[satran004]

DONT MERGE. Tracking PR for the ledger-cbor reader, the Dijkstra models and the Amaru removal across CCL (bloxbean/cardano-client-lib), Yaci and Yano. Tick items as they land.

Status (2026-11-09)

• Conformance on the ledger-cbor branches, newest pins (dataset 4fd977b2, NDC 01d11b7c), with the type-6 fix: CBOR 43,365/43,434, ledger rules 290 scenarios, UPLC 1,064, turbo 3,997 budgets — identical per sample to the previous run.
• Holistic review (Fable and Codex, independent): ledger-format reading is in CCL's ledger-cbor for CCL's models, Yaci's serializers and Yano's strict package; leftovers listed under "Follow-ups".
• Fixed: Yano accepted a Conway native script of type 6 (Dijkstra's guard) on the ledger-cbor branch; 8811854a6 on feat/ledger-cbor-reader, reviewed by both.

Decisions needed before release

• ADR 0004: status Proposed → Accepted (Dijkstra experimental); its code ships in #748
• Approve the listed differences in ADR 0003 sections 7.2 (M7a) and 7.3 (M7c)
• Indefinite-length headers: Yano accepts them (ledger 6176e413); released nodes read headers definite-only
• I10: Yano's strict reading at +30% per Conway block (about 40 µs)
• CCL #748 SonarCloud gate: 11 records with byte[] fields (array equality by reference) + 1 minor: fix, or accept in Sonar
• Yaci metadata CBOR: source bytes always (yaci-store#1215), as yaci#243 on main; done on the reader in yaci#244 (32726dc5), ADR 0003 r11
• Byron wide-head address: the harness writer fix
• ADR 0005 (CCL #750): accept, Q1–Q4 (does not block pre8)

Phase A — can merge now

• Yano chore(amaru): bump Amaru to v10.11.20261105 (368 scenarios) #240 Amaru bump — merge commit, not squash
• Yano chore: remove the Amaru WebAssembly validator (ADR-062) #241 Amaru removal (ADR-062) — retarget to main, wait for CI, merge
• Yano chore(amaru-wasm): stop attaching the wasm module to Yano releases #226 — close without merging (chore: remove the Amaru WebAssembly validator (ADR-062) #241 deletes the workflow)
• CCL #747 — rerun the native-image job (plugin-portal flake), merge
• Yaci Strict tx decoding does not enforce isMultiAssetSmallEnough (Haskell rejects at decode) #238 — version 0.5.0-pre16-SNAPSHOT on next (needed before the branch merges: next says 0.5.0-pre15, already released)

Phase B — the release chain

• B1 CCL #748: decisions above; update title/body (M0–M7c, ADR 0004); optional: one public reference-script reader (below); mark ready; merge. Then #750 retargets to master
• B2 CCL release 0.8.0-pre8
• B3 Yaci feat(ledger-rules): strict decoding through CCL's ledger-cbor reader (ADR 0003, M7c) #244, branch feat/ledger-cbor-reader (draft; 21 on next, incl. the metadata fix): CCL pre8; native-image build; mark ready; merge
• B4 Yaci release 0.5.0-pre16
• B5 Yano feat(ledger-rules): strict decoding through CCL's ledger-cbor reader (ADR 0003, M7c) #244, branch feat/ledger-cbor-reader (draft; 30 on main incl. the type-6 fix): rebase after Phase A; CCL pre8, Yaci pre16; conway3.block fixture; four reports at the newest pins; Fable + Codex; mark ready; merge
• B6 Yano release QA, then release

Follow-ups (not blocking; issues to file on approval)

CCL

• One public reference-script reader: NativeScript.deserializeScriptRef, PlutusScript.deserializeScriptRef, ReferenceScriptUtil.deserializeScriptRef each read [type, script] (V4 accepted by one, rejected by another) — candidate for pre8, as it adds public API
• RawItems.indefiniteRecordsAllowed compares eras outside EraSpec (ADR 0003 I2)
• tools/reader-diff OutcomeDiff recursion on deep unequal inputs
• #749 (aux V3 scripts with metadata; bignum/negative rationals)

Yaci

• BlockSerializer Dijkstra tx parts by position; AnchorSerializer.deserializeDI; Redeemer/Credential/StakeCredential model readers (tag 6 → null; intValue() wrap); placeholder BlockSerializerTest
• yaci#240, yaci#242

Yano (all pre-existing on main)

• Archive (YaciUtxoHistoryDecoder) and scripts API (ScriptResource) hash/return re-encoded native scripts; differs for non-canonical encodings
• ConwayTxInfoTranslator rejects bignum tag-30 rationals that validation accepts
• WitnessValidationRule reads reference scripts by hand and swallows errors; ProposalParamUpdateKeys, SyncBlock, TxIdentity re-read positions
• DefiniteLengthCbor (Scalus workaround) cites no upstream issue; the 00280 alias has no test that fails when the fixture changes
• Strict tx decoding does not enforce isMultiAssetSmallEnough (Haskell rejects at decode) #238 isMultiAssetSmallEnough: addressed in feat(ledger-rules): strict decoding through CCL's ledger-cbor reader (ADR 0003, M7c) #244 (6b0d0196e), close on merge; Dijkstra era support (experimental): ledger prerequisites #239 Dijkstra prerequisites; Low priority: check the conformance mutants against the Haskell ledger #242 Haskell oracle

Reported upstream offline: NDC 00280's expected name; dataset label bootstrap-witnesses-empty; the 64 aux-data samples.

[beanbot-gh]

Initial automated review · Aster

Hi @satran004, thanks for the PR! 🙏 This is an automatic first check by an AI reviewer. It does not replace maintainer review and may be wrong.

Reviewed commit: 8eff3d70505a42bb9402a5b6fc66dfb42e3e595e

I reviewed this draft as a tracking plan.

Issues found:

• docs/tracking/ledger-cbor-merge-plan.md:13 — Editing chore: remove the Amaru WebAssembly validator (ADR-062) #241’s base alone won’t trigger the main-targeted build, so the existing green checks can still be from the stacked base. Could we trigger fresh checks against main before merging?
• docs/tracking/ledger-cbor-merge-plan.md:23 — CCL #748’s corpus check fails because CI expects three tests and the suite now has four. Could we make a green corpus check explicit before releasing pre8?

Suggestions:

• docs/tracking/ledger-cbor-merge-plan.md:27 — Could we update corpora.env before fetching newer inputs, so the CBOR report records the dataset commit it actually tested?

Not tested: I checked the plan, source, and CI logs without running builds, tests, or reports.