Question about which tools to use, bugs, the best way to implement a function, etc should go here. Don't forget to see if your question is answered in the wiki first! When in doubt post here.
How about garbage collection for interpreted languages I intend to add this as soon as its needed, right now I just garbage collect in the main thread at regular intervals between lines of interpreted code.
You may also do some kind of benchmark, adaptive adjustment of parameters based on collected data and runs in background.
For example, cache expiry/invalidation to reclaim memory.
Basically, IMO, kthread are workaround for monolith kernel which those functionality are likewise implemented in servers for micro kernel.