Sean

This issue should be resolved as of #891 - please feel free to pull in the latest master and test to ensure this is the case.

Hey @peedrr - Apologies, this was incorrectly triaged. However, the latest hiredis-py has support for all aspects of the RESP3 which are currently in-use, so if you use that library...

Can you please check out the latest master and test with that? Note that as of #891 the client has the same API as redis-py.

Hey @artesby - Thank you for the benchmarks! I suspect the issue lies with our connections re-implementation, which is using a [port of reds-py's SocketBuffer](https://github.com/aio-libs/aioredis-py/blob/master/aioredis/connection.py#L224-L361). This has the negative affect...

Ah, yes I recall the logic now. This pipelining would be more efficient, but it made debugging extremely difficult. Thanks for the example using a Protocol, I've not used this...

Really excellent work @Andrew-Chen-Wang - I've managed to read through most of this. I have a couple comments I'll add when I'm at a keyboard, but overall I'm glad to...

Only two failing tests remain: `test_bgsave[pool]` and `test_client_list_client_id[pool]`

> @seandstewart and @Andrew-Chen-Wang, you guys are the true heroes of aioredis. Apologies that I've not had the time I planned to work on this library after offering to help...

As of the latest batch of commits, this branch is now up-to-date with redis/redis-py@4e9cc015

Okay @abrookins @Andrew-Chen-Wang - this PR is up-to-date with [email protected] and we've got a green CI