Skip to content

Merge op-v2.7.0-0 for op-reth v2.9.0 - #103

Merged
maoueh merged 134 commits into
release/optimism-2.xfrom
bump/op-reth-2.9.0
Nov 18, 2026
Merged

maoueh merged 134 commits into
release/optimism-2.xfrom
bump/op-reth-2.9.0

Conversation

@dfuse-bot

@dfuse-bot dfuse-bot commented Nov 16, 2026 •

Copy link
Copy Markdown

Merges ethereum-optimism/reth op-v2.7.0-0 (reth v2.7.0 plus OP commits), the reth pin of upstream op-reth v2.9.0, into release/optimism-2.x. To be tagged op-reth-v2.9.0-fh3.1 once merged.

streamingfast/optimism#32 pins this PR's head commit, so merge this PR first, with a merge commit (not squash or rebase), so that the pinned commit lands on release/optimism-2.x.

Merge

op-v2.7.0-0 does not descend from op-v2.6.0-2 (the OP reth fork re-applies its patches on each upstream release), so the merge conflicts in many files we never changed. Every file the Firehose fork does not touch now matches op-v2.7.0-0 exactly (git diff op-v2.7.0-0 HEAD lists only Firehose files). Firehose conflicts:

  • crates/engine/tree/src/tree/payload_validator.rs: kept mark_verified() after the last fallible step, before upstream's new BAL pairing code. The traced executor execute_and_trace_block now takes an EvmStateProvider, like upstream's execute_block.
  • Cargo.toml / Cargo.lock: took upstream's lock and re-added the Firehose crates. firehose-tracer stays at 5.5.1. The alloy-evm patch is pinned by commit (52afebec, v0.39.0-sf) instead of by tag. alloy-evm stays at 0.39, so streamingfast/evm needs no new release.
  • crates/engine/tree/Cargo.toml: upstream dropped the serde_json dev-dependency that the Firehose engine tests use, so it is added back.

Firehose review

  • Serial tracing order: reth v2.7.0 turns on the sender recovery cache by default. It keeps transaction order. The BAL parallel path only runs for blocks that carry a block access list, which OP blocks never do, and the traced branch is taken before it.
  • Block printed after validation: no new post-execution check. The traced block is still printed only after state root and post-execution validation pass.
  • RPC replays: upstream now replays debug_trace* and trace_* through the chain's block executor (Call::block_executor). That executor comes from executor_for_block, which the Firehose tracer does not hook, so RPC calls never print Firehose blocks. flashbots_validateBuilderSubmission uses execute_one, which is not traced either.
  • Upstream fields: the header, transaction and receipt types used by the mapper are unchanged in this range.

CHANGELOG

## op-reth-v2.9.0-fh3.1

### Changed

- Merged ethereum-optimism/reth `op-v2.7.0-0` (reth v2.7.0 plus OP commits), the reth pin of
  op-reth v2.9.0. The previous pin was `op-v2.6.0-2` (reth v2.6.0). revm moves to 43.0.3;
  alloy-evm stays at 0.39 and the `alloy-evm` patch stays on `streamingfast/evm` `v0.39.0-sf`,
  now pinned by commit. The Firehose engine path follows upstream's new `EvmStateProvider`
  bound; the traced block is still printed only after post-execution validation passes.
- RPC replays (`debug_trace*`, `trace_*`) now run through the chain's block executor upstream.
  They build it with `executor_for_block`, which the Firehose tracer does not hook, so they never
  print Firehose blocks.

### Fixed

- Optimism post-exec transactions (type `0x7d`, activated by the OP Stack Lagoon upgrade) are
  traced as `TRX_TYPE_OPTIMISM_POST_EXEC` instead of stopping the node. `firehose-tracer` moves
  to 5.5.1.

Open questions

  • The op-reth-v2.8.1-fh3.1-1 CHANGELOG section (the post-exec transaction fix from Trace Optimism post-exec transactions #92) was never tagged, and no op-reth release pinned it. Its entry now lives under op-reth-v2.9.0-fh3.1, the first release that ships it. If op-reth-v2.8.1-fh3.1-1 is still meant to be tagged on release/optimism-2.x, restore that section.

Test Execution Summary

Build Status

  • streamingfast/reth cargo check --workspace --all-targets: PASSED
  • streamingfast/optimism rust/: cargo check --workspace --all-targets: PASSED
  • streamingfast/optimism rust/: cargo build --release -p op-reth --bin op-reth: PASSED

Unit Tests

Repository Package Status Notes
streamingfast/reth reth-firehose PASSED 34 tests
streamingfast/reth reth-firehose-tests (prestate golden files) PASSED 2 tests, golden files unchanged
streamingfast/reth reth-engine-tree --test firehose_live_tracing PASSED 1 test
streamingfast/optimism reth-optimism-firehose PASSED no unit tests in the crate
streamingfast/optimism op-alloy-consensus --features firehose PASSED 106 tests
streamingfast/optimism reth-optimism-evm PASSED 58 tests
streamingfast/optimism reth-optimism-payload-builder PASSED 45 tests
streamingfast/optimism reth-optimism-trie (lib, tests/lib.rs, tests/live.rs) PASSED 254 + 86 + 8 tests, with --test-threads=4 (see notes)
streamingfast/optimism reth-optimism-node --test it builder PASSED 2 tests

Notes:

  • reth-optimism-trie with the default 60 test threads failed 88 tests with MDBX Cannot allocate memory (12) when opening their test databases, all at once. With --test-threads=4 all pass, so this is the test machine, not the merge.
  • The build image ships builder-playground v0.3.3-sf1; the OP devnet needs v0.3.3-sf2 (--sequencer-el, --enable-karst), built from source for this run.
  • reth v2.7.0 locks llvm-sys 221 (LLVM 22). The build machine has LLVM 23 only; it builds with LLVM_SYS_221_PREFIX=/usr/lib/llvm-23 and LLVM 23's llvm-config first on PATH.

Integration Tests

  • Status: NOT RUN (no integration suite beyond the tests above for this chain)

Battlefield Tests

  • Status: PASSED
  • Pass/Fail: 95 passing / 0 failing (43 pending: the Amsterdam, ArbOS, Base and other suites that do not apply to an OP Stack chain)
  • Compare blocks (Firehose blocks against the node's JSON-RPC): PASSED
  • Karst suite (activation block upgrade deposits, CLZ, MODEXP, P256VERIFY, BN254 pairing cap, transaction gas cap): PASSED
  • No panic or error in the Firehose op-reth log.
  • Failures: none.
  • Coverage: the devnet runs Jovian from genesis and Karst from L2 block 10, so every fork live on OP Mainnet and Unichain is covered. Operator fee parameters are zero, so the OperatorFeeVault credit is not exercised. Lagoon (not scheduled yet) is not activated by any Battlefield devnet.

Commands (battlefield-ethereum adb1921, builder-playground v0.3.3-sf2 built from source, op-node v1.19.11, fireeth v2.24.1, op-reth from cargo build --release -p op-reth --bin op-reth on this branch):

./scripts/optimism/run_optimism_devnet.sh
./scripts/run_firehose_op_reth_devnet.sh   # op-reth from this branch first on PATH
pnpm test:fh3.0:op-reth-devnet

mattsse and others added 30 commits September 17, 2026 11:49
Co-authored-by: Brian Picciano <933154+mediocregopher@users.noreply.github.com>
Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
)

Co-authored-by: Sergei Shulepov <2205845+pepyakin@users.noreply.github.com>
…27291)

Co-authored-by: Emma Jamieson-Hoare <21029500+emmajam@users.noreply.github.com>
…27297)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…ations on reorg (paradigmxyz#27301)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
Co-authored-by: grandizzy <38490174+grandizzy@users.noreply.github.com>
Co-authored-by: Alexey Shekhirin <github@shekhirin.com>
Co-authored-by: Brian Picciano <me@mediocregopher.com>
…7284)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
…7316)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
…#27312)

Co-authored-by: grandizzy <38490174+grandizzy@users.noreply.github.com>
…27321)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
…digmxyz#27327)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
…digmxyz#27304)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
Co-authored-by: Matthias Seitz <matthias.seitz@outlook.de>
)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
mattsse and others added 26 commits September 28, 2026 12:29
Squash the fork CI support from 04e73b7
and push-event support from b0d194c into
one replayable maintenance commit.

Preserve release action pins and advisory policy when resolving the
replay. Select the book build job without entering Pages deployment.

Pin the Rust toolchains the workflows install: stable to 1.97.0, nightly
to nightly-2026-07-14. A maintenance branch outlives its release by
weeks, and floating toolchains change under it. When replaying onto a new
release, resolve each workflow hunk to upstream's side and re-apply only
these pins, the llvm install in lint.yml's wasm job and the gcc-multilib
`--yes`.

Run cargo deny natively instead of through act (folded in from
c279f31): lint.yml's `deny` job calls
a reusable tempoxyz workflow that starts with tempoxyz's
secure-runner. That action requires a GitHub OIDC token unless the event
is a fork pull request, and act never receives one, so the job failed on
every push and same-repo PR. Run the same `cargo deny --all-features
check all` in a CircleCI job with a digest-pinned Rust image and a
checksum-pinned cargo-deny release, and drop `deny` from the act lint
matrix.

Build the compact reader in a separate target directory so act cannot
reuse libraries compiled from the comparison base after restoring older
source mtimes (folded in from 785b057).

Give the valid_blocks ef-tests group the ten-minute budget already used
by grouped general-state fixtures instead of the single-test sixty-second
timeout, which terminated it on every CircleCI attempt (folded in from
d9bd987).

Compare Compact vectors on pushes to an op-v<version> branch against
upstream tag v<version>, derived from the branch name, instead of a
hard-coded release commit (folded in from 447813d, #21).

Pin the docs build to the nightly the book job installs: the script reads
RUST_NIGHTLY instead of resolving `+nightly` to the newest channel, which
fails to compile boa_engine since 2026-11-01 (rust-lang/rust#165263).
Adapted from upstream 121b0f7. Ignore
RUSTSEC-2026-0391, -0398 and -0403, compression advisories pinned by
reth-zstd-compressors 0.7 whose patched releases need reth-core 0.8 and
Rust 1.96, and RUSTSEC-2026-0387, crypto-bigint 0.5 with no patched
release compatible with elliptic-curve 0.13. All four still apply on
this base (folded in from 22ab931, #30).

Update the lockfile to patched releases for the advisories cargo deny
reports on this base, re-resolved against its lockfile rather than
replayed: rustls 0.23.46, libz-sys 1.1.31, rustls-webpki 0.103.16,
tokio 1.53.3, webpki-root-certs 1.0.10, webpki-roots 1.0.10 and
zstd-sys 2.1.2. No manifest changes.

Co-authored-by: Sebastian Stammler <seb@oplabs.co>
Co-authored-by: Josh Klopfenstein <git@joshklop.com>
Co-authored-by: einar-oplabs <oplabs@einar.io>
Co-authored-by: falcorocks <14293929+falcorocks@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Brian Picciano <933154+mediocregopher@users.noreply.github.com>
(cherry picked from commit 078d026)
)

Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
(cherry picked from commit 42fa3c5)
Co-authored-by: Brian Picciano <933154+mediocregopher@users.noreply.github.com>
Co-authored-by: Brian Picciano <me@mediocregopher.com>
(cherry picked from commit 5e309af)
Co-authored-by: Sebastian Stammler <seb@oplabs.co>
Co-authored-by: Claude <noreply@anthropic.com>
(cherry picked from commit 186af6d)
Pin that `OverlayStateProvider::storage_root` reads trie nodes through
the overlay when partial persistence leaves database storage-trie nodes
older than the persisted hashed storage. OP Stack consensus depends on
this: the Isthmus withdrawals root is the L2ToL1MessagePasser storage
root, computed through `storage_root` when building and validating
blocks.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
(cherry picked from commit d70cc04)
RUSTSEC-2026-0441 marks pin-utils unmaintained with no safe upgrade.
It is reachable only on wasm targets, through alloy-provider's
wasmtimer.
Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
(cherry picked from commit f269bcc)
Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
(cherry picked from commit 0fc98b4)
Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
(cherry picked from commit e29c343)
Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
(cherry picked from commit cdc4e09)
Co-authored-by: Matthias Seitz <19890894+mattsse@users.noreply.github.com>
(cherry picked from commit 569c770)
The docs job ran an unpinned `npx playwright install`, which resolves the
newest `playwright` at run time and broke while npm had 1.65.0 of the CLI
but not of `playwright-core`. Pin it to 1.64.1, as op-v2.6.0 does since
ethereum-optimism#38. Nothing in the docs lockfile depends on a
Playwright version.

(cherry picked from the book.yml half of commit 806fa3b)
…paradigmxyz#28266) (#37)

Backports paradigmxyz#28266 to op-v2.6.0. Every replay of a block's own
transactions, canonical, raw or bad, runs through the chain's block executor
instead of a bare EVM, so traces, replayed state diffs, `debug_accountAt` and
`debug_intermediateRoots` include the per-transaction changes the executor
applies on top of the EVM result. `Call` gains `block_executor` and
`replay_transactions_until_with_executor`; `Trace::inspect_transaction_in_block`
takes the target transaction's index instead of a transaction environment and
rejects an out-of-range index; `TracingCtx` borrows the result and owns the
inspector, as upstream. Raw and bad blocks the executor rejects error instead
of returning pre-refund traces.

Differences from upstream, both forced by alloy-evm 0.39: the inspector is a
type parameter of the EVM, so tracing paths build an inspected executor and
enable the inspector only for the target transaction instead of calling
`inspect_transaction_without_commit`; and there is no `set_block_access_index`,
so the cached-BAL path still executes the target transaction on a bare EVM.

Ports upstream's end-to-end test: a node whose block executor credits a
settlement to every sender shows that prestate diffs, block traces and
intermediate roots agree with committed state.

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
(cherry picked from commit 58c6608)
clippy's type_complexity rejects the return type of Call::block_executor
since v2.7.0 wraps the provider in EvmStateProviderAdapter. Name the
database StateCacheDb reads through instead.
The traced executor takes an EvmStateProvider like upstream's execute_block,
and the engine-tree Firehose tests still need serde_json.
@dfuse-bot
dfuse-bot requested a review from maoueh November 16, 2026 18:19
@maoueh
maoueh merged commit 783aec7 into release/optimism-2.x Nov 18, 2026
2 checks passed
@maoueh
maoueh deleted the bump/op-reth-2.9.0 branch November 18, 2026 20:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.