v1.1.7

Revert logging change that caused some error information to be dropped from metrics
Add mechanism to create new ETCD client instance in case the client dies, which caused service discovery to stop working
Gracefully drain nats server connections on server shutdown

v1.1.6

Upgrade multiple libraries and dependencies
Fix make setup
Fix log volume problem caused by etcd watcher dying

v2.0.1

Added

• IsRunning on pitaya.Pitaya (#163)

v2.0.0

This major version is no more compatible with v1. So upgrading to it might require refactoring the package usage.

Changed

• The entire Pitaya package API is now non-static. More details at #133.

v1.1.5

Changed

• BeforeHandler return context (#153)
• Move log information to error message in RPC call (851525f).

v1.1.3: Add more information to nats timeout error logs (#151)

Changed

• Add more information to remote call error logs
• Do not log errors on empty messages because of EOF
• Fix race condition where agent heartbeat could write to closed channel

v1.1.2

Update Pcall in order to log a better stacktrace.

On the previous implementation, the stacktrace would be logged as a debug
log and in multiple lines. This breaks tools like Kibana that consider
line breaks as different logs.

This version also changes the stacktrace log to be error instead of
debug.

V1.1.1

This version avoids concurrent writes to the client's socket, fixing #112

v1.1.0...v1.1.1

First Stable Version

This is the first stable release of Pitaya.

We have already 4+ big games using it in production, so the API should not change much from now on.

There's also an improvement in the TCP acceptor for correctly treating it as a streaming protocol rather than a messaging one.
Acceptors are now more decoupled from the agent, so it must be even easier to create new ones.

v0.15.0...v1.1.0

v0.15.0: Lazy gRPC connections (#109)

* Improve grpc client logging

* Lazy grpc connections

* Avoiding concurrent connection

* Add locking on grpc disconnect