I have an object which has a public void method on it that just modifies internal state. Part of me feels as though it should have tests as it is a public method but how do you go about testing this? Do you bother or do you rely on the other tests which make use of the fact that the internal state has changed?
To add an example I am implementing a gap buffer which has an insert method on it and a moveCursorForward method which just modifies the internal state.
Thanks,
Aly
I think injection is your answer. Create a Cursor object and inject it into your Buffer. Unit test the cursor object so that you know it works perfectly ;). Since cursor is stateful and you've passed it into your buffer, you can assert on the cursor object when you test moveCursorForward.
Check out Michael Feather's Working Effectively with Legacy Code
This also is where mocking objects might be beneficial.
I prefer separate tests for such methods. For example, there are two case for "moveCursorForward" method:
1. cursor is already in the end of the buffer
2. cursor is not in the end of the buffer
So it is likely that case 1 will be skipped if you don't create special test for it.
In other words you can miss some boundary cases.
Related
I'm looking for guidelines to using a consistent value of the current date and time throughout a transaction.
By transaction I loosely mean an application service method, such methods usually execute a single SQL transaction, at least in my applications.
Ambient Context
One approach described in answers to this question is to put the current date in an ambient context, e.g. DateTimeProvider, and use that instead of DateTime.UtcNow everywhere.
However the purpose of this approach is only to make the design unit-testable, whereas I also want to prevent errors caused by unnecessary multiple querying into DateTime.UtcNow, an example of which is this:
// In an entity constructor:
this.CreatedAt = DateTime.UtcNow;
this.ModifiedAt = DateTime.UtcNow;
This code creates an entity with slightly differing created and modified dates, whereas one expects these properties to be equal right after the entity was created.
Also, an ambient context is difficult to implement correctly in a web application, so I've come up with an alternative approach:
Method Injection + DeterministicTimeProvider
The DeterministicTimeProvider class is registered as an "instance per lifetime scope" AKA "instance per HTTP request in a web app" dependency.
It is constructor-injected to an application service and passed into constructors and methods of entities.
The IDateTimeProvider.UtcNow method is used instead of the usual DateTime.UtcNow / DateTimeOffset.UtcNow everywhere to get the current date and time.
Here is the implementation:
/// <summary>
/// Provides the current date and time.
/// The provided value is fixed when it is requested for the first time.
/// </summary>
public class DeterministicTimeProvider: IDateTimeProvider
{
private readonly Lazy<DateTimeOffset> _lazyUtcNow =
new Lazy<DateTimeOffset>(() => DateTimeOffset.UtcNow);
/// <summary>
/// Gets the current date and time in the UTC time zone.
/// </summary>
public DateTimeOffset UtcNow => _lazyUtcNow.Value;
}
Is this a good approach? What are the disadvantages? Are there better alternatives?
Sorry for the logical fallacy of appeal to authority here, but this is rather interesting:
John Carmack once said:
There are four principle inputs to a game: keystrokes, mouse moves, network packets, and time. (If you don't consider time an input value, think about it until you do -- it is an important concept)"
Source: John Carmack's .plan posts from 1998 (scribd)
(I have always found this quote highly amusing, because the suggestion that if something does not seem right to you, you should think of it really hard until it seems right, is something that only a major geek would say.)
So, here is an idea: consider time as an input. It is probably not included in the xml that makes up the web service request, (you wouldn't want it to anyway,) but in the handler where you convert the xml to an actual request object, obtain the current time and make it part of your request object.
So, as the request object is being passed around your system during the course of processing the transaction, the time to be considered as "the current time" can always be found within the request. So, it is not "the current time" anymore, it is the request time. (The fact that it will be one and the same, or very close to one and the same, is completely irrelevant.)
This way, testing also becomes even easier: you don't have to mock the time provider interface, the time is always in the input parameters.
Also, this way, other fun things become possible, for example servicing requests to be applied retroactively, at a moment in time which is completely unrelated to the actual current moment in time. Think of the possibilities. (Picture of bob squarepants-with-a-rainbow goes here.)
Hmmm.. this feels like a better question for CodeReview.SE than for StackOverflow, but sure - I'll bite.
Is this a good approach?
If used correctly, in the scenario you described, this approach is reasonable. It achieves the two stated goals:
Making your code more testable. This is a common pattern I call "Mock the Clock", and is found in many well-designed apps.
Locking the time to a single value. This is less common, but your code does achieve that goal.
What are the disadvantages?
Since you are creating another new object for each request, it will create a mild amount of additional memory usage and additional work for the garbage collector. This is somewhat of a moot point since this is usually how it goes for all objects with per-request lifetime, including the controllers.
There is a tiny fraction of time being added before you take the reading from the clock, caused by the additional work being done in loading the object and from doing lazy loading. It's negligible though - probably on the order of a few milliseconds.
Since the value is locked down, there's always the risk that you (or another developer who uses your code) might introduce a subtle bug by forgetting that the value won't change until the next request. You might consider a different naming convention. For example, instead of "now", call it "requestRecievedTime" or something like that.
Similar to the previous item, there's also the risk that your provider might be loaded with the wrong lifecycle. You might use it in a new project and forget to set the instancing, loading it up as a singleton. Then the values are locked down for all requests. There's not much you can do to enforce this, so be sure to comment it well. The <summary> tag is a good place.
You may find you need the current time in a scenario where constructor injection isn't possible - such as a static method. You'll either have to refactor to use instance methods, or will have to pass either the time or the time-provider as a parameter into the static method.
Are there better alternatives?
Yes, see Mike's answer.
You might also consider Noda Time, which has a similar concept built in, via the IClock interface, and the SystemClock and FakeClock implementations. However, both of those implementations are designed to be singletons. They help with testing, but they don't achieve your second goal of locking the time down to a single value per request. You could always write an implementation that does that though.
Code looks reasonable.
Drawback - most likely lifetime of the object will be controlled by DI container and hence user of the provider can't be sure that it always be configured correctly (per-invocation and not any longer lifetime like app/singleton).
If you have type representing "transaction" it may be better to put "Started" time there instead.
This isn't something that can be answered with a realtime clock and a query, or by testing. The developer may have figured out some obscure way of reaching the underlying library call...
So don't do that. Dependency injection also won't save you here; the issue is that you want a standard pattern for time at the start of the 'session.'
In my view, the fundamental problem is that you are expressing an idea, and looking for a mechanism for that. The right mechanism is to name it, and say what you mean in the name, and then set it only once. readonly is a good way to handle setting this only once in the constructor, and lets the compiler and runtime enforce what you mean which is that it is set only once.
// In an entity constructor:
this.CreatedAt = DateTime.UtcNow;
Assume we have an Order class with a method called Approve. When this method is called, it checks certain conditions and either puts the Order in the state of Approved or throws an exception. In the service layer, we've got something like this:
var order = _repository.Single(o => o.ID == orderID);
order.Approve();
_context.SaveChanges(); // or _session.SaveChanges();
There are 2 ways to test this method and I'd like to hear your insight on this:
Solution 1: Stub the repository to return an Order object. Then assert the Order is in the state of "Approved".
Solution 2: Stub the repository to return a Mock Order object. Assert that Approve() method was called.
Solution 1 is easier and I personally favor state-based testing to interaction-based testing, as the latter can target implementation details and should be avoided. However, I believe testing that the given Order is in the state of Approved is not the concern of this service method. I think we need a separate test method for the Order class to test whether an exception is thrown or the Order's state is changed to Approved.
Solution 2 may sound logical as we are delegating the responsibility of Approving an Order to the Order class itself. So perhaps we need 2 tests for this service method: One to ensure it delegates the task of Approving an Order to the Order class and one to ensure it saves the changes.
What's your insight on this? Which solution do you prefer?
Cheers
Unit tests are to test whether the observed behavior is conforming to expectations/specification.
The answer to your question boils down to what you consider "expected behavior" in this case: a) if the expected behavior is that the Order is in approved state after calling the service method, then test the state; b) if the expected behavior is that the approve action is delegated, then test the method call.
You will need to test the Order object's behavior as well (so that calling Approve() changes the state to approved) in either case.
The second solution plays well as it decouples the behavior of the two objects, but if there are more than one ways that the order can be in approved state (and that's what you are testing -- case a)), then you limit the accepted behavior needlessly.
Also, I would create a separate test for testing the saving part, if that is not essential to the approval part
I am just starting to learn unit testing and I think it's a really good tool which I want to start using for all my projects, however I'm still not sure how to test some things.
As an example I am implementing a queue and one of the methods it has is AddNode, I also have a mock object representing the Node in the queue, after writing the method I don't know what I should test for. The method is void so I can't test for a return value, maybe I should test for out of mamory exception or some other exception ? or maybe there is no need to do any testing in this case.
For AddNode you could be testing the following:
queue is not empty afterwards
size has increased by one
if the queue checks for duplication, that the size has not increased after adding a duplicate
a roundtrip: new Queue, addNode, getNode returns the same Node again
if the queue checks for invalid objects (no nulls, wrong node type or something), that there is an exception when you try to add these things.
Note that some of these tests do not test AddNode in isolation (they also need to call other methods). That is not a problem. You are unit-testing the class, not individual methods.
I had to start writing some unit tests, using QualityTools.UnitTestFramework, for a web service layer we have developed, when my approach seemed to be incorrect from the beginning.
It seems that unit tests should be able to run in any order and not rely on other tests.
My initial thought was to have the something similar to the following tests (a simplified expample) which would run as an ordered test in the same order.
AddObject1SuccessTest
AddObject2WithSameUniqueCodeTest(relies on first test having created object1 first then expects fail)
AddObject2SuccessTest
UpdateObject2WithSameUniqueCodeTest(relies on first test having created object1 and thrid test having created object2 first then expects fail)
UpdateObject2SuccessTest
GetObjectListTest
DeleteObjectsTest(using added IDs)
However, there is no state between tests and no apparent way of passing say added IDs to the deletetest for example.
So, is it then the case that the correct approach for unit testing complex interactions is by scenario?
For example
AddObjectSuccessTest
(which creates an object, gets it to validate the data and then deletes it)
AddObjectWithSameUniqueCodeTest
(which creates object 1 then attempts to create object 2 with a fail and then deletes object 1)
UpdateObjectWithSameUniqueCodeTest
(which creates object 1 then creates object 2 and then attempts to update object 2 to have the same unique code as object 1 with a fail and then deletes object 1 and object 2)
Am I coming at this wrong?
Thanks
It is a tenet of unit testing that each test case should be independent of any other test case. MSTest (as well as all other unit testing frameworks) enforce this by not guaranteeing the order in which tests are run - some (xUnit.NET) even go so far as to randomize the order between each test run.
It is also a recommended best practice that units are condensed into simple interactions. Although no hard and fast rule can be provided, it's not a unit test if the interaction is too complex. In any case, complex tests are brittle and have a very high maintainance overhead, which is why simple tests are preferred.
It sounds like you have a case of shared state between your tests. This leads to interdependent tests and should be avoided. Instead you can write reusable code that sets up the pre-condition state for each test, ensuring that this state is always correct.
Such a pre-condition state is called a Fixture. The book xUnit Test Patterns contains lots of information and guidance on how to manage Fixtures in many different scenarios.
As a complement to what Mark said, yes, each test should be completely independent from the others, and, to use your terms, each test should be a self-contained scenario, which can run independently of the others.
I assume from what you describe that you are testing persistence, because you have in your steps the deletion of the entities you created at the end of the test, to clean up the state. Ideally, a unit test is running completely in memory, with no shared state between each test. One way to achieve that is to use Mocks. I assume you have something like a Repository in place, so that your class calls Repository.Add(myNewObject), which calls something like Repository.ValidateObjectCanBeAdded(myNewObject). Rather than testing against the real repository, which will add objects in the database and require to delete them to clean the state after the test, you can create an interface IRepository, with the two same methods, and use a Mock to check that when your class calls IRepository, it is exercising the right methods, with the right arguments, in the right order. It also gives you the ability to set the "fake" repository to any state you want, in memory, without having to physically add or delete records from a real storage.
Hope this helps!
The question may be a little vague but here's an example of what I want to know (pseudocode):
//start test-case for CreateObject function
{
// initialization of parameters
MyObject *obj = CreateObject();
// test results
}
//end test-case for CreateObject function
Is it necessary in this case to also deallocate the memory by calling "DestroyObject" function? [this is the particular case that gave birth to this question]
My personal opinion would be no, that I shouldn't undo what the function did, but if many tests would be carried out I could remain without memory/resources for that test-suite (not likely to happen but ...).
What do you think? In this case and also in a general case.
Thanks,
Iulian
In this case, you should deallocate the memory your test case has allocated. That way, you can use a tool that lets you run your tests and confirms that no memory was leaked. Letting your test code leak memory means that this would fail, and you wouldn't be able to tell for certain that the leak was in the test and not in your production code.
As to the more general situation, tests should clean up what they've done. Most unit testing frameworks let you implement a tearDown() method to do this. That way, if one test fails you'll know it's an issue with that test and not an interaction with another test.
You should really try to create all the mock objects on the stack (or use smart pointers). That way they get automatically destructed when test function goes out of scope.
Not directly to do with testing, but if you have C++ code that does stuff like:
MyObject *obj = CreateObject();
where "obj" is not a smart pointer or is not being managed by a class, then you have problems. If I were writing the test, I would say:
MyObject obj;
// tests on obj here
No matter what the results of your tests, obj will be correctly destroyed. Never create an object dynamically in C++ if you can possibly avoid it.
Typically, you want to test an isolated code path and an isolated piece of functionality, and you want to make it a fair test each time. That means starting fresh, setting up exactly what you need, and then discarding the testing environment when you're through. This avoids the problem of different test cases leaving behind side effects that might alter the results or behaviour of subsequent runs. It also means you can guarantee that your tests are all independent of each other, and that you can run any subset, in any order.
You'll find, however, that it's also quite common to have pre and post-suite set-up and tear-down methods, which set up a whole testing environment (such as a database or whatever) that a whole bunch of unit tests can execute against.