Periwink

Tests mainly needed on: - Behaviour when tick wraps around - Replication (clamping latest_tick and last_action), or just replication edge-cases - Sync - Inputs

Some facts: - the RealTime and VirtualTime are updated once at `First` - the FixedTime (including overstep) are updated once at `RunFixedUpdateLoop` - we reset `latest_server_tick` and `duration_since_latest_tick` in the...

Think about what to do with client inputs that happen before syncing if finished

1

- Either we discard the inputs, but then all the inputs that happened before syncing will be lost - Or we map them to the current tick and send them,...

Improve serialization

6

There are a bunch of tradeoffs regarding serialization/deserialization. The current strategy is: - to Encode, we maintain one Encode buffer per client. When we want to buffer a message to...

Stop trying to reliable send a message if we didn't receive an ack in too long

1

Just as a safety measure, on the send side, we want to prevent the reliable sender from keep trying to resend a message if we don't get any ack from...

Update the way we collect world changes that need to be replicated?

1

Currently, replication works like this: - we have systems that track entity spawn/despawn for all entities that have the Replicate component - for each component in the protocol, we have...

Currently, the server is responsible for starting Prediction/Interpolation for a given client, by specifying the `prediction_target` and `interpolation_target`. But fundamentally prediction/interpolation is purely a client concern. It would be nice...

use spawn batch

1

Benchmarks don't really show much difference...

Attempts to solve https://github.com/cBournhonesque/lightyear/issues/263 Issues: - everything in the parallel closure must be clone: - can use Rc for netservers.client_server_map since each client will only connect via one method -...