High level overview​

This week, the Mithril team completed the implementation of the publisher and consumer for the DMQ network using the n2c local message submission and n2c local message notification mini-protocols from the Pallas library.They also worked on creating a fake DMQ node to facilitate end to end testing. The team also completed the stabilization of the Cardano database v2 backend within the Mithril client CLI and library. Additionally, they kept implementing the certificate chain synchronization in the follower aggregator, prepared the upgrade to the Cardano node v.10.5, and enhanced some workflows in the CI.

Finally, they continued preparing the repository for the Rust 2024 edition upgrade and progressed with refactoring the STM cryptographic library.

Low level overview​

• Completed the issue Implement a publisher with the n2c Local Message Submission mini-protocol with Pallas #2539
• Completed the issue Implement a consumer with the n2c Local Message Notification mini-protocol with Pallas #2540
• Completed the issue Stabilize Cardano DB v2 client CLI/library - Phase 2 #2577
• Completed the issue Create and store artifacts for the headless tests in the client CI workflow #2579
• Completed the issue Use alternative to Sendgrid for email notifications in CI #2617
• Completed the issue Build Linux ARM binaries in the CI #2178
• Worked on the issue Upgrade to Rust 2024 edition #2582
• Worked on the issue Synchronize certificate chain of follower aggregator #2534
• Worked on the issue Enhance STM library readability #2568
• Worked on the issue Implement a fake DMQ node with Mithril relay #2627
• Worked on the issue Upgrade to Cardano 10.5 #2623

High-level summary​

We focused on our new release: 0.22.0, and made some quality of life improvements. We spent a good amount of time investigating a tricky issue related to deadlocks, and will continue our investigations. We will also focus on making a patch release 0.22.1 to fix the issues identified.

What did the team achieve?​

• Released version 0.22.0!
• Added the ability to run our "smoke tests" with Blockfrost #2048
• New feature to derive the hydra-scripts-tx-id parameter from a --network parameter #1441
• Investigated an issue related to a deadlock when loading a large state #2089
• Continued planning the roadmap for Hydra in 2025-2026
• Continued to support the glacier drop

What's next?​

• Release version 0.22.1!
• Further investigate the potential deadlock #2086
• Fix peers with different persistence causing hydra-node to crash #1937
• Continue working on improving etcd UX #2054
• Continue to focus on supporting the glacier drop

High-level summary​

• Added support for late input block inclusion in the Haskell implementation of the Leios simulator (#413).
• Held the Consensus Working Group call (recording), where we discussed the tie-breaker change proposed in #1548. A request for comments has been submitted to the Technical Steering Committee.
• Released cuddle on Hackage.
• Optimized handle usage in the on-disk backend of UTxO-HD (#1563) and exposed a configuration option for setting the maximum number of open handles (#6256).
• Incorporated further improvements to resource management in UTxO-HD (#1564).
• Implemented a new query to retrieve the maximum major protocol version supported by the code (#1562).
• Introduced new versions of GetPoolDistr and GetStakeDistribution queries, now returning additional information (#1540).

High level overview​

This week, the Mithril team focused on implementing the publisher and consumer for the DMQ network using the n2c local message submission and n2c local message notification mini-protocols from the Pallas library. The team also focused on stabilizing the Cardano database v2 backend within the Mithril client CLI and library. Additionally, they initiated the implementation of certificate chain synchronization in the follower aggregator and successfully completed refactoring the mithril-common crate.

Finally, they started upgrading the repository to Rust 2024 edition and continued refactoring the STM cryptographic library.

Low level overview​

• Completed the issue Enhance ledger state snapshot converter command in client CLI #2571
• Completed the issue Split Docker client download and snapshot convert in manual client test in CI #2578
• Completed the issue Client WASM npm package publication fails on versions discrepancy #2585
• Completed the issue Split mithril-common crate - Phase 2 #2392
• Completed the issue Re-organize STM library structure #2369
• Completed the issue Split Docker client download and snapshot convert in manual client test in CI #2578
• Completed the issue Ensure Cargo.lock file is updated in the CI #2549
• Completed the issue Reactivate v2 version of action-gh-release GitHub in the CI #2581
• Worked on the issue Implement a publisher with the n2c Local Message Submission mini-protocol with Pallas #2539
• Worked on the issue Implement a consumer with the n2c Local Message Notification mini-protocol with Pallas #2540
• Worked on the issue Upgrade to Rust 2024 edition #2582
• Worked on the issue Stabilize Cardano DB v2 client CLI/library - Phase 2 #2577
• Worked on the issue Synchronize certificate chain of follower aggregator #2534
• Worked on the issue Enhance STM library readability #2568

Overview of sprints 89, 90​

We released the ouroboros-network-0.21.2.0 version, which was adopted by cardano-node-10.5.

We created integration branches for ouroboros-network:cardano-diffusion with ouroboros-consensus, ekg-forward and cardano-node repositories (all called coot/cardano-diffusion-integration. We merged cardano-diffusion: integration changes to ouroboros-network.

After the release of QuickCheck-2.16, which includes one of our contributions (Added Every and Some monoid, we removed dependency on quickcheck-monoids and deprecated its API, #5142.

We merged the following pull requests:

• inbound governor performance improvement
• [lower the time to cache DNS errors][#5143]
• we fixed bootstrap test failure in sim-net

We also worked on improving our CI (Hydra) experience. In the short term, we decided not to run computationally expensive simulations on {aarch64,x86_64}-darwin architectures. These simulations are platform independent, and they quite often timed out due to not enough CPU bandwidth, see #5145

Detailed overview​

For a more detailed overview, please take a look at the sprint boards:

• Sprint 89
• Sprint 90