Madelyn Olson

> we can easily move it to happen earlier (that's the right thing to do anyway, i.e. do tests before actual expensive init). I think moving it earlier also makes...

@PingXie Given that we are after the code cut-off for Redis 7, we probably won't include this, will probably merge this after 7 goes GA.

@PingXie I see a lot of your points, and I agree with most of them. I have two concerns. Let's focus on the implementation, if it's ready we can of...

@PingXie Ok! I don't know about json, we don't have a JSON parser today in Redis and I don't think we really want to take a dependency on one. I...

> My primary point is to show that there is a way forward. I don't have a strong opinion that we should use JSON. Can you point me to your...

@ushachar Will you take a look at this since we presumably want the same "shard ID" to be the same between cluster v2 and cluster v1.

@ushachar One of the assumptions being discussed here is that ShardIDs may not necessarily be monotonically increasing, and externally imposed.

On the other thread we decided to go with the aux format: ``` 823ca5eb86404530a2cd1f6beee7ed9c00e786fb 127.0.0.1:30001@40001,host-name master aux1:val1 aux2:val2 aux3:val3 ... - 0 1656441238093 1 connected 0-5460 ```

The only other change is unit tests can now be written for cluster features so they run in the ci, in tests/unit/cluster/*, so all of your changes need to get...

> This test conversion is harder than I thought. @madolson I don't want to block your code review on the test code refactoring. I can take your feedback on other...