* metrics: fix ResettingSample Prometheus _count monotonicity (#2174)

* metrics: fix ResettingSample Prometheus _count monotonicity

* use atomic count and remove cumulative sum in resettingSample

* metrics: add Clear method for resetting sample and add more tests

* metrics: avoid cloning slice

---------

Co-authored-by: Manav Darji <manavdarji.india@gmail.com>

* ethapi: include pre-Madhugiri state-sync logs in bor_getLogs (#2155)

* ethapi: include pre-Madhugiri state-sync logs in bor_getLogs

* test: directly validate getBlockAndReceipts state-sync behavior

* security: harden gRPC server — disable by default, bind to localhost (#2078)

* security: harden gRPC server - disable by default, bind localhost

- Add 'Enabled' field to GRPCConfig (default: false), requiring
  explicit --grpc.enabled flag to start the gRPC server
- Change default bind address from 0.0.0.0:3131 to 127.0.0.1:3131
- Guard gRPC server startup with Enabled check in NewServer()
- Add --grpc.enabled CLI flag in flags.go
- Fix GetGrpcAddr() to use net.SplitHostPort instead of string slice
- Update docs and test helper for compatibility

The gRPC server currently starts unconditionally on all interfaces
(0.0.0.0:3131) with no authentication, no TLS, and no way to disable
it. This is inconsistent with HTTP-RPC and WS-RPC which are disabled
by default. An attacker with network access can invoke sensitive RPCs
including ChainSetHead (reorg), PeersAdd/Remove (eclipse attacks),
and StatusBorStatus (reconnaissance) without credentials.

* remove grpc.enabled flag, simplify changes

* internal/cli/server: update default toml for tests

---------

Co-authored-by: harshinsecurity <hi@harshinsecurity.in>
Co-authored-by: Manav Darji <manavdarji.india@gmail.com>

* eth: stop dialing before waiting on handler shutdown (#2158)

* core/state: fix GetWriteMapDump write-set output (#2156)

* metrics, build: fix time.Tick resource leaks and oss-fuzz shebang (#2145)

* metrics, build: fix time.Tick resource leaks and oss-fuzz shebang

Two fixes inherited from upstream go-ethereum:

1. metrics: replace time.Tick() with time.NewTicker() + defer Stop()
   in 6 monitoring goroutines. time.Tick() leaks ticker resources
   (SA1015). Files: debug.go, json.go, log.go, opentsdb.go,
   syslog.go, writer.go.

2. oss-fuzz.sh: fix missing '!' in shebang (#/bin/sh → #!/bin/sh)
   of generated fuzz test runner scripts.

Both fixes have been accepted upstream:
- ethereum/go-ethereum#34044 (shebang, merged)
- ethereum/go-ethereum#34046 (time.Tick, open)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* revert: remove time.Tick changes per maintainer request

Go 1.23+ garbage-collects unreferenced tickers created by time.Tick,
making the NewTicker replacement unnecessary. Revert to stay close to
upstream go-ethereum and avoid merge conflicts.

The oss-fuzz.sh shebang fix is retained as it is independent.

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

* build: bump Go from 1.26.1 to 1.26.2 to fix stdlib CVEs (#2189)

Addresses 6 vulnerabilities in the Go standard library fixed in go1.26.2:
- GO-2026-4947: Unexpected work during chain building in crypto/x509
- GO-2026-4946: Inefficient policy validation in crypto/x509
- GO-2026-4866: Case-sensitive excludedSubtrees auth bypass in crypto/x509
- GO-2026-4870: Unauthenticated TLS 1.3 KeyUpdate DoS in crypto/tls
- GO-2026-4869: Unbounded allocation for old GNU sparse in archive/tar
- GO-2026-4865: JsBraceDepth XSS in html/template

* triedb/pathdb: fix exponential memory growth in preload BFS (#2187)

* triedb/pathdb: replace exponential BFS with trie-aware child decoding in preload

The preload BFS previously called gatherChildPaths which always enqueued
all 16 nibble children for every node regardless of node type. This caused
the BFS queue and visited map to grow as O(16^depth), leading to visited
maps of ~14 GB per goroutine for large storage tries (8 GB caches). On
machines with limited RAM this caused severe swapping that stalled block
import for the duration of the preload (~2 hours for the largest caches).

Replace with decodeChildPaths which decodes the RLP-encoded trie node and
returns only the actual children:
- Branch nodes (17 elements): enqueue slots with non-nil children only
- Extension nodes (2 elements, no terminator): enqueue the single child
  path using the decoded compact key nibbles
- Leaf nodes (2 elements, with terminator): no children

Because MPT child paths are strictly longer than the parent path, the
traversal is cycle-free without a visited set. Remove the visited map
entirely. Queue size is now O(trie width) rather than O(16^depth).

Add compactKeyToNibbles helper and replace the gatherChildPaths test
with comprehensive tests covering all node types, the round-trip compact
encoding, and a cycle-free proof via a 5-node trie where all nodes must
be loaded exactly once.

* triedb/pathdb: address PR review feedback on biased cache preload

- Change nil-child check from exact 0x80 match to len <= 1 for robustness
  against non-canonical RLP encodings (both branch and extension nodes)
- Make decodeChildPaths a package-level function (receiver was unused)
- Pre-allocate nibbles slice in compactKeyToNibbles to avoid reallocations
- Replace time.Sleep with cache.wg.Wait() in two tests for deterministic sync

* triedb/pathdb: fix legacy rate-limit tests to use valid RLP trie nodes

TestAddressBiasedCache_RateLimitInterruption and ShutdownDuringRateLimitWait
were silently passing without exercising their stated scenario: both used raw
non-RLP bytes as the root node, causing decodeChildPaths to return nil
immediately and the rate limiter to never block.

Rewrite both tests to use valid branch/leaf RLP nodes so the BFS actually
traverses children and exhausts the 64KB burst, making WaitN genuinely block
before Close() is called.

* build(deps): bump go.opentelemetry.io/otel/sdk from 1.40.0 to 1.43.0 (#2182)

Bumps [go.opentelemetry.io/otel/sdk](https://github.com/open-telemetry/opentelemetry-go) from 1.40.0 to 1.43.0.
- [Release notes](https://github.com/open-telemetry/opentelemetry-go/releases)
- [Changelog](https://github.com/open-telemetry/opentelemetry-go/blob/main/CHANGELOG.md)
- [Commits](https://github.com/open-telemetry/opentelemetry-go/compare/v1.40.0...v1.43.0)

---
updated-dependencies:
- dependency-name: go.opentelemetry.io/otel/sdk
  dependency-version: 1.43.0
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* eth/fetcher: refactor witness_manager to reduce complexity (#2188)

* eth/fetcher: refactor witness_manager to reduce complexity and close test gaps

Structural refactor of witness_manager.go plus targeted tests, guided by
diffguard (https://github.com/0xPolygon/diffguard). Max cognitive
complexity 44 → 10; Tier 1 / Tier 2 mutation score 100%.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* ci: add diffguard workflow for PR quality metrics

Runs diffguard (https://github.com/0xPolygon/diffguard) on every PR to
report cognitive complexity, function/file sizes, dependency structure,
churn-weighted complexity, and mutation testing at 10% sample rate
against the base branch.

Text report goes to the step summary (visible in the PR check UI) and a
JSON artifact is retained for 14 days for offline triage. Uses the
16-core bor runner to keep mutation turnaround bounded.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* eth/fetcher: fix inverted condition in armTimerChan

The refactor introduced an inverted condition where armTimerChan
returned a nil timer channel when pending requests existed, preventing
tick() from ever firing through the timer path. The retry loop only
appeared to work because rescheduleWitness also pokes pokeCh.

Reported in PR #2188 review. Adds TestWitnessLoopDrivesFetchesForPending
to guard against regressions — it exercises the full loop→tick→
fetchWitness pipeline via real channels, which the existing TestLoop
did not.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* eth/fetcher: fix distance check drift and test nil derefs

Three issues from PR #2188 code review:

- handleNeed's distance check was accidentally tightened from
  `dist < -maxUncleDist` to `<=` during the refactor, introducing a
  boundary mismatch with block_fetcher.go (which uses `<`). A block at
  exactly dist == -maxUncleDist would be accepted by block_fetcher but
  dropped by witness_manager, potentially importing blocks without
  their witnesses. Reverted to match block_fetcher.

- TestWitnessHandleFilterResultSkipsAlreadyPending dereferenced
  `m.pending[hash].op.origin` inside the nil-check branch, panicking on
  the exact mutation it was meant to catch (pending entry removal).
  Now fails cleanly with t.Fatal when entry is nil.

- Same nil deref pattern in TestWitnessCheckCompletingSkipsAlreadyPending.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* all: fix nightly race test flakes (#2193)

The nightly-race workflow had been failing on every run for a month. Most
failures were either race-detector overhead pushing test timings past
narrow assertion margins, tests mutating shared global state that
collided with parallel tests, or a single real data race that cascaded
into many victim tests.

internal/ethapi: testBackendWithPreMadhuguriBorReceipt.ChainConfig() did
a shallow copy of params.AllEthashProtocolChanges, leaving cfg.Bor
aliased to the global BorConfig, then mutated MadhugiriBlock on it.
Concurrent tests calling IsMadhugiri raced with the write. Deep-copy
BorConfig like the sibling testBackendWithNilBorTx already does. This
single race was the root cause of the ~10 cascading ethapi test failures
(TestBorForks, TestBorGetLogs_*, TestCoinbase, TestEstimateGas, etc.).

consensus/bor/heimdall: TestFailover_SwitchOnPrimaryDown and
TestRegistry_MarkUnhealthyOnRealFailure set only getSpanFn on the
primary mock, leaving FetchStatus returning success. The registry's
background probe then raced with the test's MarkUnhealthy call and could
flip the primary back to healthy (or the active gauge back to 0) before
the assertion ran. Make the primary mock fail FetchStatus too so probe
and API are consistent.

core/state: TestConcurrentUsedParallelism measures wall-clock parallel
speedup and asserts >=2x. Race instrumentation serializes atomic/mutex
ops and skews the measurement to ~1.7x. Skip under -race via a new
race_{on,off}_test.go build-tag pair; the test still guards against the
global-lock regression in non-race runs.

core/txpool/legacypool: TestLockOrdering_{PricedHeapNoDeadlock,
ReplacePendingNoDeadlock,RemovedNoDeadlock} used a 10s deadlock-detect
timeout that is genuinely too short once -race instrumentation is added
(the bare test takes ~12s). Bump to 60s — still catches real deadlocks,
no longer fires on legitimate completion.

eth/downloader: TestBeaconSync68/69Full used a 3s sync timeout that
CI routinely exceeded under -race. Bump to 30s.
TestSkeletonSyncRetrievals had four 2s polling loops that could exit
before background serving goroutines finished incrementing the served
counter (the assertion then read a partial count). Bump the budget to
30s, cap per-iteration sleep at 500ms so exponential backoff stays
responsive, and fold the served counter into the polling condition so
we wait for both subchain state and served totals before asserting.

eth/relay: TestCheckTxStatus, TestSubmitPreconfTx, TestSubmitPrivateTx
parallel subtests had handlers sleeping for rpcTimeout-100ms, leaving
only 100ms of slack before the client-side 2s timeout fired; under
-race that slack was easily exhausted. Halve the handler sleep to
rpcTimeout/2 — still proves parallelism (3 serial calls would exceed
the unchanged 2s upper bound) with comfortable margin for -race.

metrics: TestExpDecaySampleNanosecondRegression relied on the unseeded
global RNG and wall-clock timing, making the reservoir's average a
noisy statistic that occasionally drifted outside [14, 16] under
-race. Drive the test with a seeded RNG and synthesised monotonic
timestamps; override t0/t1 after NewExpDecaySample so dt in update()
stays positive. Test still catches the priority-overflow regression
(average would stick at 10 under the buggy formula).

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* Merge pull request #2191 from 0xPolygon/psp-enable-wit-fs

internal/cli: enabled witness filestore by default

* eth: cancel milestone subscriber on shutdown (#2199)

* Merge pull request #2175 from 0xPolygon/manav/disable-pending-loop

(feat): disable pending block creation loop via flag

* rawdb: recover witness size from filesystem metadata miss (#2200)

* ci: fix kurtosis setup/cleanup action paths (#2223)

kurtosis-pos moved the setup and cleanup composite actions under
.github/actions/kurtosis/, so the bor workflows fail with "Can't find
action.yml" at the Pre/Post kurtosis run steps after checkout.

Update both kurtosis-e2e.yml and kurtosis-stateless-e2e.yml to use
./.github/actions/kurtosis/setup and ./.github/actions/kurtosis/cleanup.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* Fixing amoy static peers (#2209)

* core/types: guard (*Header).GetValidatorBytes against short Extra (#2222)

* core/types: guard (*Header).GetValidatorBytes against short Extra
(*Header).GetValidatorBytes panics with 'slice bounds out of range'
when len(h.Extra) < ExtraVanityLength+ExtraSealLength on pre-Cancun
chain configs. The post-Cancun branch and companion helpers already
guard this path; hoist the check to the top so both branches are safe.
Add TestGetValidatorBytesShortExtra as regression test.
Discovered with Zorya (concolic execution).
Fixes #2221

* Potential fix for pull request finding

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

* Potential fix for pull request finding

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

* chore: pos workspace setup (#2218)

* miner, consensus: more block building metrics (#2228)

* consensus/bor, miner: fix stale pending state for non-validator nodes (#2183)

* consensus/bor, miner: fix stale pending state for non-validator nodes

Nodes whose signer is configured but not in the active validator set
(e.g. after a validator rotation) returned "insufficient funds for
transfer" on eth_estimateGas and other RPC calls against the "pending"
block. The root cause was that Prepare() returned UnauthorizedSignerError
when the signer was not in the validator set, which caused prepareWork()
to fail on every block. Since commit() was never called, the pending
snapshot was never refreshed, and its underlying pathdb trie layers were
eventually garbage-collected — making all state reads return zero.

Fix Prepare() to not fail for unauthorized signers. Instead, default the
succession number to 0 and let the block building pipeline continue
normally. Seal() already independently rejects blocks from unauthorized
signers, so no invalid blocks can be produced. This ensures the pending
snapshot is refreshed on every chain head, keeping RPC queries functional
regardless of whether the node is actively producing blocks.

Also remove a redundant second call to GetSignerSuccessionNumber in the
Giugliano waitOnPrepare block, reusing the succession value already
computed earlier in Prepare().

* fix linter

* Merge pull request #2229 from 0xPolygon/psp-pos-3598

docs, .claude: add hardfork rollout review guidance

* build: bump Go to 1.26.3 (#2230)

* consensus/bor, internal/cli: full grpc implementation (#2194)

* implement full grpc comms with heimdall

* address comments

* address comments

* address comments

* address comments

* address comments

* revert tracers/data.csv

* internal/cli, docs: skip gRPC on empty addr, warn unauth non-loopback

* internal/cli: address comments

* internal/cli: address comment

* internal/cli: implement additional tests

* internal/cli: address comments and improve tests/validation

* internal/cli: address comments and add new tests

* internal/cli/server: map reorg & non-contiguous range sentinels; tidy cmd/keeper go.sum

mapBorAPIError now translates the two GetRootHash sentinels that previously
fell through to codes.Internal:
  - "reorg occurred while computing checkpoint root" -> codes.Aborted
    (transient/retriable; clients can retry once the chain settles)
  - "non-contiguous headers in checkpoint range" -> codes.DataLoss
    (DB inconsistency / ancient pruning; conveys unrecoverable server state
    without the "this is a bug" implication of Internal)

Also runs go mod tidy in cmd/keeper to add the missing
golang.org/x/sys v0.42.0 h1: source hash that the previous bump left out.

* internal/cli: add maxGRPCMessageSize to mirror HTTP behaviour

* internal/cli: add nil check on grpc addr. Bump go to fix vulncheck

* chore: fix govuln

* internal/cli: lower chunkSize

* chore: bump polyproto

* templates: replace dead witnessprotocol with witness.enable hint (#2204)

* templates: replace dead witnessprotocol with witness.enable hint

* ci: stabilize codecov/diffguard and skip fork-only e2e constraints

* templates: move witness hint and drop workflow edits

* prefetcher: builder-phase prefetch + streaming worker pool (#2192)

* prefetcher: streaming worker pool with unified idle/builder tx providers

Replace per-call Prefetch() invocations with a long-running PrefetchStream
that spans the block lifetime. A fixed worker pool pulls transactions from
a channel, eliminating pool-startup overhead across the ~500 builder-mode
and ~20 idle-mode calls previously issued per block.

Idle and builder phases share the same stream, acting as swappable tx
providers. Phase handoff uses a two-signal interrupt: hardKill for
permanent exit and evmAbort for soft, per-phase aborts that discard
in-flight idle work before the builder provider takes over.

Prefetch(block, ...) keeps its original signature as a thin wrapper so
blockchain.go behavior stays byte-identical.

* miner: prefetch observability — slow-tx prefetch flag, split apply timers, builder contribution metric

Correlate tail latency with prefetch miss rate so we can tell whether the
slowest block txs are the ones we failed to prefetch or whether prefetch
quality is not the lever.

Three additions:

  * slow_tx_tracker 10-min log: each of the top-K slowest txs is now
    annotated inline with MGas/s and a prefetched/not-prefetched flag.
    MGas/s uses integer math on nanoseconds — per-tx durations are in
    the tens of microseconds and float seconds lose precision.
  * worker/txApplyDuration/{prefetched,notPrefetched}: side-by-side p50/
    p95/p99 of the two populations. The aggregate worker/txApplyDuration
    stays for dashboard compatibility.
  * worker/prefetch/builder_added_percent: share of block txs that the
    builder phase prefetched, isolating the payoff of the new builder-
    phase prefetch beyond the existing overall miss-rate metric.

* prefetcher: grow stream channel buffer from 512 to 4096

Sizes the buffer to the idle provider's theoretical per-loop burst
(IdleGasBudget / 21k ≈ 4762 at default config) so the scan never
blocks on a full channel. Workers drain at ~200k tx/s so the channel
sits mostly empty in practice; the buffer exists to absorb the brief
scan-phase burst where idle dumps a heap's worth of txs.

Memory cost: 4096 * 8 B + channel header ≈ 33 KB, negligible.
Handoff drain is O(actual-occupancy), not capacity, so a larger
buffer doesn't slow phase transitions.

* prefetcher: gate builder-phase attribution on post-handoff flag + goimports fixes

Review feedback: the old onSuccess closure checked genParams.builderStarted.Load()
to decide whether to attribute a successful prefetch to the builder phase. That
check races with buildAndCommitBlock's builderStarted.Store(true), which fires
before runPrefetcher reaches the handoff — an idle-phase tx whose EVM execution
finishes in the gap between those two moments would be miscounted as builder.
Impact was metric-only but the old comment overstated the guarantee.

Route attribution through a dedicated inBuilderPhase *atomic.Bool that the
coordinator flips to true only after the handoff completes (evmAbort drain +
reset). Any onSuccess firing after that point is known to come from
post-handoff work, so builder_added_percent now reflects genuine builder-phase
contribution.

Also fix goimports formatting in core/state_prefetcher.go and miner/worker.go
flagged by CI lint.

* prefetcher: stop idle batch on full channel instead of spinning the heap

Review feedback: when the stream channel fills mid-batch, the default
branch in streamIdleBatch was dropping the tx but still calling
txs.Shift() and continuing to walk the heap. Since dropped sends don't
subtract from the gas budget or populate localPrefetched, the outer
loop's viability check (nextViableIdleTx) kept returning valid txs,
so every remaining entry burned Peek + Shift cycles only to drop.

Return on the default branch instead. The outer runIdleTxProvider loop
re-snapshots the pool every ~100ms, by which time workers have drained
the channel and there's capacity for new sends. Dropping one tx per
batch and re-entering the outer loop is strictly cheaper than walking
the rest of the heap dropping everything.

* prefetcher: nil-guard prefetchedHashes + stop charging gas for failed resolves

Review feedback on buildTxPlan / scanOverflow / collectPlanBatch:

* buildTxPlan decremented remaining before calling ltx.Resolve(), so when a
  LazyTransaction failed to resolve (tx evicted between heap listing and
  resolution) the budget was consumed for a tx that never entered the plan.
  Reorder: charge gas only after Resolve succeeds, matching scanOverflow.
  The prefetched-skip path keeps consuming gas intentionally — those txs
  are still bound for the block.
* scanOverflow and collectPlanBatch called prefetchedHashes.Load() without
  a nil guard, while the sibling buildTxPlan explicitly guarded. Production
  is safe because commitWork always initializes the map, but the API
  inconsistency is a real footgun for unit tests or future callers.
  Add the missing guards so all three functions accept a nil *sync.Map as
  "no hashes known, don't skip".

No functional change in the current call paths; tests unchanged.

* prefetcher: recover panics in worker and sendPlan goroutines

Review feedback: worker goroutines in PrefetchStream ran without their own
recover. Go's recover only catches panics in its own goroutine — the parent
runPrefetcher's recover does not protect the children. Any panic in
statedb.Copy / ApplyMessage / IntermediateRoot / precompile edge cases
would therefore kill the node even though prefetching runs on a throwaway
state copy and is best-effort by definition.

Add defer/recover per worker (new chain/prefetch/worker/panic meter) and
match it on the sendPlan goroutine in fillTransactions (reuses the existing
worker/prefetch/panic meter).

Also documents the residual attribution edge where a worker stuck in
IntermediateRoot (not interruptible by evmAbort) can still reach onSuccess
after inBuilderPhase flips. Metric-only impact, bounded by at most one tx
per block; added a comment instead of a runtime fix since the window is
tiny in practice and closing it would require a more invasive barrier.

* prefetcher: fix gas budget, interrupt accounting, panic leak, dup execution

Round 4 of review fixes — four independent issues:

* sendPlan gasLimit stale for the second call. The closure captured
  env.header.GasLimit once, but the second invocation (normal-tier txs)
  runs after commitTransactions has already consumed gas from env.gasPool.
  Plan was over-sized, wasting prefetch capacity on txs that the builder
  would reject. Thread gasLimit as an explicit argument and pass
  env.gasPool.Gas() on the second call (with a nil guard for the first,
  where env.gasPool is still nil).

* prefetchOneTx early-interrupt return skipped fails.Add(1). txIndex is
  incremented unconditionally in processTx, so every interrupt-aborted tx
  was counted as a successful prefetch in blockPrefetchTxsValidMeter. The
  function docstring already promises fails is bumped on every (0,false)
  return; the interrupt path was the one branch that didn't.

* runPrefetcher shutdown (evmAbort + close(txsCh) + <-streamDone) was
  sequential, so a panic in runIdleTxProvider or runBuilderTxProvider
  unwound past it. The PrefetchStream goroutine and its N workers would
  then block forever on `range txsCh` (hardKill is only checked after a
  dequeue). Wrap the close in a sync.Once-guarded closure and defer it
  so the channel is released on every exit path, panics included.

* Builder phase could double-execute an in-flight plan tx. scanOverflow
  checked prefetchedHashes, which is only written by onSuccess after EVM
  completion — a tx forwarded in one batch but still mid-EVM when the next
  batch's overflow scan ran could be re-emitted and picked up by a second
  worker. Track locally-sent hashes in runBuilderTxProvider (single-
  threaded, plain map is fine) and have scanOverflow consult it alongside
  prefetchedHashes.

* prefetcher: address within-iter dup, heap preservation, panic fails

- scanOverflow: break instead of Pop on gas-too-large so accumulated
  extendedBudget can cover price-leading accounts on later iterations;
  popping permanently evicted them from the long-lived overflow heap.
- runBuilderTxProvider: populate sentThisPhase with the current plan
  batch before scanOverflow so bonus collection can't re-emit a tx
  already returned by collectPlanBatch in the same window.
- PrefetchStream worker recover: increment ctx.fails so the
  {valid,invalid} meters stay consistent with txIndex when a panic
  aborts prefetchOneTx mid-flight.

* miner: unit tests for prefetch stream primitives

Covers the helpers that prior review rounds kept surfacing bugs in:

- scanOverflow: zero budget, budget growth preserves popped accounts
  (regression test for the h.Pop() → break fix), in-flight skip
  without budget cost, prefetched skip without budget cost.
- forwardTxs: sent-hash recording, dropped-on-full channel must not
  record, nil sentThisPhase is safe.
- collectPlanBatch: closed planCh, timer-fires-on-empty, freed-gas
  accumulation, prefetched filter, closed gas channel nilled out.
- runBuilderTxProvider: no-duplicate-forward invariant across
  interleaved planCh + freed-gas signals — regression test for the
  within-iteration dedup fix (sentThisPhase pre-population).

* prefetcher: signal interrupt on prepareWork early-return

Move `defer interruptPrefetch.Store(true)` to the top of
buildAndCommitBlock so the prefetcher goroutine is signaled to stop
on every return path — including the early return when prepareWork
fails (UnauthorizedSignerError, Heimdall snapshot/span failures,
stateless.NewWitness errors). Previously the defer sat after the
err check, so on prepareWork failure runIdleTxProvider would loop
until gas exhaustion, burning ~100M gas of throwaway EVM work
across the worker pool before self-terminating.

No correctness/consensus impact; fixes a real CPU-waste path on
an uncommon-but-realistic error branch. Bounded by PrefetchStream's
sync.Once shutdown (no goroutine leak), but the wasted work was
material under prolonged Heimdall issues when consecutive
commitWork calls could pile up zombie prefetchers.

* prefetcher: close scanOverflow→plan cross-iter dedup edge

Pass sentThisPhase into collectPlanBatch and skip txs already present,
mirroring the existing scanOverflow dedup pattern. Closes the third
edge of the builder-phase dedup matrix:

1. within-iter plan↔overflow — covered by pre-populate sentThisPhase
   before scanOverflow (commit 62a72c0)
2. plan→overflow cross-iter — covered by scanOverflow's sentThisPhase
   read (commit 88c4e44)
3. overflow→plan cross-iter — this commit

Scenario: scanOverflow emits tx T in iteration N (h.Shift() past T,
sentThisPhase[T]=true). Worker W1 begins multi-ms EVM on T. In
iteration N+1, collectPlanBatch reads a buffered copy of T from
planCh; prefetchedHashes is still empty because onSuccess hasn't
fired, so T slips through and gets forwarded a second time.

Impact is wasted worker capacity only (throwaway state, no
consensus/correctness effect), but the race opens wider exactly on
contract-heavy workloads where builder-phase prefetch matters most.
Kurtosis run #2 scenario E didn't catch this because value-transfer
EVM finishes in ~10µs, well inside the 2ms batch window — so
prefetchedHashes raced fast enough to plug the gap.

Added TestCollectPlanBatch_SkipsInflight as the regression test.

* prefetcher: address review feedback (constants, naming, sendPlan)

- hoist prefetch tunables (chan buf, idle loop interval, gas-pct
  default/cap) into the file's top const block
- rename sentThisPhase to inFlightHashes — the variable tracks txs
  forwarded on txsCh whose onSuccess hasn't fired yet
- promote sendPlan from a closure inside fillTransactions to a free
  function alongside buildTxPlan

* prefetcher: disable IntermediateRoot per tx; add evaluation benchmarks

The streaming prefetcher's per-tx IntermediateRoot call (introduced via the
intermediateRootPrefetch flag, hardcoded true) was found to add 80–130%
prefetch wall time for at most ~10% commit speedup (≈0.1 ms in absolute
terms). With snapshots active, EVM reads bypass the trie entirely, so the
only path to value is warming pebble's block cache for the subsequent
Commit. Under realistic clean-cache sizes that working set is already
resident, leaving the warming work redundant in nearly every regime.

Upstream go-ethereum's prefetcher likewise does not compute intermediate
roots — its Prefetch is (block, statedb, cfg, interrupt) only.

Flips the call site to false (parameter retained for API stability) and
checks in three Go benchmarks documenting the methodology and findings:
  - TestIntermediateRootPrefetch_AccuracyVsCost: in-memory baseline,
    three access patterns (hot 3-slot, unique-spread, per-sender counter)
  - TestIntermediateRootPrefetch_PebbleAccuracyVsCost: real pebble disk
    with prewarm + Commit, hashdb meter deltas
  - TestIntermediateRootPrefetch_HeavyColdContract: 100k pre-populated
    storage slots with a constrained clean cache, modelling fat contracts
    that haven't been touched recently

Across all three, flag=true produced zero cache-hit benefit measurable in
StorageHitFromPrefetch / clean cache deltas, with consistent prefetch CPU
overhead. The residual ~10% commit speedup observed in the heavy-cold
regime is attributable to pebble block-cache warming and amounts to
≈0.13 ms per affected block.

* prefetcher: kill diffguard mutation survivors

Extract resolveEvmInterrupt helper so the evmAbort/hardKill fallback
wiring is unit-testable. Add focused unit tests for preloadReaderForTx
(bad-signature, contract-To code preload, EOA/nil-To) and
streamIdleBatch (per-loop gas budget). Brings T1 logic and T2 semantic
mutation kill rate to 100% on the diff.

* prefetcher: fix lint (unconvert + trailing newline)

Drops a redundant uint64() conversion in the streamIdleBatch test and
trims the trailing blank line goimports flagged.

* prefetcher: nil-guard genParams in sendPlan

The function nil-checked genParams before reading prefetchedTxHashes but
then unconditionally dereferenced genParams.planWg.Add(1). Currently
unreachable — callers in fillTransactions derive builderPlanCh from
genParams.builderPlanCh so genParams is always non-nil — but the
inconsistent guard is misleading and would panic on the block-production
path if a future caller drops that invariant.

* core/vm: add hardfork precompile continuity checks (#2234)

* core/vm: add hardfork precompile continuity checks

* docs: clarify Pro precompile continuity guidance

* packaging/docs: replace deprecated rpc.batchlimit key (#2216)

* packaging/docs: replace deprecated rpc.batchlimit key

* packaging/docs: set rpc.batch-request-limit default to 1000

* p2p: add flag to disable snap/1 serving (#2152)

* flag to disable snap serving

* p2p: add clarifying comment on NoSnapServing behaviour

Document that NoSnapServing=true disables snap/1 entirely (no serving
and no snap-syncing from peers), while the in-memory snapshot tree
remains active for local state reads.

* miner, consensus/bor: fix leaked-wedge family in worker state machine (#2220)

* miner, consensus/bor: fix leaked-wedge family in worker state machine

Closes four silent stall paths in the producer state machine

1. miner.mainLoop: when PeerCount==0 on production chains, the dropped
   newWorkReq branch now clears pendingWorkBlock instead of leaking it.
2. miner.commitWork: the defer that clears pendingWorkBlock is now
   registered above the early syncing-check return.
3. miner.taskLoop: interrupt() now deletes the previous sealhash's
   pendingTasks entry. Bor.Seal's stop-branch returns silently without
   posting to resultCh, so resultLoop never cleans the entry.
4. consensus/bor.Seal: the result-delivery goroutine's second select
   no longer has a silent default branch — it blocks on send or exits
   via <-stop.

Regression tests cover each path (four unit + one integration):
- TestMainLoopClearsPendingWorkBlockOnPeerCountZero
- TestCommitWorkLeaksPendingWorkBlockWhenSyncing
- TestTaskLoopInterruptCleansStalePendingTasks
- TestSeal_BlocksOnFullResultChannelInsteadOfSilentDrop
- TestProducerRecoversAfterMiningRestart (tests/bor, integration tag)

* miner: gate PeerCount check on heimdall presence, not chain-ID

The previous chain-ID gate (BorMainnet/Mumbai/Amoy) blocked the
PeerCount==0 drop from being exercised by kurtosis chaos tests on
chain ID 4927. Replacing it with chain-ID list omissions broke
dev/test setups (Clique single-node, intentional-disconnection
tests like TestValidatorWentOffline) where peer count is
legitimately zero by design.

Gate on heimdall presence instead. The HeimdallClient is nil on
all test/dev setups (--bor.withoutheimdall, Clique, Ethash) and
non-nil for any real-network Bor node, so the semantic 'this is
a real network node' is captured directly. Forward-compatible:
new production chain IDs don't need to be added to a list.

Also restore the !IsRio removal in the veblop fallback: the
post-Rio kurtosis bootstrap needs the periodic retry trigger
after the PeerCount drop. Gate the fallback on !isBor instead so
Clique/Ethash don't accidentally fire it.

Fixes CI failures introduced by the prior commit:
- TestServer_DeveloperMode (internal/cli/server)
- TestCommand_DebugBlock   (internal/cli)
- TestValidatorWentOffline (tests/bor, integration)

* miner: fix data race on pendingTasks length in stall warning

Caller in newWorkLoop's veblopTimer case captured len(w.pendingTasks)
under pendingMu, but the helper warnIfStalled read len(w.pendingTasks)
again unguarded in its log statement. Concurrent taskLoop /
resultLoop mutations would race with the read.

Capture the count once under the lock and pass it through to the
helper instead.

Also drop the redundant DevFakeAuthor=false write from
TestMainLoopClearsPendingWorkBlockOnPeerCountZero — NewFakeBor
already constructs with DevFakeAuthor=false, and writing after
mainLoop has captured the value at startup is itself a data race
that -race flagged.

Reported in PR #2220 review.

* miner, tests/bor: update regression-test docstrings to match heimdall-gate

The PeerCount==0 fix was originally chain-ID-gated (BorMainnet/Mumbai/Amoy)
and the test docstrings reflected that. After review the gate was changed
to heimdall-presence (`bor.HeimdallClient != nil`), but the docstrings
were not updated.

This commit corrects the inaccuracies:
- TestMainLoopClearsPendingWorkBlockOnPeerCountZero now correctly
  describes the heimdall-presence trigger and notes that the test uses
  the mock heimdall client (non-nil) from NewFakeBor — no ChainID
  override exists.
- TestProducerRecoversAfterMiningRestart now correctly explains that
  Bug 1's drop branch isn't exercised because the test uses
  withoutHeimdall=true (HeimdallClient nil), not because of chain-ID.

* miner, consensus/bor: scope pendingTasks cleanup to Seal stop-branch exits

The previous taskLoop fix made interrupt() unconditionally call
deletePendingTask(prev), which races with Bor.Seal's success path:
when the goroutine has already delivered the result to resultCh but
resultLoop is busy, interrupt() deletes the entry resultLoop is about
to look up, and the !exist branch silently drops the validly-sealed
block.

Move cleanup into Bor.Seal itself, gated to stop-branch exits, via a
new SealWithStopHook(..., onStopExit func()) method. The existing
Seal becomes a thin wrapper passing nil to preserve consensus.Engine.

taskLoop type-asserts to *bor.Bor and uses SealWithStopHook with a
per-sealhash cleanup closure. interrupt() now only closes stopCh.

Adds three bor-level tests covering both stop-branch exits and the
success path, and rewrites the worker-level test to assert the new
contract: interrupt() must NOT delete pendingTasks.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* miner, tests/bor: update docstrings for SealWithStopHook cleanup direction

Two docstrings still described the pre-revision design where
interrupt() deleted pendingTasks directly:

- tests/bor/bor_test.go bullet #3 named a unit test
  (TestTaskLoopInterruptCleansStalePendingTasks) that no longer
  exists; the actual test is TestTaskLoopInterruptPreservesPendingTasks
  asserting the opposite semantic.
- miner/worker.go deletePendingTask docstring said "Used by
  taskLoop.interrupt" but the caller is the per-task onStopExit
  closure passed to Bor.SealWithStopHook.

Documentation-only.

* miner: drop PeerCount==0 gate in mainLoop (#2238)

* miner: drop PeerCount==0 gate in mainLoop

The peer-count drop branch in `mainLoop` was introduced in #977 to
delay block production until at least one peer was connected, scoped
to BorMainnet/Mumbai (Amoy was added later in e8157bbe6). #2220
widened the gate from a chain-ID list to "any heimdall-connected
node", which drew kurtosis devnets into the gate for the first time:
in a single-validator kurtosis devnet PeerCount() is 0 by design (no
other node to peer with), the gate drops every newWorkReq, and the
veblop fallback timer re-fires straight back into the same drop.

Veblop's primary/backup producer election already handles outage
safety at the protocol level — peer count adds no safety, only a
foot-gun. Strip the gate; `mainLoop` now routes every newWorkReq
directly to `commitWork` (except the pre-existing
DisablePendingBlock-on-non-validator short-circuit).

Verified end-to-end on a kurtosis 1-validator devnet (chain ID 4927,
veblop active).

* miner: drop stale PeerCount==0 gate references in test backends

Follow-up to the gate removal. With no test exercising peer count
anymore, the `testWorkerBackend.peerCount` atomic.Int32 plumbing in
`worker_test.go` was just ceremony — drop the field, the `Store(1)`
default, and simplify `PeerCount()` to return a constant. The
`mockBackend` / `mockBackendBor` PeerCount comments in `miner_test.go`
and `fake_miner.go` still named the removed gate; rewrite them to
describe what the function does rather than what it no longer guards.

No production change.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* miner: drop dangling incident references in syncing-leak test docstring

The doc comment and t.Fatalf message for
TestCommitWorkLeaksPendingWorkBlockWhenSyncing framed it against the
"PeerCount==0 race fixed for the 2026-05-07 incident" and the related
val4 stall — context that disappeared from the tree with the gate
removal and the sibling regression test in this branch's earlier
commits. Rewrite the docstring to describe the leak on its own terms
and trim the t.Fatalf tail so a future reader hitting the failure
doesn't grep for a name that no longer exists in-tree.

No behavior change.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* tests/bor: drop dangling gate references in producer-recovery docstring

TestProducerRecoversAfterMiningRestart's docstring framed itself
against four leak paths; item (1) named the deleted unit test
TestMainLoopClearsPendingWorkBlockOnPeerCountZero and the deleted
gate (`realNetworkNode := bor.HeimdallClient != nil`). The
integration-level limitations clause also leaned on Bug 1 / that
gate.

Drop both. Renumber the remaining three leak paths and add a short
historical note that the fourth was closed by removing the gate
itself in earlier commits on this branch.

No behavior change.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* tests/bor: drop dangling Bug 1 references in producer-recovery test

After the previous commit renumbered the four-bug list down to three,
three remaining "Bug 1" mentions in the docstring header and the test
body's Phase 2 / Phase 3 comments became dangling (the enumeration no
longer has a Bug 1 entry).

Reword each to point at the generic "leak family" instead. The test's
behavior is unchanged.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* miner: format build-block log times as RFC3339Nano, gate on IsRunning (#2239)

* consensus/bor: fix race in SpanStore.PurgeCache (#2235)

* consensus/bor: fix race in SpanStore.PurgeCache

PurgeCache cleared latestSpanCache via atomic Store(nil) but did not stop
the background polling goroutine started by NewSpanStore. That goroutine
ticks every 200ms and writes the latest span back into latestSpanCache,
silently undoing the purge whenever a tick lands between the clear and
the caller's next read.

Fix: extract the loop into runPollLoop, track it with a WaitGroup, and
have PurgeCache stop and join the goroutine before resetting state.
Close uses the same path. PurgeCache no longer restarts the loop —
on-demand reads via getLatestSpan fall back to updateLatestSpan inline,
so callers still get fresh data without the race window.

Adds TestSpanStore_PurgeCache_RaceWithPollLoop, a deterministic
reproducer that sleeps past one tick before asserting. Fails reliably on
develop, passes with the fix.

* consensus/bor: clear heimdallStatus in PurgeCache

With the poll loop stopped after PurgeCache, a stale heimdallStatus
(typically CatchingUp:false) would persist and let
waitUntilHeimdallIsSynced return immediately without refreshing against
a freshly-swapped heimdall client. Clear it alongside the other
atomics, and extend the reproducer to assert the invariant.

* build: bump x/net for govulncheck (#2240)

* core/forkid: include polygon-specific forks in wire forkid (#2237)

* core/forkid: include polygon-specific forks in wire forkid

NewID computes the eth p2p handshake forkid by calling gatherForks
(lowercase), which uses reflection over the outer ChainConfig struct
and is blind to the polygon-specific block fields nested under
config.Bor (RioBlock, MadhugiriBlock, DandeliBlock, LisovoBlock,
LisovoProBlock, GiuglianoBlock, ChicagoBlock). The polygon-aware
GatherForks (uppercase) was added in #2063 to expose fork data via
the new bor_forks RPC, but the wire path was never updated to use it.

As a result, bor's wire forkid does not include any polygon-specific
fork activation in its checksum. On chains where polygon forks
activate at non-zero blocks (devnets, testnets, mainnet post-fork),
bor's wire forkid is inconsistent with what the bor_forks RPC reports
for the same node.

This was latent until erigon v3.6.0 started including the new
polygon forks (Lisovo, LisovoPro, Giugliano, Chicago) in its own
GatherForks. erigon's wire forkid now correctly hashes those blocks
while bor's still does not, causing the eth handshake to fail with
'fork ID rejected: local incompatible or needs update' on any
deployment combining bor v2.8.0 with erigon v3.6.0 on a chain where
those forks activate above block 0.

Switch NewID to call GatherForks so the wire forkid includes the
polygon-specific forks and matches what bor_forks RPC reports.

* core/forkid: include polygon-specific forks in inbound filter

newFilter (used by NewFilter / NewStaticFilter) builds the local table
of valid fork checksums that the inbound eth handshake validator
matches incoming peers against. It was calling lowercase gatherForks,
which is the same polygon-blind reflection-only path the previous
commit fixed in NewID.

Without this, a patched bor (with the fixed NewID) would emit a forkid
that includes the polygon-specific forks, but its own inbound filter
would compute the local sums[] table without them. Incoming peer
forkids would never match the table by exact rule #1 of EIP-2124 and
would only be accepted via the rule #3 "remote is a superset"
forgiveness path. The validator's accept/reject behaviour would be
correct by accident rather than by design, and any nuance that depends
on rule #1 (matching local fork state) would be wrong.

Switch newFilter to GatherForks so the validator's sums[] table is
symmetric with what NewID emits. After this, the eth handshake's
forkid check is end-to-end polygon-aware on bor.

* (feat): tracing support for state-sync transactions (#2236)

* eth/tracers: add state-sync tracing support

Add live tracing and refactor tracing methods for state-sync transactions.

Co-authored-by: Milan Došlić <milan.doslic@tenderly.co>

* core: call OnTxEnd hook only if state-sync is added

* consensus/bor: remove duplicate import

* internal/cli: use stream context

* consensus/bor: use deterministic time for state-sync traces

* eth/tracers: remove TODO

* consensus/bor: fix build

* eth/tracers: make cumulative gas used semantics same for tracing

* core/types: removed unused system address

* eth/tracers: sort statedb logs

* eth/tracers: fix comment

* eth/tracers: add transaction index to state-sync reciept

* address claude comments

* eth: move tx index check

* core/types: use bor system address as sender address of state-sync transactions

* eth/tracers: add state-sync custom tracer to handle invariants

* core, eth, consensus: fix live state-sync tracing

* eth/tracers: add state sync wrapped hooks test

* eth, consensus: address claude comments

* tests/bor: add e2e tests

* eth/tracers: fix lint

* core/types: undo sender change

* consensus/bor: remove comments

* eth, consensus: apply some simplifications

* consensus/bor: apply diffguard changes

* eth, core, consensus: increase code coverage

* eth/tracers: fix prestate tracer for state-sync transactions

* eth/tracers: add tests to run all tracers against state-sync transaction

* eth, core: explain state-sync address representation in comments

* core/types: add tests to validate empty values in state-sync tx getters

* eth: disable live tracer completely if parallel evm is enabled

* eth/tracers: record gas used for state-sync transactions in root frame

* eth/tracers: add explainer log for sorting state db logs

* consensus/bor: reuse reference

* pass vmconfig while bor consensus creation

* core: check for start and end hooks before tracing

* consensus, eth: pass ctx in state-sync execution for timeout, add bloom to trace receipts

* core/types: fix typo

* eth/tracers: correct timeout ctx

* eth/tracers: add more tests

* eth, consensus: address bot comments

* builder, packaging/templates: update all config.toml files

---------

Co-authored-by: Milan Došlić <milan.doslic@tenderly.co>

* Fix triejournaldirectory wiring (#2213)

* wire trie journal directory from server config

* server: guard trie journal path resolution

* server: deduplicate trie journal default resolution

---------

Co-authored-by: Lucca Martins <lucca_martins30@yahoo.com.br>

* miner: fix TestPrefetchRaceWithSetExtra hanging on unbuffered channel send (#2233)

Assisted-by: Claude:claude-sonnet-4-6

Co-authored-by: Lucca Martins <lucca_martins30@yahoo.com.br>

* miner: disable tracer in vm.Config to prevent conflicts during block synchronization (#2206)

* miner: disable tracer in vm.Config to prevent conflicts during block synchronization

* Update miner/worker.go

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* miner: add test to verify vmConfig strips tracer during live tracing

---------

Co-authored-by: StreamingFast Agent <agent@streamingfast.io>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* ethclient: bound tx indexing wait in tests (#2211)

* chore: use Claude Opus 4.7 1M context in workflows (#2241)

* ci: remove claude github actions (#2244)

Replaced by the built-in Claude review bot now enabled on this repo.

* params: version bump to v2.8.3-beta

* v2.8.2 (#2245)

* miner: track env.size in commitTransaction

env.size feeds txFitsSize for the params.MaxBlockSize cap. It was
seeded from header.Size() in makeEnv but never updated as transactions
were committed, so the cap was effectively inert.

* miner: set maxBlockSizeBufferZone to 1MB

* Bump up version to v2.8.2

* params: version bump to v2.8.3-beta2

* params: version bump to v2.8.3-beta3

* Revert "core/forkid: include polygon-specific forks in wire forkid (#2237)"

This reverts commit 5dc5adb9eebe7cfcd2707d54194a10e34d21481a.

* params: version bump to v2.8.3-beta4

* Revert "core/forkid: include polygon-specific forks in wire forkid (#2237)" (#2248)

This reverts commit 5dc5adb9eebe7cfcd2707d54194a10e34d21481a.

* consensus: fix milestone-mismatch rewind deadlock (#2246)

* params: bump version to v2.8.3-beta5

* params: bump version to v2.8.3

* build: bump Go to 1.26.4 to fix stdlib CVEs (GO-2026-5039, GO-2026-5037)

* Revert "build: bump Go to 1.26.4 to fix stdlib CVEs (GO-2026-5039, GO-2026-5037)"

This reverts commit b5a182b7906b066509af6db62f3e9527a73896a2.

---------

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: Lake <lake@dunamu.com>
Co-authored-by: Manav Darji <manavdarji.india@gmail.com>
Co-authored-by: Larry Arnault <Larryarnaultla@gmail.com>
Co-authored-by: Harsh kumar <51978908+harshinsecurity@users.noreply.github.com>
Co-authored-by: harshinsecurity <hi@harshinsecurity.in>
Co-authored-by: Aaron Ajose <aaronajosee@gmail.com>
Co-authored-by: haoyu-haoyu <85037553+haoyu-haoyu@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Pratik Patil <pratikspatil024@gmail.com>
Co-authored-by: Jerry <jerrycgh@gmail.com>
Co-authored-by: Krishang <109511742+kamuikatsurgi@users.noreply.github.com>
Co-authored-by: Sanket Saagar Karan <ssaagar@polygon.technology>
Co-authored-by: Karolina G. <60556109+kajaaz@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Marcello Ardizzone <marcelloardizzone@hotmail.it>
Co-authored-by: Andrei Petrov <andreipetrovqs@gmail.com>
Co-authored-by: Olivier Dupont <olivierdupontvier@gmail.com>
Co-authored-by: Léo Vincent <28714795+leovct@users.noreply.github.com>
Co-authored-by: Milan Došlić <milan.doslic@tenderly.co>
Co-authored-by: Jonas Schneider <jonasschnierderrrrrr@gmail.com>
Co-authored-by: William Morriss <wjmelements@gmail.com>
Co-authored-by: Matthieu Vachon <matt@streamingfast.io>
Co-authored-by: StreamingFast Agent <agent@streamingfast.io>
Co-authored-by: Adam Dossa <adossa@polygon.technology>