Jean Boussier

> the internal usage of MessagePack.unpack feels a bit wrong. Yes it is. You want to use a recursive type: https://github.com/msgpack/msgpack-ruby#extension-types Also you'd probably save yourself a lot of trouble...

> Does it give the extra packer/unpacker argument? What extra packer? What argument? `recursive: true` allow the extension proc to read or write the stream. So that allow you to...

> won't support `reconnect_delay` and `reconnect_delay_max` arguments, is it planned to adopt this? No, this feature was wonky, so I replaced it by `reconnect_attempts` accepting a list of sleep durations,...

Thanks, all known outstanding issues are fixed. I'll leave open in case someone else find something.

@jmthomas indeed and that is by design. Redis 4.x would call to_s on all arguments, and this was the source of many errors and mistakes. (See the recent GitLab CVE)....

As for client.call is was never public API. If you want to add your own commands you can look at the send_command helper (sorry hard to link on mobile)

@jmthomas no. This is not an oversight, it's a deliberate change. From the Redis point of view commands are list of strings, it's ok to cast Numerics to string because...

> Let me know if there's a ticket open for this or how I can help. I think there is a misunderstanding here. `redis-rb` won't cast booleans for you, it's...

Yeah, sorry I should have linked to more info. Here's the GitLab issue: https://gitlab.com/gitlab-org/gitlab/-/issues/371098 In short they were passing API responses as `Sawyer::Resource` (a bit like Hashie or OpenStruct), so...

Another reason why I'm convince this is the right move, is that I've seen several instance of: ```ruby redis.mapped_hmset("hash", "key" => "val", "other-key" => nil) redis.hgetall("hash") # => { "key"...