Asynchronix logic

[nullchinchilla]

The asynchronix branch has logic ported from asynchronix, but there is a panic at the line

        assert_ne!(active_workers & (1 << worker_id), 0);

line in pool_manager.rs. @sbarral mentioned that this might be due to incorrect logic elsewhere, but the relevant logic in new_executor.rs is ported directly over.

[sbarral]

I see now where try_set_worker_inactive is called, I missed it before because the call was already commented out in the last commit.
I has a second look and couldn't see anything obvious, but seeing your spawn function made me realize that asynchronix is pretty specific and was not designed to support spawning tasks while it is running, which is probably something you need. I wouldn't be surprised if spawning during execution is racy.