Add latest flashblock simulation mode by karankurbur · Pull Request #923 · worldcoin/world-chain

karankurbur · 2026-07-27T21:01:22Z

Summary

Add optional useLatestFlashblock request parameter to simulate_unsignedUserOp, defaulting to false.
Wire the simulate RPC to use the latest executed flashblock state when requested, while preserving the existing latest/full-block execution path otherwise.
Validate flashblock usage at the endpoint boundary: reject when flashblocks are disabled, when no latest flashblock exists yet, or when an explicit block is also provided.
Add endpoint-level integration tests for the new flashblock guard behavior.

Validation

cargo fmt
cargo test -p world-chain-rpc
cargo check -p world-chain-node
cargo check
git diff --check

cargo clippy -p world-chain-rpc --all-targets -- -D warnings could not run because this pinned nightly reports cargo-clippy is not applicable to nightly-2026-07-01-aarch64-apple-darwin.

Note

Medium Risk
Changes how simulation state is chosen for a new RPC flag and depends on flashblock pending state staying consistent with ETH pending behavior; default behavior is unchanged.

Overview
Adds optional useLatestFlashblock (default false) on simulate_unsignedUserOp so simulations can run against the latest executed flashblock instead of a canonical block.

When the flag is set, block/state resolution reads the pending flashblock from a shared watch receiver, uses that block’s header, and builds state from the parent block plus a BlockState overlay over the executed flashblock. The existing block-id path is unchanged when the flag is off.

The node threads the same pending_block() receiver used for flashblocks ETH APIs into WorldChainAddOns and the Simulate RPC via with_latest_flashblock. Requests are rejected if flashblocks are disabled, no flashblock is available yet, or block is set together with useLatestFlashblock.

Unit and integration tests cover deserialization, param validation, and the RPC error paths.

^{Reviewed by Cursor Bugbot for commit a018d3e. Bugbot is set up for automated code reviews on this repo. Configure here.}

cursor

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 5d05519. Configure here.}

Copilot

Pull request overview

Adds an opt-in simulation mode for simulate_unsignedUserOp that runs against the latest executed flashblock state (when available), while keeping the existing “latest/full-block” resolution as the default behavior.

Changes:

Adds useLatestFlashblock (default false) to SimulateUnsignedUserOpRequest and validates it (rejects when block is also provided).
Wires the RPC simulate implementation to optionally build a state provider from the latest executed flashblock via a watch::Receiver.
Adds unit + endpoint-level integration tests for serde defaults and invalid-parameter guard rails.

Reviewed changes

Copilot reviewed 6 out of 7 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
crates/rpc/src/simulate.rs	Adds the `useLatestFlashblock` request flag and routes state/header selection through a new flashblock-backed path when enabled.
crates/rpc/src/lib.rs	Re-exports `LatestFlashblockReceiver` for node integration.
crates/node/src/context.rs	Clones the pending-block receiver so it can be supplied to both flashblocks `eth_` and `simulate` wiring.
crates/node/src/add_ons.rs	Threads the pending-block receiver into `Simulate::from_eth_api` when the simulate namespace is enabled.
crates/rpc/tests/simulate_latest_flashblock.rs	New integration tests asserting `INVALID_PARAMS` errors for flashblock-disabled, missing flashblock, and conflicting `block` inputs.
crates/rpc/Cargo.toml	Adds `world-chain-test-utils` as a dev-dependency for the new integration test.
Cargo.lock	Lockfile update for the new dev-dependency.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

karankurbur · 2026-07-27T22:44:00Z

Addressed the parent state lookup review note in latest_flashblock_state: the flashblock overlay now uses state_by_block_hash(header.parent_hash) instead of history_by_block_hash(...), so it can resolve live tip/in-memory parent state.

Local follow-up validation:

cargo fmt passed
git diff --check passed
Cargo compile/test rerun was blocked in this fresh checkout by local disk space while building librocksdb-sys (No space left on device, filesystem had ~1.1GiB free).

karankurbur · 2026-07-27T22:46:45Z

Backwards compatibility pass pushed in 97f95b63:

Restored the original Simulate::new(client, evm_config, task_pool, task_guard) signature.
Restored the original Simulate::from_eth_api(client, evm_config, eth_api) signature.
Restored the original WorldChainAddOns::new(...) signature.
Added explicit .with_latest_flashblock(...) opt-in methods for the new flashblock receiver wiring.
Added an endpoint-level integration test proving requests that omit useLatestFlashblock still default to the latest full-block path and do not require flashblocks.

cargo fmt, cargo metadata --no-deps, and git diff --check pass locally. Full cargo test/check reruns remain blocked by local disk space while compiling librocksdb-sys in the fresh checkout (No space left on device, about 1.0GiB free).

crystalt · 2026-07-28T01:46:14Z

@@ -290,6 +310,7 @@ where
            simulate_enabled,
            witness,
        )
+        .with_latest_flashblock(pending_block)


nit: is there a way we could avoid adding this plumbing type code to all these places? codex said we could add a trait to FlashblocksEthApi? Something like this

pub trait PendingFlashblockSource { fn pending_flashblock_receiver(&self) -> Option<LatestFlashblockReceiver>; }

codex said,

FlashblocksEthApi could implement this by cloning its optional watch receiver. Then Simulate::from_eth_api could require SpawnBlocking + PendingFlashblockSource and obtain the receiver from registry.eth_api(). That should eliminate the pending_block field and with_latest_flashblock method on WorldChainAddOns, including the repeated forwarding through its generic builder transformations.

just a nit though.

0xOsiris · 2026-07-28T17:59:29Z

 /// Cross-request cache for resolved token metadata. LRU-bounded.
 type MetadataCache = Arc<Mutex<LruCache<Address, AssetInfo>>>;

+pub type LatestFlashblockReceiver =


Remove this type alias, it's just an indirection that makes the logic harder to read

0xOsiris · 2026-07-28T18:03:11Z

@@ -290,6 +310,7 @@ where
            simulate_enabled,
            witness,
        )
+        .with_latest_flashblock(pending_block)


0xOsiris · 2026-07-28T18:04:43Z

+
+    /// Configures access to the latest flashblock-backed pending block, when
+    /// flashblocks are enabled.
+    pub fn with_latest_flashblock(


Remove this, just add it to WorldChainAddOns::new

0xOsiris · 2026-07-28T18:05:21Z

            simulate_enabled,
            witness,
        )
+        .with_latest_flashblock(pending_block)


0xOsiris · 2026-07-28T18:05:31Z

            simulate_enabled,
            witness,
        )
+        .with_latest_flashblock(pending_block)


0xOsiris · 2026-07-28T18:05:38Z

            simulate_enabled,
            witness,
        )
+        .with_latest_flashblock(pending_block)


0xOsiris · 2026-07-28T18:05:48Z

            simulate_enabled,
            witness,
        )
+        .with_latest_flashblock(pending_block)


0xOsiris · 2026-07-28T18:07:57Z

+    /// Whether to simulate against the latest flashblock state instead of the
+    /// latest full block state. Defaults to `false`.
+    #[serde(default)]
+    pub use_latest_flashblock: bool,


You are breaking the wire format backwards compatibility of the request by not making this an Option

Add latest flashblock simulation mode

5d05519

karankurbur temporarily deployed to dev July 27, 2026 21:01 — with GitHub Actions Inactive

karankurbur marked this pull request as ready for review July 27, 2026 21:21

karankurbur requested review from 0xForerunner, 0xOsiris, Dzejkop, Kemperino, alessandromazza98, cichaczem, kilianglas, murph and piohei as code owners July 27, 2026 21:21

cursor Bot reviewed Jul 27, 2026

View reviewed changes

Comment thread crates/rpc/src/simulate.rs

crystalt requested a review from Copilot July 27, 2026 21:29

Copilot started reviewing on behalf of crystalt July 27, 2026 21:29 View session

Copilot AI reviewed Jul 27, 2026

View reviewed changes

Use live parent state for flashblock simulation

2dcc34b

karankurbur temporarily deployed to dev July 27, 2026 22:43 — with GitHub Actions Inactive

Preserve simulate constructor compatibility

97f95b6

karankurbur temporarily deployed to dev July 27, 2026 22:46 — with GitHub Actions Inactive

crystalt reviewed Jul 28, 2026

View reviewed changes

crystalt approved these changes Jul 28, 2026

View reviewed changes

Kemperino approved these changes Jul 28, 2026

View reviewed changes

0xOsiris requested changes Jul 28, 2026

View reviewed changes

0xOsiris reviewed Jul 28, 2026

View reviewed changes

Merge branch 'main' into codex/use-latest-flashblock-simulate

a018d3e

karankurbur temporarily deployed to dev July 30, 2026 23:21 — with GitHub Actions Inactive

Uh oh!

Conversation

karankurbur commented Jul 27, 2026 • edited by cursor Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Validation

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

karankurbur commented Jul 27, 2026

Uh oh!

karankurbur commented Jul 27, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

karankurbur commented Jul 27, 2026 •

edited by cursor Bot

Loading