github agglayer/aggkit v0.11.0-rc12

latest releases: v0.12.0-rc4, v0.12.0-rc3, v0.12.0-rc2...
pre-releaseone month ago

aggkit v0.11.0-rc12

Changes since v0.11.0-rc11 (3 commits).

🚀 What's New

  • A permanently disabled source no longer pins the l2-to-lx block cursor (#1882) — headline fix. autoclaim/bridgedetector/l2_to_lx.go keeps one block-window cursor (l2-to-lx) shared by every source network, and processSource treated any fetcher.GetURL error as retryable. Since bridgeservicefinder.GetURL returns a disabled error for every network in IgnoreNetworkIDs — a deliberate, permanent exclusion that can never clear — a single ignored network with a VerifyBatches row anywhere in the scanned range wedged the shared cursor forever, and no L2-origin claim request was ever created for any source. Deployments whose ignore list covers most registered rollups saw L1-origin requests exclusively. A permanently excluded source is now treated as sourceUpToDate ("nothing to do") instead of sourceRetryLater, so it no longer contributes to retryBlock and no longer holds the window back. A genuinely retryable GetURL failure (service not yet announced, or unhealthy) keeps its existing hold-the-window behaviour, unchanged.
    • The permanence distinction is now expressed in the fetcher contract: ClaimCandidatesFetcher gains a detector-local ErrSourceDisabled, the permanent counterpart of ErrURLNotFound, and ServiceFetcher.GetURL is the designated translation point. A test pins the mapping so a future refactor cannot silently reintroduce the deadlock.
    • An excluded source is now reported once per source rather than on every poll.
    • Operational caveat: the excluded source's per-(source, destination) LER cursors are deliberately left untouched, so un-ignoring a network later still covers everything bridged while it was excluded — provided the source publishes another verify row. The shared window only rewinds by overlapBlocks (default 1), so to un-ignore a source that has stopped producing verified batches, reset the l2-to-lx block-window cursor so its old verify row falls inside the scanned range again.
  • Exact settlement tx resolution for WaitL1SettledGER (#1866): the bridge tracker now resolves the specific settlement transaction for the WaitL1SettledGER step instead of an approximation.
  • PR/release templates gain an API Updates section (#1867): repository templates only, no runtime change.

⚠️ Breaking Changes

  • None. No config, API, CLI or storage changes in this release candidate.

📋 Configuration Updates

  • None.

🔌 API Updates

🔌 Bridge service API

  • No changes.

🔌 Proxy API

  • No changes.

🔌 Others API

  • No changes. #1882 adds ErrSourceDisabled to the internal ClaimCandidatesFetcher contract in autoclaim/bridgedetector; this is an internal package with no out-of-repo implementers, so it is not a breaking interface change.

✅ Testing & Validation

  • 6 new unit tests in autoclaim/bridgedetector covering the disabled-source cursor fix, including TestL2ToLxDisabledSourceAtWindowStartDoesNotPinCursor (the production regression) and TestL2ToLxTransientURLErrorStillHoldsBlockCursor (guards the unchanged retryable path). Confirmed to be genuine regression tests: 4 of the 5 behavioural tests fail against the pre-fix code.
  • go test ./autoclaim/... ./bridgeservicefinder/... — all packages ok; golangci-lint run — 0 issues.
  • No new e2e test cases.

📦 Full Changelog

  • 4dfdeda2 fix(autoclaim): don't let a permanently disabled source pin the l2-to-lx block cursor (#1882)
  • af2f650d fix(bridgetracker): resolve exact settlement tx for WaitL1SettledGER (#1866)
  • c5b0be5b chore(templates): add API Updates section to PR/release templates (#1867)

Full diff: v0.11.0-rc11...v0.11.0-rc12

Don't miss a new aggkit release

NewReleases is sending notifications on new releases.