You might be asking for too much, but basically - test assertions are tracked via the clojure.test/report multimethod - if you can call this report method, and if you can await for each spawned thread to finish before reporting everything, then you might be able to do it.

@U3Y18N0UC many thanks for you reply 🙏:skin-tone-2: sweet I’ll keep hammering this. I think it might actually be working, and there’s simply a test-runner display issue. I forgot to mention I’m using Kaocha, so I’ll take another look at their docs to see what else I might do to improve the test statistics. Thanks again for your suggestions!

How would you design a system so that you minimized testing to only those things which changed? I know this is a hard problem*TM, what practical steps do you take to try and get 80% of the way there?

@U0DJ4T5U1 honestly I just try to get it to work first, then try to make tests vaguely sane, which depends greatly on the libs/toolkits you are using. I’m using Lacinia/pedestal for the main stack, websockets for graphql subscriptions. In terms of design you can become limit by the stack you use, the choices made, and the tooling available - I’m not even sure minimising is even worth it - how do you know that a change in domain entities won’t affect messages sent out of a websocket to the client? I run my tests all the time to make sure I assume as little as possible - ~200 tests, runs within 5 mins, faster if done in parallel. HTH

Thanks rowland, errrr i didn't realize my question got asked inside a thread. I didn't mean to do that.