Holochain 0.6 is now released and ready for developing on! We’ve written about it a couple times already, so I won’t cover it in depth, but the biggest new piece is a functioning immune system that blocks bad agents if their data fails validation. The validation and networking code have been thoroughly overhauled and tested to make sure they function correctly.
The immune system isn’t complete yet — currently, agents with an invalid membrane proof can initially join an app’s network and start communicating before their membrane proof has been validated — but validation is now correctly hooked into the network transport to ensure that invalid actions are responded to with network-level blocking. Additionally, warrants are delivered to anyone who queries a bad agent’s public key, so they can find out they need to block the agent even if they’re not a validator.
There’s still work to be done on the immune system in releases down the road. First, warrants are currently only delivered to the agent public key authorities, so you have to check for warrants using get_agent_activity. Second, membrane proof checking is currently only enforced via normal validation, not during handshaking, so unauthorised agents are able to join a network and access it for a short time before being warranted and blocked.
The dev team is also continuing to work on assessing Holochain’s performance with Wind Tunnel, a distributed test runner. Some of the insights from this have worked their way into the Holochain 0.6 codebase.
We invite you to upgrade your hApp to 0.6 right away (or scaffold your first hApp) and report any issues you encounter. Happy hApping!
