Is it expected that io.aleph.dirigiste.Pool.acquire can block for a long time (experienced ~1 second blocks on a cold connection pool), basically making aleph.http/request not really asynchronous sometimes?

Looking at the code, I see that it's intended that the caller will create the object if it can't be obtained from the queue. Then I don't quite follow what is the purpose of the callback API here if the object acquisition is synchronous anyway.