v7.2.1

Network Discovery and Connection Handling Overhaul

This release introduces a substantial overhaul of the network discovery and connection handling mechanisms, focusing on enhanced performance, stability, and maintainability.

Key Changes:

• Refactored Network Discovery (discover_network_nodes):
  • Introduced NetworkRange enum (enum NetworkRange { V4((Ipv4Net, usize)), V6((Ipv6Net, usize)) }) to encapsulate IP network ranges and scan sizes, improving type safety and code clarity.
  • Moved network range parsing from within the discovery tasks to the main discover_network_nodes function, significantly reducing redundant computations.
  • Simplified IPv6 address generation using a more deterministic approach based on prefix manipulation, reducing randomness within the scan loop and improving testability.
  • Improved concurrency control with Semaphore and atomic counters for managing active connections and peer count.
  • Implemented a timeout mechanism using tokio::select! to prevent indefinite discovery attempts.
• Enhanced Connection Handling (handle_connection and handle_network_event):
  • Implemented the handle_peer_message function, responsible for deserializing incoming messages using bincode and dispatching them for processing.
  • Corrected the handling of NetworkEvent::ChainRequest within handle_network_event. The event now includes the requesting peer's SocketAddr, enabling the sending of NetworkEvent::ChainResponse directly back to the requester.
  • Added peer_txs: RwLock<HashMap<SocketAddr, mpsc::Sender<NetworkEvent>>> to the Node struct to efficiently manage senders for connected peers. This eliminates the redundant passing of the tx sender to handle_peer_message.
  • Improved error handling within ChainRequest to manage potential errors during block retrieval from the local blockchain.

Bug Fixes:

• Fixed a critical bug where ChainResponse messages were not being sent to peers requesting chain data, preventing proper chain synchronization.
• Mitigated a potential race condition where the number of connected peers could briefly exceed MAX_PEERS due to asynchronous task execution. While the issue is largely mitigated with the new concurrency controls, further refinements for absolute guarantees may be explored in future releases.

Technical Details:

• Utilized Ipv4Net and Ipv6Net for structured IP address range representation and manipulation.
• Employed tokio::sync::Semaphore for controlled concurrency during network scans.
• Leveraged tokio::select! for efficient timeout management.
• Used RwLock for safe concurrent access to peer information and senders.

Full Changelog: v7.1.5...v7.2.1