Draft
Conversation
* chore(rust): merge all the rust workspaces * chore(ci): formatting, fixing dockerfile and ci * chore: more ci fixes * Update .circleci/continue/rust-ci.yml Co-authored-by: Sebastian Stammler <seb@oplabs.co> * Update .circleci/continue/rust-e2e.yml Co-authored-by: Sebastian Stammler <seb@oplabs.co> * Update .circleci/continue/rust-e2e.yml Co-authored-by: Sebastian Stammler <seb@oplabs.co> * chore: pr comments * nit: fix kona path in op-challenger * chore: fix prestate artifacts path --------- Co-authored-by: Sebastian Stammler <seb@oplabs.co>
…nned tag (ethereum-optimism#19183) Now that kona lives in the monorepo, the prestate build no longer needs to clone the optimism repo at a pinned commit to build cannon. Instead, cannon is built from the local monorepo source via a Docker named build context, removing the cannon_tag config file and simplifying the build. Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…-optimism#19163) * proofs: port TestInteropFaultProofs action test to devstack * fix(kona/client): produce InvalidTransition when derivation falls short of target block * Update Makefile Co-authored-by: graphite-app[bot] <96075541+graphite-app[bot]@users.noreply.github.com> * fix makefile indent * revert Makefile changes * revert rust/kona/justfile changes --------- Co-authored-by: graphite-app[bot] <96075541+graphite-app[bot]@users.noreply.github.com>
…#19101) * chore(rust): move docs to unified rust/docs/ directory * chore(rust): update repo links from alloy-rs/paradigmxyz to ethereum-optimism
…optimism#19155) - Add tag classification to distinguish Rust crate groups from Docker-only groups - Add publish-crates CI job for crates.io publishing with dependency-ordered releases - Add op-reth Docker bake target and compute-git-versions entry - Add cargo-release configuration for alloy-op-hardforks and op-reth - Add .dockerignore for rust/ build context - Fix op-reth DockerfileOp manifest paths
Co-authored-by: Amp <amp@ampcode.com>
Co-authored-by: Georgios Konstantopoulos <me@gakonst.com> Co-authored-by: Amp <amp@ampcode.com>
…radigmxyz/reth#21686) Co-authored-by: Amp <amp@ampcode.com>
…e` (op-rs/op-reth#204) - Adds new crates `reth-optimism-exex` and `reth_optimism_trie` - Moves `reth_exex::external_proofs` -> `reth_optimism_exex` - Moves `reth_exex::external_proofs::storage` -> `reth_optimism_trie`
Seems like when `reth-optimism-trie` depends on `reth-db-api` indirectly, the doctest crate couldn’t resolve `reth_db_api::…` and failing. Closes op-rs/op-reth#233 Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com>
This PR separates storage and account cursors which makes sense because these two cursors may have very different seek/next implementations. In the case of in-memory, a single cursor impl can still handle both, but for MDBX, it makes sense to implement separate cursors. Closes op-rs/op-reth#236 --------- Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com>
Based on op-rs/op-reth#203 , op-rs/op-reth#204 This PR implements `StateProvider` given a `OpProofsStorage` instance. It reads most data from the external database, but falls back on the latest provider for block hashes and code by hash similar to the existing historical provider in Reth. This is an important part to implementing live syncing since we're running the sync process on the DB being created. In the Reth implementation, `proof.rs` is contained in `trie/db`, so I think it makes sense to go in our DB crate. The `provider.rs` is in the `reth-provider` crate, but I don't think a separate provider crate helps us here, so I think we should also include that in the trie crate as well. --------- Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com>
Based on op-rs/op-reth#229 Adds tests to ensure that storing TrieUpdates that include deletions actually deletes the nodes at that block height, and that updates take precedence over deletions (same as `write_trie_updates` in Reth). Closes op-rs/op-reth#238
…p-reth#198) Based on op-rs/op-reth#197 This PR implements the live state collector on top of the `OpProofsStateProvider` created in the previous PR. This sync process tries to re-execute all blocks from the current block of the external database to tip. --------- Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com>
…#222) Closes op-rs/op-reth#164 --------- Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com>
Fixes op-rs/op-reth#225 (does not track errors as these are fatal currently)
…#251) Fixes op-rs/op-reth#237 Adds more concrete error variants for `OpProofsStorageError` Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com>
…eth#248) Fixes op-rs/op-reth#247 Removes the `block_number` parameter from: - store_account_branches - store_storage_branches - store_hashed_accounts - store_hashed_storages Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com>
…p-rs/op-reth#253) Fixes op-rs/op-reth#215 --------- Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com>
…op-reth#250) Ref op-rs/op-reth#241 - Fixes comment from deps fix in op-rs/op-reth#229 (fix correct but comment misleading) - `serde-bincode-compat` is a feature which must be enabled when the `bincode` dependency is used. Asides for dev-deps, this only happens in `reth-stages`. `reth-exex` runs in `reth-stages`. In this wokrstream we use the `TrieUpdate` in exex. `serde-bincode-compat` is already implemented for `TrieUpdates` in `reth-trie-common`, this PR simply makes that feature accessible via `reth-trie` and in turn `reth-optimism-trie`, in order to enable it in `reth-optimism-exex`.
Closes op-rs/op-reth#212 --------- Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com>
…mism#19292) * ci: add cannon-builder image to Docker CI builds Add the kona cannon-builder (Rust MIPS64r1 toolchain) image to the branch and tag Docker build workflows. This publishes the image to the shared artifact registry so it can be consumed by prestate builds. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * rust: bump op-reth to v1.11.1 and update cannon build infrastructure Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: geoknee <georgeknee@googlemail.com>
…align on single ComponentID type (ethereum-optimism#18877) * stack: remove backward compatibility type aliases for Phase 6 cleanup Complete the ID type system refactor by removing all backward compatibility code. All code now uses ComponentID directly with typed constructor functions. Changes: - Remove 19 type aliases from component_id.go (L1ELNodeID, L2BatcherID, etc.) - Remove Kind = ComponentKind alias - Update KindProvider interface to use ComponentKind return type - Update context.go functions to use ComponentKind instead of Kind - Fix test files using incorrect constructor patterns The type system is now fully unified: - Single ComponentID type with kind, shape, key, and chainID fields - Typed constructors (NewL2BatcherID, NewL2ELNodeID, etc.) return ComponentID - ComponentKind enum for type discrimination - Simplified Matcher[E] interface with single type parameter * fix(op-devstack): drop misplaced kona tests and finalize componentid cleanup * fix(op-devstack): restore Go build after ComponentID refactor
…9290) The OP Stack supervisor is deprecated. Remove all supervisor-related Rust crates, the supervisor binary, Go e2e tests, and Grafana dashboard recipe. Update workspace configuration, justfiles, and documentation. ci: remove kona-supervisor CI jobs and env vars Remove the kona-supervisor-e2e-tests job, kona-supervisor-e2e workflow, and RUST_BINARY_PATH_KONA_SUPERVISOR env exports from CircleCI configs. Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat(kona): add Karst hardfork support Add Karst between Jovian and Interop in the fork ordering. Karst has an empty NUT bundle and requires empty activation blocks, matching the Go implementation in op-node (PR ethereum-optimism#19250). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(kona): chain Karst to Interop in op_fork_activation When karst_time is None, op_fork_activation should chain to Interop rather than returning Never. This maintains consistency between the is_karst_active() method and op_fork_activation(). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(kona): don't chain Karst to Interop in op_fork_activation Revert ebd614d. Interop is an independent fork that doesn't participate in the sequential activation chain. Karst is the terminal fork before Interop, matching how Jovian was terminal before Karst was added. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix rust formatting * Chain karst activation to interop * fix(kona): address Karst review feedback * fix(kona): add missing karst_time to registry test configs and fix doc link --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…timism#19341) `variant.rs` defined `EthereumDataSourceVariant` and implemented `AsyncIterator` for it, but `AsyncIterator` is not defined anywhere in the kona codebase and the file was never included in `sources/mod.rs`. The type was therefore completely unreachable and unusable at runtime. `EthereumDataSource` in `ethereum.rs` already covers both the calldata and blob cases through the `DataAvailabilityProvider` trait, making `EthereumDataSourceVariant` redundant. Fixes ethereum-optimism#19340 Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
ethereum-optimism#19355) The `load_calldata` match on `TxEnvelope` used a catch-all `_ => return None` that silently dropped EIP-4844 (type 3) transactions. The derivation spec (derivation.md:504) and op-node's `isValidBatchTx` both explicitly accept type-3 transactions, so kona must do the same to avoid divergent L2 state in the pre-Ecotone window. Add a `TxEnvelope::Eip4844` match arm that extracts `to` and `input` like the other accepted transaction types, and update the corresponding test from asserting the tx is ignored to asserting its calldata is included. Closes ethereum-optimism#19352 Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
… panics on truncated input (ethereum-optimism#19361) * fix(kona/protocol): add bounds checks in span batch decode to prevent panics on truncated input Multiple span batch decode functions panic on truncated input instead of returning an error. This adds explicit length checks before each unsafe slice operation: - prefix.rs: decode_parent_check and decode_l1_origin_check now check r.len() >= 20 before split_at(20) - transactions.rs: decode_tx_sigs now checks r.len() >= 64 before indexing r[..32] and r[32..64] - transactions.rs: decode_tx_tos now checks r.len() >= 20 before indexing r[..20] On short input, each function returns SpanBatchError::Decoding(...) instead of panicking, allowing the batch to be dropped gracefully — consistent with Go op-node's io.ReadFull behavior. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * style: fix rustfmt formatting in span batch decode Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * style: fix clippy field_reassign_with_default in span batch tests Use struct initialization syntax instead of Default::default() followed by field reassignment. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * style: fix rustfmt formatting for struct initialization Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…alidator transition (ethereum-optimism#19360) * fix(kona/derive): copy origin field during Holocene BatchQueue→BatchValidator transition (BQ-12) On Holocene activation BatchProvider::attempt_update() transitions from BatchQueue to BatchValidator. It was copying l1_blocks but not the origin field, leaving BatchValidator.origin as None. Go's BatchMux.TransformHolocene() copies both fields (batch_mux.go:67-68): bs.l1Blocks = slices.Clone(bp.l1Blocks) bs.origin = bp.origin Without the origin copy, the first update_origins() call after transition always enters the self.origin != self.prev.origin() branch (None != Some(…)). This causes two failure modes: 1. Normal case: current L1 block is pushed onto l1_blocks again, creating a duplicate entry that corrupts the two-slot epoch window. 2. Lagging case: l1_blocks.clear() discards all transferred history, then next_batch() returns MissingOrigin.crit(), halting derivation. Fix: copy batch_queue.origin into bv.origin alongside l1_blocks, matching the Go reference implementation. Closes ethereum-optimism#19356 * fix: apply rustfmt formatting fix Wrap long comment line in batch_provider.rs to satisfy nightly rustfmt check. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: rename test to reflect expected behavior The test verifies that origin IS transferred during Holocene transition, so rename from _not_transferred to _transferred. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * Remove dangling reference to random issue ID. --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
… in BlobData::fill (ethereum-optimism#19362) * fix(kona/derive): return Reset instead of Critical on blob under-fill in BlobData::fill When the blob provider returns fewer blobs than there are blob-hash placeholders (under-fill), BlobData::fill previously returned BlobDecodingError::InvalidLength which mapped to PipelineErrorKind::Critical — permanently terminating the pipeline and requiring operator restart. This is wrong. Under-fill is a transient condition: - The provider may not have all blobs cached/synced yet for a newly confirmed block - The L1 block may have been reorged out, causing the provider to return no blobs In both cases the correct response is to reset the pipeline and retry from the last safe L1 block, exactly as op-node does: fillBlobPointers wraps the same error as NewResetError at blob_data_source.go:110. Changes: - Add ResetError::BlobsUnderFill(usize, usize) to pipeline.rs - Add BlobProviderError::NotEnoughBlobs(usize, usize) to sources.rs, mapped to ResetError::BlobsUnderFill.reset() in From<BlobProviderError> for PipelineErrorKind - Change BlobData::fill return type from Result<bool, BlobDecodingError> to Result<bool, BlobProviderError>, returning NotEnoughBlobs on under-fill - Update blobs.rs caller to use the simplified .into() conversion - Update tests to assert under-fill produces a Reset-level error Fixes ethereum-optimism#19359 * fix(kona/derive): remove unused BlobProviderError import The refactor to use Into::<PipelineErrorKind>::into made the direct BlobProviderError import unnecessary, triggering a clippy warning. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: apply cargo fmt to blobs.rs import ordering Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(kona/derive): address review feedback on blob under-fill PR - Wrap BlobProviderError in ResetError::BlobsUnderFill instead of duplicating fields - Use named fields for BlobProviderError::NotEnoughBlobs - Remove unnecessary map_err in BlobSource fill loop - Add Clone to BlobDecodingError and BlobProviderError Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…ethereum-optimism#19358) * fix(kona/derive): make syscfg update failure non-fatal in L1Traversal When `update_with_receipts` fails to apply system config changes from L1 receipts, both `PollingTraversal::advance_origin` and `IndexedTraversal::provide_next_block` were returning a Critical pipeline error, permanently halting the pipeline. op-node's reference implementation (`l1_traversal.go:78-82`) treats this failure as non-fatal with the comment: "failure to apply is just informational, so we just log the error and continue." This commit aligns kona's behaviour with op-node: on syscfg update error, log a warning and continue advancing the origin rather than halting. Fixes ethereum-optimism#19353 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: apply rustfmt to indexed.rs test line added in this fix The block2 struct literal added in this fix exceeds the 100-char line width limit. Wrap it to comply with rustfmt requirements. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
…et constructors (ethereum-optimism#19452) * refactor(op-devstack): simplify acceptance testing around direct preset constructors fix acceptance and kona e2e test harness regressions fix same-timestamp cycle test and kona shared runtime wrappers address review feedback and restore devstack test coverage fix(ci): stabilize rust e2e websocket tests * refactor(op-devstack): validate preset option support * fix(ci): reduce rust subprocess log volume
…reum-optimism#19364) * fix(kona/derive): add over-fill check in BlobSource::load_blobs After the blob-pointer fill loop completes, add a post-loop check: if `blob_index < blobs.len()` the provider returned more blobs than were requested. Return `ResetError::BlobsOverFill` (→ PipelineErrorKind::Reset) rather than silently discarding the extras. This mirrors op-node's `fillBlobPointers` check at blob_data_source.go:162-163 which returns `fmt.Errorf("got too many blobs")` wrapped as `NewResetError`. Over-fill can occur with buggy blob providers (e.g. third-party RPC services that ignore the requested hash list) or in rare L1 reorg scenarios where the blob set shifts between hash collection and fetch. Changes: - Add `ResetError::BlobsOverFill(usize, usize)` variant to `pipeline.rs` (symmetric with the existing `BlobsUnderFill` variant). - Import `ResetError` in `blobs.rs` and add the post-loop guard. - Add `should_return_extra_blob` flag to `TestBlobProvider` for testing. - Add `test_load_blobs_overfill_triggers_reset` regression test. Fixes ethereum-optimism#19363 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * chore: use function names instead of line numbers in Go references Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(kona/derive): use named fields for BlobsOverFill variant Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(kona/derive): rename blob_index to filled_blobs in BlobSource (ethereum-optimism#19480) Rename the `blob_index` variable in `load_blobs` to `filled_blobs` for clarity, as the variable tracks the number of blob placeholders that were filled rather than serving as a traditional index. Addresses review feedback from optimism#19364. Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * chore: remove references to Go implementation in comments Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(kona/derive): remove useless BlobProviderError conversion in map_err The `From<BlobProviderError> for PipelineErrorKind` impl already exists, so `?` handles the conversion automatically. The explicit `map_err` was redundant. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: add fix-rust-fmt Claude Code skill Adds a skill that fixes Rust formatting CI failures by running `just fmt-fix` with the correct nightly toolchain via mise. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * Apply suggestion from @ajsutton * Update .claude/skills/fix-rust-fmt/SKILL.md Co-authored-by: Sebastian Stammler <seb@oplabs.co> * docs: link /fix-rust-fmt skill from kona CLAUDE.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs: reference fix-rust-fmt skill in rust-dev.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Sebastian Stammler <seb@oplabs.co>
Bump all reth dependencies from v1.11.2 to v1.11.3 in the rust workspace, including op-reth crate versions and the Cargo.lock.
* chore(cannon): migrate Makefile to justfile Migrate cannon build targets from Make to Just. The Makefile now delegates to just with a deprecation warning, preserving backwards compatibility for existing make invocations. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(cannon): add missing lint target and include justfiles in Docker context - Add `lint` to DEPRECATED_TARGETS and justfile (CI compatibility stub) - Copy justfiles/ into kona cannon-repro.dockerfile for deprecated.mk shim - Install `just` binary in cannon Docker build for the Make shim Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(cannon): remove [default] attribute for just <1.38 compat The Alpine 3.21 just package is v1.37.0 which doesn't support the [default] attribute. Move cannon recipe to first position instead. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(cannon): include justfiles/ in op-program Docker build context The cannon Makefile deprecated shim requires justfiles/deprecated.mk, which is resolved relative to cannon/ inside the Docker container. The op-program Dockerfile.repro.dockerignore was excluding justfiles/ from the build context, causing the cannon make shim to fail. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * docs(cannon): update README to use just instead of make Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(cannon): call just directly in cannon-repro.dockerfile Instead of going through the deprecated Make shim, invoke just cannon directly in the Docker build. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(cannon): update Dockerfile.diff to use just diff-cannon directly The diff-%-cannon Make pattern target was converted to a parameterized just recipe (just diff-cannon VM). Update the Dockerfile to call just directly instead of through make, which would fail since the deprecated shim doesn't support pattern targets. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(cannon): add diff-%-cannon pattern target to deprecated Makefile Preserves backwards compatibility for make diff-<vm>-cannon invocations (used by Dockerfile.diff and potentially other scripts) by translating the pattern to just diff-cannon <vm>. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: install just via system package manager in Dockerfiles - cannon/Dockerfile.diff: use `apk add just` instead of curl install script, drop unnecessary `make` dependency - cannon-repro.dockerfile: switch cannon-build stage from ubuntu:22.04 to golang:1.23.8-alpine3.21, matching the monorepo's Go builder image, so just can be installed via `apk add` instead of curl install script Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(cannon): bump Go to 1.24.10 in cannon-repro.dockerfile The golang Docker image sets GOTOOLCHAIN=local which prevents automatic toolchain downloading. Since go.mod requires go 1.24.0, the 1.23.8 image fails to build. Match Dockerfile.diff which already uses 1.24.10. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…ptimism#19252) * feat: add binary entry point for external proofs in OP (op-rs/op-reth#222) Closes op-rs/op-reth#164 --------- Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com> * feat: add support for `eth_getProof` (op-rs/op-reth#257) Adds support for `eth_getProof` RPC method. This required reworking the launch command to still work with an alternative DB provider. Closes op-rs/op-reth#173 --------- Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com> * perf(trie): feature gate `reth-optimism-trie` metrics (op-rs/op-reth#282) Closes https://github.com/op-rs/bin/issues/281 - Feature gates `reth-optimism-trie` metrics - Moves cursor impls out of proofs module into new module `cursor` - Moves cursor factory impls into new module `cursor_factory` - Updates cursor factory impls to return cursor types with metrics wrapper if metrics feature is enabled * fix(test): Enable live collector tests with metrics feature (op-rs/op-reth#291) Closes https://github.com/op-rs/bin/issues/283 Enable live collector tests when metrics feature is enabled * feat: implement `debug_executePayload` (op-rs/op-reth#276) Closes https://github.com/op-rs/bin/issues/189 --------- Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com> * feat: live collector integration (op-rs/op-reth#306) Closes op-rs/op-reth#296 --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com> * fix: rebase conflicts (op-rs/op-reth#367) Fix conflicts rebasing onto latest upstream main * feat: implemented `OpProofStorage` Database metrics (op-rs/op-reth#407) Closes op-rs/op-reth#224 Closes op-rs/op-reth#387 --------- Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com> * feat: Implemented `OpProofStoragePrunerTask` (op-rs/op-reth#375) Closes op-rs/op-reth#361 Closes op-rs/op-reth#395 --------- Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com> * chore: inmem proof storage removed (op-rs/op-reth#465) This PR removes the CLI support for in-memory proof storage (proofs_history_storage_in_mem). The in-memory storage implementation was primarily intended for unit testing and is not feasible for running a node due to the large memory requirements of storing historical trie nodes. Closes op-rs/op-reth#466 * chore(exex): Add metrics feature in `reth-optimism-exex` (op-rs/op-reth#438) Closes https://github.com/op-rs/bin/issues/427 --------- Co-authored-by: itschaindev <jagrutk@protonmail.com> * feat: prune cli added (op-rs/op-reth#507) Closes op-rs/op-reth#452 * refactor(trie): return `OpProofsStorageError` from `execute_and_store_block_updates` (op-rs/op-reth#535) Closes op-rs/op-reth#523 --------- Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com> * fix: reduce default proofs pruning interval (op-rs/op-reth#560) Closes op-rs/op-reth#559 * feat: add verification interval for integrity check (op-rs/op-reth#577) Closes op-rs/op-reth#449 The approach used is to perform full block verification after every N blocks to ensure the state is still correct. --------- Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com> * chore: getProof benchmark utility added (op-rs/op-reth#550) Utility for op-rs/op-reth#446 * chore: mv proof args to rollup node (op-rs/op-reth#625) Closes op-rs/op-reth#613 * chore: ExEx config builder (op-rs/op-reth#642) Closes op-rs/op-reth#641 --------- Co-authored-by: Himess <semihcvlk53@gmail.com> * chore: moved proof initialization to `reth-optimism-node` (op-rs/op-reth#640) Closes op-rs/op-reth#612 --------- Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com> * chore(rust): fix compilation errors * docs(op-reth): add historical proofs README to exex crate Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * chore(op-reth): move proof-bench to rust/op-reth/bin/proof-bench Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(op-reth): fix go-lint and rust-clippy CI failures - Replace big.Int.Uint64() calls with bigs.Uint64Strict() in proof tests - Fix goimports ordering in proof test files (move ethereum-optimism/* before ethereum/go-ethereum/*) - Fix const alignment in preset.go - Fix uninlined_format_args clippy lint in proof-bench report.rs Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> fix(op-reth): update proofs preset to use stack.ComponentID Replace removed type aliases (L1NetworkID, L1ELNodeID, L2CLNodeID, etc.) with unified stack.ComponentID type. Also fix L2NodeMatcher generics and match.Not calls to use stack.ByID wrapper. fix(op-reth): remove duplicate crates/tests/proofs preset This was a partial copy of rust/op-reth/tests/proofs/utils with no go.mod, no test files, and no other sources. It caused go-lint CI failures due to stale type references. --------- Co-authored-by: jagroot <4516586+itschaindev@users.noreply.github.com> Co-authored-by: Arun Dhyani <dhyaniarun7@gmail.com> Co-authored-by: Julian Meyer <julianmeyer2000@gmail.com> Co-authored-by: Emilia Hane <elsaemiliaevahane@gmail.com> Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Sadiqur Rahman <sadiqurr8@gmail.com> Co-authored-by: itschaindev <jagrutk@protonmail.com> Co-authored-by: Himess <95512809+Himess@users.noreply.github.com> Co-authored-by: Himess <semihcvlk53@gmail.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
…-reth (ethereum-optimism#19408) * feat(kona): implement L2PayloadWitness hint for interop Adds support for the L2PayloadWitness hint in kona interop mode, enabling execution witness pre-fetching via debug_executePayload RPC. Changes: - Add enable_experimental_witness_endpoint flag to InteropHost config - Add ExecutionWitness and OpPayloadAttributes imports - Implement L2PayloadWitness handler with chain_id routing - Extracts chain_id from last 8 bytes (variable-length pattern) - Routes to correct L2 provider via providers.l2(&chain_id) - Calls debug_executePayload RPC with silent failure for unsupported clients - Stores witness preimages (state, codes, keys) with keccak256 keys - Add unit tests for hint data parsing and validation Follows the same pattern as single-chain implementation with multi-chain routing added for interop. All tests pass. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * Fix package reference in interop acceptance tests * feat(op-challenger): add --cannon-kona-experimental-witness-endpoint flag Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * remove redundant witness endpoint acceptance test * fix(kona): add L2PayloadWitness proactive hint and improve interop tests - Register L2PayloadWitness as a proactive hint in interop cfg.rs to match single-chain behavior - Extract parse_l2_payload_witness_hint() helper and update tests to call it directly rather than reimplementing the parsing logic Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(kona): extract preimages iterator before kv lock, add WithL2ELKind tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(op-reth): wire proof-history exex and RPC overrides into main entrypoint main.rs was calling launch_with_debug_capabilities directly, bypassing launch_node_with_proof_history entirely, so the proofs-history exex and eth_getProof/debug_executePayload RPC overrides were never installed. Made launch_node_with_proof_history generic over DB to work with the bare DatabaseEnv the CLI provides. Removed --rpc.eth-proof-window from devstack since the exex override now handles all eth_getProof requests. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Lint fix * fix(op-reth): make supervisor URL optional, disable interop validation when not set Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * chore(rust): apply nightly rustfmt to op-reth changes Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Lint fixes * fix(op-devstack): always enable proof history for op-reth, window=10000 Ports the proof history always-on change to mixed_runtime.go (the new location after develop's devstack refactoring). Removes the ELProofHistory toggle — proof history is now unconditionally enabled for all op-reth nodes with a 10000-block window. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(op-devstack): use op-reth for all interop tests, always enable proofs-history Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(op-devstack): remove supervisor RPC from op-reth node startup Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(op-devstack): use op-reth for supernode interop tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(op-node): handle reorged super authority safe head gracefully Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(op-devstack): skip op-reth tests gracefully when binary not available Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(op-devstack): revert startL2ELNodeWithSupervisor to op-geth * fix(op-challenger): clarify --proofs-history requirement in flag usage * fix(op-reth): fix imports and Cargo.toml after rebase conflict resolution * review: fail on missing op-reth binary, clarify supervisor deprecation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "fix(op-node): handle reorged super authority safe head gracefully" This reverts commit c054e41004b3e9d61850e1e1bf63d246ba4df87d. * fix(kona): log info when debug_executePayload is unavailable, add tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * chore(kona): fix import ordering from nightly rustfmt * Lint fix * fix(kona): make is_rpc_method_not_found const to satisfy clippy * fix(op-devstack): respect DEVSTACK_L2EL_KIND in supernode runtime startMixedOpRethNode was hardcoded so memory-all-opn-op-geth CI job (which sets DEVSTACK_L2EL_KIND=op-geth) was still running op-reth. Add startSupernodeEL helper that reads the env var and falls back to op-reth when not set. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…ributesBuilder (ethereum-optimism#19503) System config update errors in `prepare_payload_attributes` were propagated as critical pipeline errors, halting derivation. One or more receipts may contain malformed or invalid config update logs, which should not be fatal. Change the error handling to log a warning and continue derivation instead. Closes: AQ-10 Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
ethereum-optimism#19549) ReorgTriggered previously checked both that a reorg occurred on an exact block. When a reorg happens earlier than the target block and the target block is not the first in the epoch (e.g. due to an L1 reorg), the parent hash check fails because the reorged block's parent was itself reorged. Introduce ReorgExactFn which preserves the strict parent hash check, and relax ReorgTriggeredFn to only verify the block hash changed. Tests that assert a reorg at a specific divergence point now use ReorgExact; ReorgTriggered is available for cases where the reorg depth may vary. Extract shared reorg-detection logic into a reorgTriggered helper. Fixes TestUnsafeGapFillAfterSafeReorg CI failure where the reorg propagated deeper than the target block, causing a spurious parent hash mismatch. CircleCI: https://app.circleci.com/pipelines/github/ethereum-optimism/optimism/119359/workflows/34010516-d6b0-4a0d-a424-f688235c28c5/jobs/4573838/tests
Add a `deny` target to the Rust justfile that runs `cargo deny --all-features check all`, and update CI to use it instead of invoking cargo deny directly. Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…ism#19553) The large-kona-sequencer config (1 kona sequencer + 4 reth validators + 4 geth validators) runs the same tests as simple-kona-sequencer with no additional test logic — the tests just iterate over more nodes. The only difference is scale, but with 20+ processes on xlarge (8 vCPU, 16GB), kona-node engine API calls intermittently time out under resource pressure, causing TestSyncSafe to flake. The simple-kona-sequencer variant already covers the kona-sequencer + mixed-validator topology with 1 of each type. Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* chore: migrate rust/op-reth Makefiles to just Port all targets from the three Makefiles under rust/op-reth/ to justfiles, following the same deprecation pattern used by Go packages. - rust/op-reth/justfile: expanded with all targets from the Makefile (build, install, cross-compilation, docker, lint, test, profiling) - rust/op-reth/tests/justfile: new file with Go e2e test targets - rust/op-reth/crates/tests/justfile: new file with Go e2e + docker targets - All three Makefiles replaced with deprecation shims that forward to just - CI updated to use `just test-e2e-sysgo` instead of `make test-e2e-sysgo` Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: make FEATURES overridable via env in op-reth justfile Address review feedback: use env() for FEATURES so that docker-build-push-tags can override it in sub-invocations. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: subshell cd in build-release-tarballs, remove invalid deprecated targets - Wrap `cd BIN_DIR` in subshells so cwd doesn't persist across subsequent just/cp commands in the bash shebang block - Remove build-native and build-cross from deprecated targets — these are parameterized just recipes that require a target argument and didn't exist as standalone Make targets (the original had pattern rules build-native-% and build-%) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: add manual shims for build-% and build-native-% pattern rules The generic deprecated-target macro can't handle Make pattern rules. Add manual shims that forward make build-native-<target> to just build-native <target> and make build-<target> to just build-cross <target>, matching the original Makefile behavior. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * Fix default target. * docs: update op-reth READMEs to reflect make-to-just migration Update all three READMEs under rust/op-reth/ to reference `just` recipes instead of `make` targets. Also fix the tests/README.md which described a stale Kurtosis-based workflow that no longer matched the actual Makefile targets — replaced with the current sysgo-based workflow matching the justfile. Remove the stale `make ef-tests` reference from the main README (no such target existed in the Makefile or justfile). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore(ci): capture JUnit XML test results for rust-e2e jobs Wrap bare `go test` invocations with `gotestsum --junitfile` in the kona tests justfile and op-reth tests Makefile so that all rust-e2e CI jobs produce JUnit XML. Add `store_test_results` and `store_artifacts` steps to all 4 jobs in rust-e2e.yml so CircleCI Test Insights can surface per-test timing, flake detection, and failure details. The kona-proof-action-tests job already produced JUnit XML via gotestsum but never uploaded it — this change adds the missing CI upload steps. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: theo <80177219+theochap@users.noreply.github.com>
…#19598) * chore(rust): remove stale RUSTSEC-2026-0002 ignore from deny.toml The lru crate advisory (RUSTSEC-2026-0002) no longer matches any crate in the dependency tree, causing cargo-deny to fail with "advisory was not encountered". The vulnerable lru versions (0.9.0–0.16.2) have been patched — lru 0.16.3 is already in Cargo.lock. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: Update lz4_flex to 0.12.1 lz4_flex 0.12.0 suffers from RUSTSEC-2026-0041 https://rustsec.org/advisories/RUSTSEC-2026-0041 --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: wwared <541936+wwared@users.noreply.github.com>
…ism#19536) * refactor: move OP-specific EVM code into alloy-op-evm newtypes Introduce OpTx and OpTxError newtype wrappers in alloy-op-evm to hold all OP-specific EVM implementations that were previously in alloy-evm's `op` module. This satisfies Rust's orphan rule while decoupling alloy-evm from Optimism-specific dependencies. Changes: - Add OpTx newtype wrapping OpTransaction<TxEnv> with all necessary trait impls (Transaction, TransactionEnv, FromRecoveredTx, FromTxWithEncoded, IntoTxEnv, OpTxEnv, TryIntoTxEnv) - Add OpTxError newtype wrapping OpTransactionError with InvalidTxError and map_op_err helper - Move spec ID mapping and EvmEnv constructors as free functions (evm_env_for_op_block, evm_env_for_op_next_block, evm_env_for_op_payload) - Add rpc module with TryIntoTxEnv<OpTx> for OpTransactionRequest - Update all downstream crates (op-reth, kona, custom-node example) - Remove alloy-evm "op" feature usage from kona crates Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore(rust): move newtypes to op-reth and kona --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat: split gotestsum JSON logs into per-test files Add two scripts under ops/scripts/: - split-test-logs.sh: takes a gotestsum JSON log file and splits it into one log file per test, organized as per-test/<package>/<TestName>.log next to the source JSON file. - gotestsum-split.sh: drop-in gotestsum wrapper that runs gotestsum, then calls split-test-logs.sh. Preserves gotestsum's exit code so splits happen even on test failure. If --jsonfile isn't provided, adds one automatically (tmp/testlogs/log.json) so per-test logs are always generated. All gotestsum invocations in justfiles and CI config now use the wrapper. The one exception is the xargs-based parallel sharding in kona action tests, which calls gotestsum directly (since xargs invokes it multiple times appending to one jsonfile) and then calls split-test-logs.sh once at the end. Per-test logs land under tmp/testlogs/per-test/ which is already captured by existing CircleCI store_artifacts steps. Also migrates rust/op-reth/crates/tests from plain `go test` to gotestsum via the wrapper for consistency. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: fail on split-test-logs.sh errors instead of silently ignoring split-test-logs.sh now exits non-zero when the JSON file is missing or python3 is unavailable. gotestsum-split.sh propagates split failures (preferring the split exit code over gotestsum's if both fail). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: correct script paths in kona tests justfile SOURCE is rust/kona/tests, so the repo root is ../../../ not ../../. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(ci): use literal path for store_artifacts in go-tests job store_artifacts doesn't support glob patterns — `testlogs*.tar.gz` was treated as a literal filename, so no artifacts were ever captured for parallel go-tests jobs. Use a consistent `testlogs.tar.gz` name since the node index is already in the filenames inside the tarball. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ad of silently swallowing them (ethereum-optimism#19604)
ClaytonNorthey92
force-pushed
the
clayton/rust-cherry-pick
branch
from
cdf7603 to
4ad33a4
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.