test: end-to-end deposit-from-CEX EIP-7702 sweep demo against a local Ethereum node by gregorydemay · Pull Request #10670 · dfinity/ic

gregorydemay · 2026-07-07T06:11:32Z

A hermetic, monorepo-native integration test that validates the "deposit from CEX" sweep (design in rs/ethereum/cketh/docs/deposit_from_cex.md) end-to-end against a local anvil node, through the minter's productive EIP-7702 transaction layer (#10671).

It sweeps unfunded deposit addresses to the minter in minter-paid EIP-7702 transactions, checks each emits the canonical deposit event with the right IC principal, measures gas across batch sizes (1/10/20), and covers both sweeper designs — caller-gated and permissionless-with-minter-attestation — including rejection of unauthorized sweeps.

foundry and solc are vendored via Bazel (no docker, IC replica, or alloy).

Copilot

Pull request overview

Adds a standalone, runnable Foundry/Alloy proof-of-concept under rs/ethereum/cketh/docs/ that demonstrates the “deposit from CEX” EIP-7702 sweeping flow referenced by the design doc, including single and batched sweeps, helper-based sweeps that emit canonical events, and a rejection case for non-minter sweeping.

Changes:

Introduces a Rust CLI demo that deploys mock contracts, simulates CEX deposits, and performs EIP-7702 single/batched sweeps (direct and via the real helper), asserting gas usage.
Adds Solidity demo contracts (sweeper delegates + mock USDT) and embeds their compiled bytecode artifacts for reproducible runs.
Documents how to run and rebuild the artifacts, plus Foundry configuration and local ignore rules.

Reviewed changes

Copilot reviewed 12 out of 13 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
rs/ethereum/cketh/docs/deposit_from_cex_demo/src/main.rs	Runnable Alloy-based demo driving the full sweep flow (single/batch, direct/via helper, attacker rejection) and asserting gas usage.
rs/ethereum/cketh/docs/deposit_from_cex_demo/README.md	Runbook and explanation of the demo, variants, and artifact regeneration steps.
rs/ethereum/cketh/docs/deposit_from_cex_demo/foundry.toml	Configures Foundry for Prague/Pectra-compatible EVM settings.
rs/ethereum/cketh/docs/deposit_from_cex_demo/contracts/CkSweeper.sol	Variant A sweeper delegate + batch entry point for direct-to-minter token/ETH sweeps.
rs/ethereum/cketh/docs/deposit_from_cex_demo/contracts/CkSweeperViaHelper.sol	Variant B sweeper delegate + minter-gated batch entry point that routes deposits through the helper contract.
rs/ethereum/cketh/docs/deposit_from_cex_demo/contracts/MockUSDT.sol	Mock USDT-like ERC-20 with non-standard return behavior to exercise tolerant transfer/approve handling.
rs/ethereum/cketh/docs/deposit_from_cex_demo/Cargo.toml	Declares a standalone (non-workspace) Rust crate for the demo.
rs/ethereum/cketh/docs/deposit_from_cex_demo/Cargo.lock	Locks dependencies for reproducible demo builds.
rs/ethereum/cketh/docs/deposit_from_cex_demo/artifacts/CkSweeper.bin.hex	Embedded compiled bytecode for the Variant A delegate/batcher.
rs/ethereum/cketh/docs/deposit_from_cex_demo/artifacts/CkSweeperViaHelper.bin.hex	Embedded compiled bytecode for the Variant B delegate/batcher.
rs/ethereum/cketh/docs/deposit_from_cex_demo/artifacts/MockUSDT.bin.hex	Embedded compiled bytecode for the mock USDT token.
rs/ethereum/cketh/docs/deposit_from_cex_demo/artifacts/CkDeposit.bin.hex	Embedded compiled bytecode for the real helper contract used in Variant B.
rs/ethereum/cketh/docs/deposit_from_cex_demo/.gitignore	Ignores Foundry/Cargo build outputs for the demo directory.

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

+fn encode_principal(principal_text: &str) -> FixedBytes<32> {
+    let principal = Principal::from_text(principal_text).expect("valid principal");
+    let bytes = principal.as_slice();
+    let mut encoded = [0u8; 32];
+    encoded[0] = bytes.len() as u8;
+    encoded[1..=bytes.len()].copy_from_slice(bytes);
+    FixedBytes::from(encoded)
+}


github-actions

Auto approving mainnet-icos-revisions.json changes.

Stacked on top of the design-doc PR (#10652). Adds the runnable alloy/foundry demo of the EIP-7702 sweep mechanism (both variants, gas assertions, attacker rejection) that the design document references. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

@foundry

Vendors the foundry v1.7.1 release binaries (anvil/cast/forge/chisel) via http_archive and adds a monorepo-native rust_test that spawns anvil and probes it over JSON-RPC. No docker, no InternetComputer, no alloy: the skeleton foundation for hand-rolling the deposit-from-CEX PoC's Ethereum layer on the already-vendored k256/sha3/rlp/ethnum crates. The pinned foundry release is >= v1.0 so anvil supports EIP-7702, unlike the v0.3.0 @foundry oci image used by //rs/tests/cross_chain:ic_xc_cketh_test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Vendor foundry binaries for all four host platforms (linux/darwin x amd64/arm64) and select the matching one, so the test runs on macOS as well as linux CI. Previously only linux_amd64 was vendored: on an arm64 Mac the Linux ELF spawned (via the execvp /bin/sh fallback) and died immediately, but the readiness loop kept polling a dead process until the 30s timeout. The loop now detects an early child exit and fails fast with the exit status. Inline the archive's build_file content instead of a separate third_party/BUILD.foundry.bazel. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Relocate the hermetic anvil smoke test to rs/ethereum/cketh/minter/tests and excludes it from the integration_tests suite, which runs against the IC test machinery rather than a local foundry node. The platform-selecting anvil alias and its dedicated rust_test target move alongside it into the minter BUILD. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…-7702 tx layer Fills in the migrated test with the decided variant B flow, driving a local anvil node through the minter's own transaction layer (Eip7702TransactionRequest, Authorization/SignedAuthorization, Signed) rather than a hand-rolled encoder or alloy. Signing uses ic_secp256k1 locally as a stand-in for threshold ECDSA; ABI calldata, JSON-RPC transport and event decoding are minimal helpers. The test deploys a USDT-style ERC-20, the real ckETH helper and the sweeper delegate, then sweeps unfunded deposit EOAs in single and batched EIP-7702 transactions (gas paid by the minter), asserts the canonical ReceivedEthOrErc20 events carry the right IC principal, and shows a non-minter sweep is rejected before the minter finalizes it with a plain EIP-1559 transaction. Gas used by each sweep is asserted against deterministic constants. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Drops the checked-in .bin.hex artifacts and compiles the demo contracts from their Solidity sources at test time with the vendored solc, mirroring how ic_xc_cketh_test provisions its contracts. The real helper (CkDeposit) is compiled from its canonical DepositHelperWithSubaccount.sol rather than a duplicated bytecode blob, and the two demo-only sources are consumed directly from the PoC's contracts directory. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Vendors the latest stable solc (0.8.35) pinned exactly by sha256 and compiles the demo contracts with an explicit --evm-version prague, the fork that defines EIP-7702 and that anvil runs. Pinning both the compiler and the EVM version keeps the bytecode (and the asserted gas) reproducible instead of drifting with solc's default target (now osaka/fusaka). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Makes the demo test's solc host-platform aware like the anvil binary: vendors solc 0.8.35 for linux-amd64/linux-arm64/macos and selects it through a `solc` alias reusing the existing config_settings. The macOS build is a universal (x86_64 + arm64) binary, so the test runs natively on Linux and on both Intel and Apple Silicon Macs. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Follows the repository convention for host-platform-selected tool binaries (bitcoind, dfx, ruff, ...): defines //:anvil and //:solc as aliases selecting on @bazel_tools//src/conditions in the top-level BUILD, and drops the bespoke config_settings from the minter BUILD. The demo test consumes //:anvil and //:solc. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…y PoC Moves the two contracts the test compiles (MockUSDT.sol, CkSweeperViaHelper.sol) into minter/tests/deposit_from_cex_demo/, next to the test that consumes them, and removes the now-redundant standalone alloy PoC (docs/deposit_from_cex_demo: main.rs, Cargo.*, README, foundry.toml, the pre-compiled artifacts/ bytecode and the unused variant-A CkSweeper.sol). The productive minter test is now the single deposit-from-CEX demo, with no duplicated contracts or bytecode. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Refactors the deposit-from-CEX demo into a table-driven test over the two batch sizes from the design doc's cost-estimation section: a single deposit swept in one batch (B=1) versus a batch of twenty (B=20). Each scenario funds fresh unfunded deposit EOAs from the CEX and sweeps them all in one EIP-7702 transaction, asserting the canonical events and that per-deposit gas amortizes as the batch grows (94_932 for B=1 vs 55_668/deposit for B=20). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Re-adds the attacker-rejection as a focused second test: the minter installs a deposit EOA's delegation without sweeping (its authorization rides in an empty batch), then a non-minter's sweepErc20 attempt reverts and the funds stay put, while the minter's own sweep still succeeds via a plain EIP-1559 transaction (the delegation persists). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Review renamed Eip1559Signature to TransactionSignature and Signed::raw_transaction_hex to raw_transaction_bytes before #10671 landed. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Adds an average_gas_used field to BatchScenario (total / deposits) and asserts it alongside the total, making the amortization explicit: 94_932 per deposit at B=1 versus 55_668 at B=20. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Third point between B=1 and B=20 (570_151 gas, 57_015 per deposit), keeping the per-deposit amortization curve monotonic: 94_932 -> 57_015 -> 55_668. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…tion The first sweep installs the delegations (EIP-7702); a second sweep of the same addresses, after re-funding, reuses the persisted delegations and so needs only a regular EIP-1559 transaction. This isolates the EIP-7702 delegation-install overhead from the steady-state sweep cost. Per batch (total / per-deposit): B=1: EIP-7702 94_932/94_932 EIP-1559 63_252/63_252 B=10: EIP-7702 570_151/57_015 EIP-1559 390_151/39_015 B=20: EIP-7702 1_113_373/55_668 EIP-1559 753_373/37_668 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

) Design doc for [DEFI-2096](https://dfinity.atlassian.net/browse/DEFI-2096): allow onramping ckUSDC/ckUSDT (and generally any ckERC20, later ckETH) directly from a centralized exchange withdrawal. Deposits currently require calling the helper smart contract, which a CEX cannot do: a CEX withdrawal is a plain transfer from a shared hot wallet, carrying no IC principal. The proposed design gives each IC account a unique tECDSA-derived deposit address and sweeps funds to the minter using EIP-7702 (Pectra), so deposit addresses never need ETH for gas and remain key-recoverable independently of any contract code. The doc covers the requirements, claim-based deposit detection, the sweeper delegate, fee model, a phased delivery (ckERC20 first, then ckETH with its balance-based detection and fixed-21k-gas constraints), and the discussed alternatives (CREATE2 forwarders, ERC-4337, permit-based sponsoring). The runnable proof-of-concept the doc references lives in a stacked follow-up, dfinity#10670. 🤖 Generated with [Claude Code](https://claude.com/claude-code) [DEFI-2096]: https://dfinity.atlassian.net/browse/DEFI-2096?atlOrigin=eyJpIjoiNWRkNTljNzYxNjVmNDY3MDlhMDU5Y2ZhYzA5YTRkZjUiLCJwIjoiZ2l0aHViLWNvbS1KU1cifQ --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com> Co-authored-by: IDX GitHub Automation <infra+github-automation@dfinity.org>

Adds CkSweeperAttested, an EIP-7702 delegate matching the proposed one-time self-attestation design: sweeping is permissionless but each sweep carries an attestation — a secp256k1 signature by the deposit address' own key (the minter's role via threshold ECDSA) over keccak256("ck-deposit-owner" || chainid || helper || principal || subaccount), verified with ecrecover == address(this). Runs the same table-driven batch gas test through the attested sweeper with a non-minter relayer submitting the transactions, plus a test that a forged attestation is rejected while the genuine one credits the attested account. The attestation adds ~4k gas/deposit over the caller-gated variant (per batch of 20: 59_645 vs 55_668 for the delegating sweep, 41_645 vs 37_668 for the steady-state EIP-1559 sweep). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…led codec Replaces the custom Token enum + head/tail ABI encoder (including the TupleArray special case) and the manual return/event decoding with ethers-core's ethabi (encode/decode/Token/ParamType), already a dependency of the sibling integration_tests. The attestation preimage stays an explicit byte concatenation (it is a hash preimage, not ABI, and abi.encodePacked semantics differ). Gas figures are unchanged, confirming the encoding is byte-for-byte equivalent. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

The style_edition=2024 cargo-fmt check orders imports case-sensitively while the autofix bot re-sorts them case-insensitively, so a mixed-case group (e.g. {ParamType, Token, decode, encode}) ping-pongs between them and fails Cargo Lint. Import only the uppercase items and fully-qualify the lowercase ethers_core::abi::{encode,decode} and serde_json::json!, leaving no group whose ordering depends on case. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

The Cargo Lint CI job runs clippy with -D clippy::unseparated_literal_suffix, so underscore-separate the u8 literals (0u8 -> 0_u8). Also declare reqwest as a dev-dependency of the crate (it was only in the Bazel target), which cargo needs to resolve the test, and record the edge in Cargo.lock. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

zeropath-ai · 2026-07-09T08:55:35Z

✅ No security or compliance issues detected. Reviewed everything up to 8993359.

Security Overview

🔎 Scanned files: 9 changed file(s)
🔗 Scan Link: https://zeropath.com/app/repositories/783a9bd4-fb4e-4e10-ab3c-c2fb2ddfb92f?scanId=cd55b62f-d854-4dfc-b5ec-a3450c0da818&codeScanTypes=PrScan&tab=issues

Detected Code Changes

Change Type	Relevant files
Enhancement	► rs/ethereum/cketh/minter/BUILD.bazel + Deposit_from_cex_demo_test target and related test wiring ► rs/ethereum/cketh/minter/Cargo.toml + Add reqwest dependency ► rs/ethereum/cketh/minter/tests/deposit_from_cex_demo.rs + New integration test file for deposit-from-CEQ demo flow
Enhancement	► rs/ethereum/cketh/minter/BUILD.bazel + Exclude certain tests in test suite and add a dedicated test name deposit_from_cex_demo_test
Enhancement	► MODULE.bazel + Add solc http_file definitions for solc v0.8.35 (linux/linux arm64/macosx) and Foundry binaries (linux amd64/arm64, darwin amd64/arm64)
Enhancement	► CARGO.lock/rs dependencies (indirect in diff) + Include reqwest in Rust workspace for the new test flow

github-actions Bot added the docs label Jul 7, 2026

gregorydemay requested a review from Copilot July 7, 2026 06:11

Copilot started reviewing on behalf of gregorydemay July 7, 2026 06:12 View session

gregorydemay mentioned this pull request Jul 7, 2026

docs: design for supporting deposits from CEX (DEFI-2096) #10652

Merged

Copilot AI reviewed Jul 7, 2026

View reviewed changes

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from eb1ca42 to 27806c5 Compare July 7, 2026 19:42

gregorydemay temporarily deployed to CREATE_PR July 7, 2026 19:42 — with GitHub Actions Inactive

pr-creation-bot-dfinity-ic Bot temporarily deployed to CREATE_PR July 7, 2026 19:47 Inactive

github-actions Bot approved these changes Jul 7, 2026

View reviewed changes

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from 60e0c3c to 8f0238a Compare July 8, 2026 05:09

gregorydemay temporarily deployed to CREATE_PR July 8, 2026 05:09 — with GitHub Actions Inactive

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from 8f0238a to 64abe70 Compare July 8, 2026 05:10

gregorydemay temporarily deployed to CREATE_PR July 8, 2026 05:10 — with GitHub Actions Inactive

gregorydemay changed the base branch from ic_DEFI-2096_design-support-deposit-from-cex to ic_DEFI-2919_eip7702-tx-layer July 8, 2026 05:10

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch 3 times, most recently from ce3d105 to 382bfca Compare July 8, 2026 07:23

Base automatically changed from ic_DEFI-2919_eip7702-tx-layer to master July 8, 2026 07:31

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from ef92401 to d7f815a Compare July 8, 2026 07:38

gregorydemay and others added 11 commits July 8, 2026 07:48

Automatically fixing code for linting and formatting issues

9d27dc0

Automatically fixing code for linting and formatting issues

d436812

test: rename anvil smoke test to deposit_from_cex_demo

d5121fe

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

gregorydemay and others added 5 commits July 8, 2026 07:48

test: adapt to the EIP-7702 tx API as merged in master (#10671)

05a19ab

Review renamed Eip1559Signature to TransactionSignature and Signed::raw_transaction_hex to raw_transaction_bytes before #10671 landed. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from ee71e0e to 05a19ab Compare July 8, 2026 07:50

gregorydemay and others added 2 commits July 8, 2026 08:01

test: add a batch-of-10 scenario

3521d47

Third point between B=1 and B=20 (570_151 gas, 57_015 per deposit), keeping the per-deposit amortization curve monotonic: 94_932 -> 57_015 -> 55_668. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from e13e8bf to 3521d47 Compare July 8, 2026 08:20

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from 35d0086 to fa6fac1 Compare July 8, 2026 08:55

gregorydemay force-pushed the ic_DEFI-2096_deposit-from-cex-demo branch from 33f2df6 to f55bd97 Compare July 8, 2026 09:33

IDX GitHub Automation and others added 3 commits July 8, 2026 09:37

Automatically fixing code for linting and formatting issues

567f1d1

gregorydemay changed the title ~~docs: runnable foundry PoC for the deposit-from-CEX design~~ test: end-to-end deposit-from-CEX EIP-7702 sweep demo against a local Ethereum node Jul 9, 2026

github-actions Bot added test and removed docs labels Jul 9, 2026

gregorydemay marked this pull request as ready for review July 9, 2026 08:54

gregorydemay requested review from a team as code owners July 9, 2026 08:54

github-actions Bot added @idx @defi labels Jul 9, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

test: end-to-end deposit-from-CEX EIP-7702 sweep demo against a local Ethereum node#10670

test: end-to-end deposit-from-CEX EIP-7702 sweep demo against a local Ethereum node#10670
gregorydemay wants to merge 24 commits into
masterfrom
ic_DEFI-2096_deposit-from-cex-demo

gregorydemay commented Jul 7, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

github-actions Bot left a comment

Uh oh!

zeropath-ai Bot commented Jul 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

gregorydemay commented Jul 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

zeropath-ai Bot commented Jul 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

gregorydemay commented Jul 7, 2026 •

edited

Loading