I remember testing chapter from your lectures, and I really liked it. This chapter looks excellent to me! I have a few comments, maybe they don't have any sense:

- It would be great if the text is also covered with the GitHub repository with working examples.

- With event sourcing, we can easily decouple our code base into a lot of small code bases, since events are only single source of truth. Each event is language agnostic so we can basically build each command/event handler in a different language (while they share the same event schema). With technologies like OpenFaas, gRPC, Docker we can easily assemble system like this. Having said that, I think we can move the test framework one step higher so we can basically test REST API instead of a command handler in a specific language. We could have unified tests, written for example as Postman collection, and we can use header value to switch between in-memory or real event store.

- How to write a test that describes a situation where we have multiple events generated in multiple steps with the same correlation ID and possibly generated in different streams? We can write multiple tests for each stream/aggregate but it would be nice to have a built-in link between those tests somehow.

- In the last project we have been building in my company we use "stream" terminology instead of "aggregate". You chose to use "aggregate". Is there some important reason for this :)

Kind regards,
Stanko Culaja.