Skip to content

log :: 2026‐06

Eric Torreborre edited this page Jun 15, 2026 · 2 revisions

2026-06-12

Weekly Update (@etorreborre)

What did you work on this week?

Tightening-up some chain store operations:

  • child_tips was not operating on a snapshot of the database.
  • remove_blocks was not transactional
  • I removed the "optimistic transaction" mode because our design guarantees that we cannot have concurrent writes (we still need to batch writes to make them atomic).

Starting a seemingly small task: properly check the opcert last sequence number when validating a header:

  • The check itself involves accessing both the chain store and the ledger. Our current code was a bit convoluted so I did a refactor to fully delegate the responsibility for checking a header to a ledger effect.
  • Along with the check I improved the testing of headers based on the existing data files.
  • I added some persistence inside the ledger volatile state + db for opcert sequence numbers.
  • I started working on the extraction of those numbers from ledger states during bootstraping.

What outcome/key result did it support?

  • Being a reliable relay node

What's immediately next?

  • Testing the full chain which might imply the recreation of snapshots for the end to end tests.
  • Prepare PRs. I might have to do a separate PR just for the refactoring because it is quite large.
  • Resume the work on using cuddle and antigen to test the decoding of data types.

2026-06-12

Weekly Update (@KtorZ)

What did you work on this week?

  • Setting up the new release & distribution pipeline:

  • Make Amaru's bootstrap and initial start more flexible to allow for easier execution in Antithesis: PR#886

What outcome/key result did it support?

  • Distributing Amaru to SPO and preparing the ground for a more professional software delivery,

What's immediately next?

  • Investigate and address ledger bugs visible from syncing
  • Restore ledger test epoch snapshots for at least Preview and PreProd

Clone this wiki locally