I am dealing with a case where my tests pass or fail based on the order of declaration. This of-course points to not properly isolated tests. But I am stumped about how to go about finding the issue.
The thing is my junit tests derive from a class that is belongs to a testing framework built on junit and has some dependency injection container. The container gets reset for every test by the base class setup and so there are no lingering objects at least in the container since the container itself is new. So I am leaning toward the following scenario.
test1 indirectly causes some classA which sets up classA.somestaticMember to xyz value. test obj does not maintain any references to classA directly- but classA is still loaded by vm with a value xyz when test1 ends.
test2 access classA and trips up on somestaticmember having xyz value.
The problem is a) I dont know if this is indeed the case- how do I go about finding that ? I cannot seem to find a reference to a static var in the code...
b) is there a way to tell junit to dump all its loaded classes and do it afresh for every test method ?
You can declare a method with #Before, like
#Before public void init()
{
// set up stuff
}
and JUnit will run it before each test. You can use that to set up a "fixture" (a known set of fresh objects, data, etc that your tests will work with independently of each other).
There's also an #After, that you can use to do any cleanup required after each test. You don't normally need to do this, as Java will clean up any objects you used, but it could be useful for restoring outside objects (stuff you don't create and control) to a known state.
(Note, though: if you're relying on outside objects in order to do your tests, what you have isn't a unit test anymore. You can't really say whether a failure is due to your code or the outside object, and that's one of the purposes of unit tests.)
Related
I have a hard time to grasp how to implement unit tests in a class where all my fields are private.
The class is calculating a user's position with BLE and CoreLocation - not that important. I have a protocol, which when a new location is found I'm calling it and all the classes which conform to that protocol will receive a room id and room name. So, what that means is that literally all the fields in my class are private, because yeah, there's no reason any outside class should access them right? But that also means I can literally test nothing in that class, even though there are quite a few functions which I would like to test. I mean, I could just make the variables internal instead of private, but it just seems wrong to do that just to unit test. I've heard about dependency injection, but it just seems like so much effort.
For example I have this function:
private var beacons: [AppBeacon] = []
private var serverBeacons:[Beacon] = []
private func addBeacons(serverBeacons: [Beacon]){
for beacon in serverBeacons {
let beacon = AppBeacon(id: beacon.id, uuid: beacon.uuid, building: beacon.building, name: beacon.name)
beacons.append(beacon)
}
}
there's no way I can test whether the beacons array was actually filled up as I wanted to or not for example. The public features of my class are basically a function called startLocating and the result which is the room id and name and I know in black box testing which unit testing imitates (right?) I should not care about the intermediate steps, but honestly, with this much functionality should I just say, doesn't matter? And assume i did populate the beacons with some rssi values of my choice, the actual location algorithm is executed on a node.js server, so I honestly don't know what to test client side?
It's classic MVC and there's no way I can change it architecture until the deadline that I have, so I don't know what's the best way to go from here? Just don't test the functionalities? Make the fields internal instead of private? We do testing of the algorithm itself server side, so testing whether the the room id is the expected room id, is already tested.
I read on another post the following:
"Unit testing by definition is black box testing, which means you don't care about the internals of the unit you test. You are mainly interested to see what's the unit output based on the inputs you give it in the unit test.
Now, by outputs we can assert on several things:
the result of a method
the state of the object after acting on it,
the interaction with the dependencies the object has
In all cases, we are interested only about the public interface, since that's the one that communicates with the rest of the world.
Private stuff don't need to have unit tests simply because any private item is indirectly used by a public one. The trick is to write enough tests that exercise the public members so that the private ones are fully covered.
Also, one important thing to keep in mind is that unit testing should validate the unit specifications, and not it's implementation. Validating implementation details adds a tight coupling between the unit testing code and the tested code, which has a big disadvantage: if the tested implementation detail changes, then it's likely that the unit test will need to be changed also, and this decreases the benefit having unit test for that piece of code."
And from that I essentially understand it as that I should just not unit test this?
If you have a private var that would help you write unit tests, change it to private(set) var so that it can be read (but not changed).
Revealing the innards may bother you. If it does, it's possible that there's another type waiting to be extracted from the view controller.
First, the definition of 'unit-test' you have quoted is quite unusual: All definitions from the literature that I am aware of consider unit-testing a glass-box/white-box testing method. More precisely, unit-testing as such is actually neither black-box or white-box - it is the method used for designing the test cases that makes the difference. But there is no reason not to apply white-box test design techniques for unit-testing. In fact, some white-box test design techniques make only sense when applied for unit-testing. For example, when you investigate on unit-testing, you will encounter a lot of discussions about different code coverage criteria, like, statement coverage, branch coverage, condition coverage, MC/DC - for all of which it is essential to know the implementation details of the code and consider these implementation details during test case design.
But, even taking that into account, it does not change much for your particular case: The variables in your code are still private :-) However, you are approaching the testing problem in a too restrictive way: You try to test the function addBeacons for itself, in total isolation even from the other functions in that component.
To test addBeacons, your test cases should not only contain a call to addBeacons, but also some other function call, the result of which would show you if the call to addBeacons was successfull. For example, if you also have a function to find a beacon by name or position, you would first call addBeacon and, to check that this has succeeded, call one of these function to see if your added beacon will be found. These two calls would be part of the same test case.
I am writing the Unit test cases for my sqlite database class. I have five public API in that class.
My test cases goes something like below:
+ (void)setUp { // Note, this is class method and hence called only once during start of this test suite.
// Code to delete the existing sqlite DB file.
}
- (void)testDBManagerSingletonInstance {
DBManager *dbMgr = [DBManager getSharedInstance];
DBManager *dbMgr1 = [[DBManager alloc] init];
XCTAssertEqualObjects(dbMgr, dbMgr1);
}
- (void)testSaveAndDeleteNicknameAPI {
// Multiple Assert statements in this test.
}
- (void)testAllAccountStatusAPIs {
// Multiple Assert statements in this test.
}
Each of the single unit test is executed without any errors. But it fails when whole test suite is executed.
Probably, I know the Root Cause for failing. It is because when entire test suite is executed then all test runs in parallel and there is simultaneous update-delete happens in the database. Hence, when all unit tests runs it will fail.
But I don't know how can I fix that, because this is not Async, and hence I cannot use XCExpectation class.
Need assistance to resolve & understand the problem.
Tests based on XCTests don't run in parallel - they are run sequentially. To quote the docs :
Tests execute synchronously because each test is invoked independently one after another.
Since you've shown very little code, it's hard to say what is the real problem. It is very likely that you were close with your assumption - you should either improve your setUp (maybe switch to instance version from class version) and tearDown methods or introduce mocks, and perform your tests on a mocked database if possible.
You shouldn't work with singletons in a codebase you want to test. I think that your code uses the sharedSingleton instance of DBManager internally. Each unitest will change the internal state of that instance, all following unitests are therefore corrupted. You need to reset all changes after each test case in teardown.
But my suggestion is to avoid singletons in your base code by dependency injection. When creating an instance of your class inject the singleton instance of your DBManager. That makes unit testing easier. If you inject a protocol instead of an object you may even test your code with a protocol fake implementation.
I've a problem and I don't know the solution even after several hours of try and error and googling and stackoverflowing.
I have a view controller. I would like to pass per dependency injection an object. This object derives from a protocol. In general it is not a problem to setup a unit test. Also mocking works and the unit tests are running. So where's the problem?
I am testing only one class in my primary target. This class has absolutely nothing to do with view controller. But the code coverage is showing me a decent percent value of covering the view controllers. After a while I found out that when I hit the "test" button the project gets executed as if I push the "run" button. And because of that the view controller gets initialized and created and I have no chance to pass another dependency first, or before the tests getting executed.
So I need a method to distinguish between a test run and a real run, to pass in one case a real object and in the other case the fake object.
And my question is, how to to that? I wonder why nobody have this problem. I mean what gives me the code coverage tool if it shows me that methods are covered even though I haven't tested them.
The one and only class that I am testing:
And these are the coverage results (The bars are just gray because Xcode lost focus during screenshot. Otherwise they are blue.):
So I was expecting to see covered in the results just the class I am testing and not everything else. I know why this problem persists. The view controller has a dependency and this dependency after it gets initialized creates some more classes and so on. What I would like to do is to pass a fake object during unit testing and a real object during a real run. Just like It works in Visual Studio for non ui tests: If the tests are executed the application does NOT start up. The test runner just initialize the subjects under test and that's all. And this is what I want to achieve for iOS unit tests. I guess I've missed sth. very important :(
For all of us who have or will have the same problem. The solution is to specify environment variables for the test run. After that you can check if you are running your unit tests with code like this (assuming you have created an environment variable called "InTestMode" and set the value to "1" during test run:
let dict = NSProcessInfo.processInfo().environment
if let env = dict["InTestMode"] as? String?
{
return env == "1"
}
return false
I have a class which can be configured to do 2 slightly different things. I want to test both paths. The class is a descendant of UIViewController and most of the configuration takes place in Interface Builder. i need to verify that both Storyboard Scenes and their Outlets are wired up in the same way, but also need to check for the difference in behavior.
I'd like to use shared XCTest suites for this purpose.
One is aimed for use with the left hand, one for the right. Both appear after another when using the app. The first one (right hand) triggers a segue to the other. The last one (left hand) should trigger a different segue. This is where it differs, for example.
Now I want to verify the segues with tests. I'd like to create a BothHandSharedTests suite which both view controller instance tests use to verify everything they have in common. However, the BothHandSharedTests class is treated as a self-containing test suite, which it clearly isn't.
I came up with these strategies:
inherit from an abstract XCTest descendant, like described above (doesn't seem to be that easy),
write a test auite for the common properties and use one of the two as the Object Under Test, and add two smaller suites for the differences.
How would you solve this problem?
Here's a solution in Swift:
class AbstractTests: XCTestCase {
// Your tests here
override func perform(_ run: XCTestRun) {
if type(of: self) != AbstractTests.self {
super.perform(run)
}
}
}
I don't have a conclusive answer, but here's what I ended up doing.
I first tried the subclassing route. In the parent test (the "AbstractTestCase") I implemented all the tests that would be executed by the the AbstractTestCase subclasses, but added a macro so they don't get run by the actual parent test:
#define DONT_RUN_TEST_IF_PARENT if ([[self className] isEqualToString:#"AbstractTestCase"]) { return; }
I then added this macro to the start of every test, like so:
- (void)testSomething
{
DONT_RUN_TEST_IF_PARENT
... actual test code ...
}
This way, in the ConcreteTestCase classes which inherit from AbstractTestCase, all those tests would be shared and run automatically. You can override -setUp to perform the necessary class-specific set-up, of course.
However – This turned out to be a crappy solution for a couple reasons:
It confuses Xcodes testing UI. You don't really get to see a live representation of what's running, and tests sometimes don't show up as intended. This makes clicking through to debug test failures difficult or impossible.
It confuses XCTest itself – I found that tests would often get run even when I didn't ask them too (if I were just trying to run a single test) and the so the test output wouldn't be what I would expect.
Honestly it felt a little janky to have that macro – macros which redirect flow control are never really that good of an idea.
Instead, I'm now using a shared object, a TestCaseHelper, which is instantiated for each test class, and has has a protocol/delegate pattern common to all test cases. It's less DRY – most test cases are just duplicates of the others – but at least they are simple. This way, Xcode doesn't get confused, and debugging failures is still possible.
A better solution will likely have to come from Apple, unless you're interested in ditching your entire test suite for something else.
so I think I'm perhaps not fully understanding how you would use an IOC container for doing Integration tests.
Let's assume I have a couple of classes:
public class EmailComposer : IComposer
{
public EmailComposer(IEmailFormatter formatter)
{
...
}
...
public string Write(string message)
{
...
return _formatter.Format(message);
}
}
OK so for use during the real application (I'm using autofac here) I'd create a module and do something like:
protected override void Load(ContainerBuilder containerBuilder)
{
containerBuilder.RegisterType<HtmlEmailFormatter>().As<IEmailFormatter>();
}
Makes perfect sense and works great.
When it comes to Unit Tests I wouldn't use the IOC container at all and would just mock out the formatter when I'm doing my tests. Again works great.
OK now when it comes to my integration tests...
Ideally I'd be running the full stack during integration tests obviously, but let's pretend the HtmlEmailFormatter is some slow external WebService so I decide it's in my best interest to use a Test Double instead.
But... I don't want to use the Test Double on all of my integration tests, just a subset (a set of smoke-test style tests that are quick to run).
At this point I want to inject a mock version of the webservice, so that I can validate the correct methods were still called on it.
So, the real question is:
If I have a class with a constructor that takes in multiple parameters, how do I make one of the parameters resolve to a an instance of an object (i.e. the correctly setup Mock) but the rest get populated by autofac?
I would say you use the SetUp and TearDown (NUnit) or ClassInitialize and ClassCleanup (MSTest) for this. In initialize you register your temporary test class and in cleanup you restore to normal state.
Having the DI container specify all the dependencies for you has the benefit of getting an entire object graph of dependencies resolved. However if there's a single test in which you want to use a different implementation I would use a Mocking framework instead.