unit test a method inside MyClass that called another method inside MyClass; for example:

public class MyClass : IMyClass
{
	public bool Method1()
{
		int rowCount = ReadNumberOfRowsInFileOnDisk();
		Return rowCount > 10;
	}
	
	public int ReadNumberOfRowsInFileOnDisk()
	{
		// Opens a file, reads it, and returns the number of rows
	}
}

(Let’s not get bogged down in how realistic this scenario is, or whether or not this is good practice – it illustrates a point)

If you want to unit test Method1(), but don’t want to actually read a file from the disk, you’ll want to replace ReadNumberOfRowsInFileOnDisk(). The only real way that you can do this is to subclass the class; for example:

public class MyClassMock : MyClass

You can now test the behaviour on MyClass, via MyClassMock; however, you still can’t* override the method ReadNumberOfRowsInFileOnDisk() because it isn’t virtual. If you make the method virtual, you can override it in the subclass.

The same is true with NSubstitute – if you want to partially mock a class in this way, it follows the same rules as you must if you would to roll your own.

Footnotes

* There may, or may not be one or two ways to get around this restriction, but let’s at least agree that they are, at best, unpleasant.

Live Testing in Visual Studio 2017

While working on this project, I decided that I would see how useful the new feature of “Live Testing” is for TDD.

Once you’ve created your suite of tests, turning live testing on gives you some interesting information:

As you can see, there are some slight contradictions here in what it’s telling me. The first test is failing, the test for Count has a line (which indicates not covered). However, code lens is telling me there are two covering tests.

My suspicion here is that live testing is being too clever for its own good; for example, looking at the tests themselves, shows that only the first Enqueue statement is covered:

My thoughts are that the engine works out where the test fails, and so, in the case above, it’s telling me that the tests are not running the line of code that tests the Count functionality. What I’m guessing is that this is because the tests are failing before that.

I can prove this by creating a further unit test:

And, indeed, this now shows coverage:

As I start to add the functionality, I can quickly see that the tests are now passing:

Changing the way code is written

The interesting thing here is that is promotes two things:
1. Defined tests – if a test covers more than one thing, it will show as uncovered initially
2. Tests first – obviously, there need to be tests to show test coverage

Caveats

1. Annoyingly, by default, you continually have to opt in for live unit testing. It only starts when you select to “Start” live unit testing, and this must be done each time you launch VS. This is configurable, and can be switched on by default in options:

2. If the build fails, it shows nothing.

3. Live testing is only available in Visual Studio Enterprise.

NUnit

One thing that only occurred to me as notable while I was writing this is that the tests that I’ve written are in NUnit. So the Live Testing clearly works with none MSTests, and probably works with anything that’ll show up in Test Explorer / Code Lens.

References

https://blogs.msdn.microsoft.com/visualstudio/2017/03/09/live-unit-testing-in-visual-studio-2017-enterprise/

NUnit TestCaseSource

While working on this project, I found a need to abstract away a base type that the unit tests use (in this instance, it was a queue type). I was only testing a single type (PriorityQueue); however, I wanted to create a new type, but all the basic tests for the new type are the same as the existing ones. This led me to investigate the TestCaseSource attribute in NUnit.

As a result, I needed a way to re-use the tests. There are definitely multiple ways to do this; the simplest one is probably to create a factory class, and pass in a string parameter. The only thing that put me off this is that you end up with the following test case:

        [TestCase("test", "test9", "test", "test2", "test3", "test4", "test5", "test6", "test7", "test8", "test9"]
        [TestCase("a1", "a", "a1", "b", "c", "d", "a"]
        public void Queue_Dequeue_CheckResultOrdering(
            string first, string last, params string[] queueItems)
        {

Becoming:

        [TestCase("PriorityQueue", "test", "test9", "test", "test2", "test3", "test4", "test5", "test6", "test7", "test8", "test9"]
        [TestCase("PriorityQueue2", "test", "test9", "test", "test2", "test3", "test4", "test5", "test6", "test7", "test8", "test9"]
        [TestCase("PriorityQueue", "a1", "a", "a1", "b", "c", "d", "a"]
        [TestCase("PriorityQueue2", "a1", "a", "a1", "b", "c", "d", "a"]
        public void Queue_Dequeue_CheckResultOrdering(
            string queueType, string first, string last, params string[] queueItems)
        {

This isn’t very scaleable when adding a third or fourth type.

TestCaseSource

It turns out that the (or an least an) answer to this is to use NUnit’s TestCaseSource attribute. The NUnit code base dog foods quite extensively, so that is not a bad place to look for examples of how this works; however, what I couldn’t find was a way to mix and match. To better illustrate the point; here’s the first test that I changed to use TestCaseSource:

        [Test]
        public void Queue_NoEntries_CheckCount()
        {
            // Arrange
            PQueue.PriorityQueue<string> queue = new PQueue.PriorityQueue<string>();

            // Act
            int count = queue.Count();

            // Assert
            Assert.AreEqual(0, count);
        }

Which became:

        [Test, TestCaseSource(typeof(TestableQueueItemFactory), "ReturnQueueTypes")]
        public void Queue_NoEntries_CheckCount(IQueue<string> queue)
        {
            // Arrange


            // Act
            int count = queue.Count();

            // Assert
            Assert.AreEqual(0, count);
        }

(For completeness, the TestableQueueItemFactory is here):

    public static class TestableQueueItemFactory
    {
        public static IEnumerable<IQueue<string>> ReturnQueueTypes()
        {
            yield return new PQueue.PriorityQueue<string>();
        }
    }

However, when you have a TestCase like the one above, there’s a need for the equivalent of this (which doesn’t work):

        [Test, TestCaseSource(typeof(TestableQueueItemFactory), "ReturnQueueTypes")]
        [TestCase("test", "test9", "test", "test2", "test3", "test4", "test5", "test6", "test7", "test8", "test9")]
        [TestCase("a1", "a", "a1", "b", "c", "d", "a")]
        public void Queue_Dequeue_CheckResultOrdering(string first, string last, params string[] queueItems)
        {

A quick look at the NUnit code base reveals these attributes to be mutually exclusive.

Compromise

By no means is this a perfect solution, but the one that I settled on was to create a second TestCaseSource helper method, which looks like this (along with the test):

        private static IEnumerable Queue_Dequeue_CheckResultOrdering_TestCase()
        {
            foreach(var queueType in TestableQueueItemFactory.ReturnQueueTypes())
            {
                yield return new object[] { queueType, "test", "test9", new string[] { "test", "test2", "test3", "test4", "test5", "test6", "test7", "test8", "test9" } };
                yield return new object[] { queueType, "a1", "a", new string[] { "a1", "b", "c", "d", "a" } };
            }
        }

        [Test, TestCaseSource("Queue_Dequeue_CheckResultOrdering_TestCase")]
        public void Queue_Dequeue_CheckResultOrdering(
            IQueue <string> queue, string first, string last, params string[] queueItems)
        {

As you can see, the second helper method doesn’t really help readability, so it’s certainly not a perfect solution; in fact, with a single queue type, this makes the code more complex and less readable. However, When a second and third queue type are introduced, the test suddenly becomes resilient.

YAGNI

At first glance, this may appear to be an example of YAGNI. However, in this article, Martin Fowler does state:

Yagni only applies to capabilities built into the software to support a presumptive feature, it does not apply to effort to make the software easier to modify.

Which, I believe, is what we are doing here.

References

http://www.smaclellan.com/posts/parameterized-tests-made-simple/

http://stackoverflow.com/questions/16346903/how-to-use-multiple-testcasesource-attributes-for-an-n-unit-test

https://github.com/nunit/docs/wiki/TestCaseSource-Attribute

http://dotnetgeek.tumblr.com/post/2851360238/exploiting-nunit-attributes-valuesourceattribute

https://github.com/nunit/docs/wiki/TestCaseSource-Attribute

Testing for Exceptions using the Arrange Act Assert Pattern in C# 7

Unit testing massively benefits from following the Arrange / Act / Assert pattern. I’ve seen tests that are not written in this way, and they can be sprawling and indecipherable, either testing many different things in series, or testing nothing at all except the .Net Framework.

I recently found an issue while trying to test for an exception being thrown, which is that Nunit (and probably other frameworks) test for an exception by accepting a delegate to test. Here’s an example:

        [Test]
        public void Test_ThrowException_ExceptionThrown()
        {
            // Arrange
            TestClass tc = new TestClass();

            // Act / Assert
            Assert.Throws(typeof(Exception), tc.ThrowException);
        }

We’re just testing a dummy class:

    public class TestClass
    {
        public void ThrowException()
        {
            throw new Exception("MyException");
        }
    }

C# 7 – Inline functions

If you look in the references at the bottom, you’ll see something more akin to this approach:

        public void Test_ThrowException_ExceptionThrown2()
        {
            // Arrange
            TestClass tc = new TestClass();

            // Act
            TestDelegate throwException = () => tc.ThrowException();            

            // Assert
            Assert.Throws(typeof(Exception), throwException);
        }

However, since C# 7, the option on a local function has arrived. The following has the same effect:

        [Test]
        public void Test_ThrowException_ExceptionThrown3()
        {
            // Arrange
            TestClass tc = new TestClass();

            // Act
            void CallThrowException()
            {
                tc.ThrowException();
            }

            // Assert
            Assert.Throws(typeof(Exception), CallThrowException);
        }

I think that I, personally, still prefer the anonymous function for this; however, the local function does present some options; for example:

        [Test]
        public void Test_ThrowException_ExceptionThrown4()
        {
            void CallThrowException()
            {
                // Arrange
                TestClass tc = new TestClass();

                // Act
                tc.ThrowException();
            }

            // Assert
            Assert.Throws(typeof(Exception), CallThrowException);
        }

Now I’m not so sure that I still prefer the anonymous function.

References

http://stackoverflow.com/questions/33897323/nunit-3-0-and-assert-throws

https://pmbanugo.wordpress.com/2014/06/16/exception-testing-pattern/

http://stackoverflow.com/questions/24070115/best-approach-towards-applying-the-arrange-act-assert-pattern-when-expecting-exc