RxJS Unit Testing: what's the functionality of the callback passed to TestScheduler

Question

I'm trying to use TestScheduler to write unit test for my RxJs functions. There are many posts related to it, and I'm following this post : https://medium.com/@kevinkreuzer/marble-testing-with-rxjs-testing-utils-3ae36ac3346a

There is one confusing point for the following part:

I don't understand why we need to pass the callback function into TestScheduler constructor. In the post, it mentioned that it is assertDeppEqual function which tells the TestScheduler how to compare values. The methods used to compare values depends on your testing framework.

But this explanation is not clear for me. I don't figure out what's the relationship between this function and the actual test logic we write, for example the following case:

in the above case, it assert that expectObservable().toBe(). what's the relationship between them?

Answer 1

Let's try to see what expectObservable(observable, subscriptionMarbles) and run(callback) methods do.

expectObservable() subscribes to a passed observable. The callback methods passed to subscribe() method populate actual array (this is what you get in a callback passed to TestScheduler constructor).

actual array gets pushed to this.flushTests array (wrapped in flushTest object). flushTests is an array of tests that have to be executed (remember, you can call multiple times expectObservable() or expectSubscriptions() methods). If you're using run() method (and you do based on the image), after a run() callback is executed, this.flush() method is called. This method runs your tests by calling this.assertDeepEqual() callback that you provided to TestScheduler constructor.

The actual and expected values from the test are passed in. expected is populated in the same method where actual is populated, just after it.

So, calling expectObservable() only prepares and converts Observable values to something comparable (an array of actual and expected values). It is your responsibility, as the official documentation says, to do the comparison based on the testing framework you're using.