Arthur Silva

- [x] accept/change configuration through the protocol - [ ] persist/load queue configuration

Amazon SQS has the concept of a receipt that's different from the Message ID, its used to delete (ack) the message afterwards. Also, if a messages timeout a new receipt...

This is what Floki does right now **RPUSH** push one or more messages `RPUSH queue_name message1 [message2, ...]` Returns the number of messages inserted **HMGET** get one or more messages...

Add loom tests

3

Loom https://github.com/tokio-rs/loom This can be a significant amount of work but in my experience it is MUCH more effective catching concurrency bugs compared to stress tests. Eg. https://github.com/tokio-rs/tokio/blob/master/tokio/src/sync/tests/loom_broadcast.rs

Using Timeout/Interval with millis >= i32::MAX results in incorrect behavior

1

### Describe the Bug Millisecond arguments are taken as u32 and internally cast (unchecked) to i32. Leading to the timer/interval firing after the wrong duration. Look for `millis as i32,`...

At least for reads, right now a MGET will immediately fail if it spans multiple partitions.

Support CRDTs datastructures

1

Sucredb would be a great fit for those structures, even if it probably can't support all commands reasonably. - [x] Sets - [ ] Sorted Sets - [x] Counters -...

It might make sense to support more commands.

Right now it's non-existent, except for `FabricError::NoRoute`