this prevents the crash log from filling up when redis is unreachable
Add option to not reconnect and timeouts on initial connection
This prevents starting the client from taking forever if Redis is not responsive (which happened to one of our Redis instances during the recent EC2 outage)
This is useful because the automatic reconnect spawns a process that is not linked and could run forever (if the server is permanently unavailable). Robust clients will need to have their own reconnect logic anyway to handle failures with the initial connect.
pub/sub - fix the issue that controlling process not able to receive message after socket reconnect.
…he socket sending more messages than wanted. At large volumes of messages, we get into a situation where we can no longer handle the messages fast enough. Thanks to pmembrey for reporting.
Added documentation for pattern subscribe and example functions
Added basic psubscribe support
…on the punsubscribe tests though - seems to require too many ack_message() for my liking...
Fix return type spec