@eskimo Thanks for the reply! I understand this much, and I guess what I'm asking is whether it's possible to work with these data streams. If it's not feasible, why is ZLIB an option with AA? The use case would seem to be exclusively about compatibility. LZFSE has an equivalent ratio while being much quicker. I could certainly create my own async streams with Compression, like you mentioned, but I'm attempting to exhaust this route first.

I suppose one could use zlib in a different context, compressing strings or otherwise using something other than whole individual archive containers. Oh well - I know this falls into the "unsupported" bucket of dumb questions, but being that AA is part of Accelerate I've wondered whether there is some more magic happening than simply using GCD and Compression to split up stream work items. At the end of the day, I have a thousand other priorities, but my curiosity will remain.