Fork me on GitHub

@plexus Wondering about parallelization in Kaocha -- is significant restructuring still a prerequisite? (, you last commented in April) Can you summarize what that restructuring would be?


hi @U3JH0P5LJ, seems I missed this message earlier. There's a #kaocha channel that I monitor more actively. In short: nothing has changed, and things are unlikely to change unless I decide to seek funding for another round of Kaocha improvements.


There are two sides to this: parallelization in the kaocha runner, and support for parallelization on the test type level. The former is probably not that hard, but it will impact a bunch of things. Test runs are no longer deterministic, so randomization no longer makes sense. fail-fast will only work approximately. On the test type we have to see where there is an assumptions that things run consecutively. E.g. kaocha-cljs will require a lot of work, because we would have to spin up multiple JS environments and distribute work. Not sure what the assumptions are for Cucumber, but I wouldn't be surprised if it blows up when you start using it from multiple threads. For clojure.test tests things should be relatively straightforward.