(u)pmap blocks when the pool size is 0

[jumarko]

I'm wondering if there's any rationale for accepting 0 as pool size. I've found this basically blocks forever (claypoole version 1.1.4).

(cp/upmap 0 (constantly 1) [])

As you may have guessed, I didn't want to actually pass 0 as thread pool size - it was a bug in my code. But I would prefer if claypoole functions would throw instead of silently blocking the current thread.