Community Update 40

Owned by Medha Parlikar (Unlicensed)
Last updated: Sept 26, 2018
2 min read
  • Medha
    • Specified what needs to be done to shut down a node properly, so a node can recover the state on restart.
    • Team entered new sprints on Monday.  Sprints are large, and the team is aware of how aggressive the goals are.
    • PM team is focused on ensuring no blockers or churn comes up on work.
    • Medha is working on the Mercury release specification.
    • Medha is working on the Node 0.7 release announcement
    • PM team will also be testing Validator bonding later today.
    • During next week's Tuesday test we plan to test the process for bonding new validators to a network.
      • Look for updates in the node-testing channel in Discord on how to participate.  There will likely be some pre-requisite hoops to jump through around provisioning test Rev for bonding.
  • Rholang team:
    • Cost Accounting PR's have been cut.  Team is moving on to testing and documentation.  
    • Compromise specified to solve the matching issue in the tuplespace related to cost accounting.  Team is reviewing & will present to Greg later today.
    • New block storage architecture design and specification, initial draft will be done this week.
    • Now that the registry is complete, we are focusing on using it.  This involves porting contracts to use the registry, as well as crypto functions.
    • Removing the ability to send or listen on forgeable names. PR is in review
    • Progress made in making the interpreter stack secure.  All variants of AST are stack safe.  Pretty Printer is stack safe.
    • Enhanced the algorithm to find the most common ancestor in the DAG, addresses the InvalidBondsCache error
    • The Clique safety oracle uses the new in memory representation of the DAG.
    • Validator bonding is merged, including support for min / max fees.  Implementing Validator bonding fees.
    • Refactoring Casper to use new comms status messages. 
  • Storage/Node team:
    • Fine grained locking (disjoint channels) initial work is complete, PR is merged.  We observed an improvement in processing speed, but not throughput.
    • Node shutdown - We know Node operators want a graceful and reliable process. We also know there is complexity in the process of shutting down Casper, RSpace, and RNode.  Work planned for next sprint to orchestrate the shutdown across the components and provide a good experience for RNode operators.
    • A Node now reports its GitHash version.
    • Kademlia & RChain protocol separated.  Bootstrap node address format has changed as a result. See RNode bootstrap addresses for the new format.
    • Message queues (chunking & unlimited block size), meeting held with Casper team, Communications team will take responsibility for delivery of messages 
    • Work on threading improvements for RNode, RSpace, and Casper is a high priority for this sprint.
  • SRE team
    • Looking forward to having a new CI system in place by end of sprint
    • In process of upgrading to Java 10 with a plan to be done by end of sprint.
    • Working with the RSong team to bring the project into RChain CI
  • REV issuance
  • Incentivized testnet
    • Work in progress to develop this program. Look for more information in Oct. and the start of the challenges around Dec.
  • Website - https://developer.rchain.coop 
Date