Repository navigation
kona-engine,op-acceptance-tests: insert pre-Canyon payloads with newPayloadV2 and unskip the op-reth sync tests - #24092
Merged
Conversation
The test asserted that a block executed by engine_newPayload but not yet canonicalized by a forkchoice update is served by eth_getBlockByHash. The Engine API does not require that: op-geth persists such side-chain blocks, op-reth keeps them in its engine tree and does not serve them over RPC. Drop that assertion and the op-reth skip. The rest of the test (newPayload VALID, not canonical by number until the FCU, canonical afterwards) holds on any EL. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The test only ran on op-geth and mostly pinned op-geth skeleton-sync internals that the Engine API leaves open: that newPayload drops payloads it answers SYNCING instead of buffering them (op-reth buffers and later connects them), and that side-chain blocks are served by hash before an FCU. Its spec-mandated parts already run on op-reth in the same package: TestSafeDoesNotAdvanceWhenUnsafeIsSyncing_NoELP2P covers newPayload VALID on an appendable block, FCU promoting it to canonical, FCU to a head with missing ancestors returning SYNCING without moving the heads, and the FCU turning VALID once the EL is peered. TestELP2PFCUUnavailableHash covers FCU-triggered EL sync over EL P2P reaching VALID. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The op-reth skip is stale for op-node: on op-reth the chain activates every fork from Bedrock to Jovian after genesis, and the sync-tester verifier follows it over CL P2P and by derivation. kona-node still cannot sequence this chain. Its engine inserts pre-Canyon payloads, which have no withdrawals, with engine_newPayloadV1. op-reth only serves engine_newPayloadV2 and later, as the OP Stack Engine API uses them, so every insert fails with -32601 and the unsafe head stays at genesis. Skip the test on kona-node until that is fixed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TestL2ELP2PCanonicalChainAdvancedByFCU was its only caller. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A pre-Canyon payload has no withdrawals and decodes to a V1 envelope, which the engine inserted with engine_newPayloadV1. op-reth serves only engine_newPayloadV2 and later, so on op-reth every pre-Canyon insert failed with -32601: a kona-node could neither sequence nor CL-sync pre-Canyon blocks. engine_newPayloadV2 accepts a V1 payload, and op-node already sends every pre-Ecotone payload with it, with or without withdrawals. With no caller left, drop EngineClient::new_payload_v1. The unit tests that insert a V1 envelope now expect engine_newPayloadV2; the ones that check the parameters pin that the request carries no withdrawals. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
kona-node now inserts pre-Canyon payloads with engine_newPayloadV2, so a kona-node sequencer on op-reth gets past the Bedrock genesis of this chain and the test passes in both CL lanes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Both payload arms of the insert now call engine_newPayloadV2, so the method name no longer tells them apart. Pin that a V2 envelope's insert sends the envelope's withdrawals, and word the V1 arm's comment by what it matches: the envelope's shape. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
joshklop
approved these changes
Nov 13, 2026
joshklop
added a commit
that referenced
this pull request
Nov 17, 2026
develop now inserts pre-Canyon payloads with engine_newPayloadV2 (#24092), so finalization_target_survives_resets_below_it must expect that method. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
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.
kona-node inserts pre-Canyon payloads (which have no withdrawals) with
engine_newPayloadV1, which op-reth does not serve. Against op-reth, a kona-node can neither sequence nor CL-sync pre-Canyon blocks: every insert fails with "method not found" and the unsafe head never moves. kona-node now sends them withengine_newPayloadV2, as op-node does; the Shanghai Engine API lets V2 carry a V1 payload. The payload version for Canyon and later is unchanged.EngineClient::new_payload_v1has no caller left and is removed.The bug surfaced while resolving the last three
SkipOnOpRethtests. CI runs acceptance tests only on op-reth, so each such test is dead coverage, and the skip helper has to go with the in-process op-geth L2 EL (#21744); #23827 and #24073 resolve the other four.TestSyncTesterHardforks(Bedrock at genesis, every fork up to Jovian activated in-chain) runs on op-reth with both op-node and kona-node. With op-node the skip was stale; with kona-node it failed on the bug above.TestVerifierManualSyncruns on op-reth. It no longer asserts that a block executed byengine_newPayloadbut not yet canonicalized is served byeth_getBlockByHash: the Engine API doesn't require that, and op-reth returns null for such a block until the forkchoice update, where op-geth serves it from its database.TestL2ELP2PCanonicalChainAdvancedByFCUis deleted, with its now-unused DSL helperFinishedELSync. It pins op-geth skeleton-sync internals (SYNCING payloads dropped rather than buffered, side-chain blocks served by hash, geth log lines), and parts of it would fail on op-reth, which buffers such payloads. What the spec requires of it (a forkchoice update to a head with missing ancestors returns SYNCING without moving the heads, then VALID once peered) is covered on op-reth byTestSafeDoesNotAdvanceWhenUnsafeIsSyncing_NoELP2PandTestELP2PFCUUnavailableHashin the same file.One difference from op-node remains and is left alone: for a pre-Canyon forkchoice update with attributes, op-node sends V1 and kona-node sends V2, which op-reth serves.
Part of #23317
Part of #21744
🤖 Co-created with Claude Opus 5.5 and Claude Fable 5.2