test(ledger-suite-orchestrator): drop the StateMachine test fixture

[gregorydemay]

Completes the DEFI-2262 PocketIC migration for rs/ethereum: with the ckETH minter's integration tests now running on PocketIC, nothing in rs/ethereum still uses the ledger-suite-orchestrator's StateMachine-based test fixture. This PR deletes it, along with the now-unused ic-state-machine-tests, ic-types, ic-management-canister-types-private, and assert_matches dependencies, leaving PocketIC as the orchestrator's sole test harness.

• Removed the StateMachine-based LedgerSuiteOrchestrator/flow/universal_canister fixture from ledger-suite-orchestrator/test_utils; the pocket_ic module (added in test(ledger-suite-orchestrator): migrate integration tests to PocketIC #10949) is now the crate's only fixture.
• Kept the environment-agnostic helpers the pocket_ic module and cketh still import at their existing paths (wasm loaders, init-arg builders, usdc()/usdt(), constants).
• Dropped the now-unused ic-state-machine-tests, ic-types, ic-management-canister-types-private, and assert_matches dependencies from Cargo.toml/BUILD.bazel.

📚 PR stack

1. test(ledger-suite-orchestrator): migrate integration tests to PocketIC #10949 — test(ledger-suite-orchestrator): migrate integration tests to PocketIC
2. test(cketh): deduplicate signed-transaction literals in integration tests #10950 — test(cketh): deduplicate signed-transaction literals in integration tests
3. test(cketh): migrate integration tests to PocketIC #10955 — test(cketh): migrate integration tests to PocketIC
4. test(ledger-suite-orchestrator): drop the StateMachine test fixture #10956 — test(ledger-suite-orchestrator): drop the StateMachine test fixture 👈 you are here

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Removes the legacy ic-state-machine-tests-based test fixture from rs/ethereum/ledger-suite-orchestrator/test_utils, leaving the existing PocketIC fixture as the only supported test harness and cleaning up now-unused dependencies.

Changes:

• Deleted the StateMachine-based flow and universal_canister fixtures and removed the StateMachine LedgerSuiteOrchestrator helper implementation from test_utils/src/lib.rs.
• Kept the environment-agnostic helpers (wasm loaders, init-arg builders, token helpers/constants) at their existing crate-root paths.
• Dropped unused dependencies (ic-state-machine-tests, ic-types, ic-management-canister-types-private, assert_matches) from Cargo + Bazel, and updated Cargo.lock.

Reviewed changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
rs/ethereum/ledger-suite-orchestrator/test_utils/src/universal_canister.rs	Deleted the StateMachine-based universal canister helper (PocketIC version remains under pocket_ic/).
rs/ethereum/ledger-suite-orchestrator/test_utils/src/lib.rs	Removed StateMachine fixture exports/impls; retained shared helper APIs/constants used by PocketIC + ckETH.
rs/ethereum/ledger-suite-orchestrator/test_utils/src/flow.rs	Deleted the StateMachine-based flow/assertion helpers (PocketIC equivalents remain under pocket_ic/).
rs/ethereum/ledger-suite-orchestrator/test_utils/Cargo.toml	Removed unused StateMachine-era dependencies.
rs/ethereum/ledger-suite-orchestrator/test_utils/BUILD.bazel	Removed unused Bazel deps corresponding to the deleted fixture.
Cargo.lock	Reflected the dependency removals for ic-ledger-suite-orchestrator-test-utils.

---

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

[gregorydemay]

🤖🧐 VERDICT: CHANGES_REQUESTED — 0 blockers, 0 mediums, 3 nits; CI pending (Bazel Test All, Bazel Test All on RBE, Cargo Build Linux, Cargo Lint Linux, Build IC, Bazel Run Fuzzers still running). The review substance is clean — the only thing standing between this and READY is green CI.

Review details

Verdict rationale

Nothing in the diff needs to change. CHANGES_REQUESTED is bookkeeping: my rules make a non-green CI a hard block on READY, and six jobs were still pending when I looked. Everything that has reported so far passes (Bazel Test arm64-linux, Bazel Test arm64-darwin, Autofix, Lock Generate, Candid compatibility checks, Check PULL_REQUEST_BAZEL_TARGETS, …). Re-review once the remaining jobs land.

The three questions a deletion PR has to answer

1. Was anything still needed deleted? No.
I enumerated every symbol surviving in test_utils/src/lib.rs and found a live consumer for each — the pocket_ic module (GIT_COMMIT_HASH, MAX_TICKS, the four wasm loaders, default_init_arg, MINTER_PRINCIPAL, the Ledger* re-exports), tests/tests.rs (GIT_COMMIT_HASH_UPGRADE, NNS_ROOT_PRINCIPAL, cketh_installed_canisters, tweak_ledger_suite_wasms, usdc/usdt/usdc_erc20_contract/usdt_erc20_contract), and cketh (supported_erc20_tokens, usdc, CKERC20_TRANSFER_FEE, pocket_ic::LedgerSuiteOrchestrator, pocket_ic::flow::call_ledger_icrc1_total_supply). The remaining private items (modify_wasm, is_gzipped_blob, ledger_init_arg, archive_wasm, ledger_get_blocks_disabled_wasm) are used inside the crate.

Conversely, every deleted symbol is gone repo-wide: new_state_machine, out_of_band_upgrade, stop_canister, assert_reply, AddErc20TokenFlow, ManagedCanistersAssert, UniversalCanister, CanisterHttpQuery have no reference outside the pocket_ic module's own (differently-typed) namesakes, and no file anywhere imports …test_utils::flow or …test_utils::universal_canister at the old root paths.

2. Was anything dead kept? Three pre-existing dead deps — see the inline nit on Cargo.toml. They predate the stack, so leaving them is legitimate surgical scope; I flagged them as a nit rather than a medium.

3. The sweep claims — one holds as stated, one needs restating.

Grep sweep: holds. grep -rn "state_machine_tests\|StateMachine\|state_machine" rs/ethereum returns 10 hits: 9 are historical git-log excerpts inside cketh/mainnet/*_upgrade_*.md proposal records (immutable by nature) and 1 is the explanatory comment at cketh/test_utils/src/lib.rs:941 — a cketh file, out of this PR's component scope. No Rust code in rs/ethereum mentions StateMachine.

Bazel sweep: the substantive criterion is met, but the plan's literal wording is unachievable. rdeps(//rs/ethereum/..., //rs/state_machine_tests) is not empty — it lists the four integration_tests targets. The reason is not a leftover: somepath(//rs/ethereum/ledger-suite-orchestrator:integration_tests, //rs/state_machine_tests) shows exactly one edge chain, integration_tests → //:pocket-ic-server (data) → //rs/pocket_ic_server:pocket-ic-server → //rs/state_machine_tests. The PocketIC server binary reuses the state-machine internals, so any target that runs PocketIC reaches //rs/state_machine_tests through its data attribute — the plan's → empty criterion could never hold for a PocketIC-based suite. The criterion that actually matters is compile-time linkage, and it is clean: attr(deps, "state_machine_tests", //rs/ethereum/...) returns nothing, and rdeps(//rs/ethereum/..., //rs/state_machine_tests) except rdeps(//rs/ethereum/..., //:pocket-ic-server) leaves no rs/ethereum target at all. No Rust library or test in rs/ethereum compiles against ic-state-machine-tests any more. I consider the acceptance criterion met in substance; the plan's wording should be read as "no compile dependency", not "no node in the transitive graph".

Maintainability rundown

• Duplication: none found — and the PR removes the ~1000-line dual-fixture duplication the plan itself listed as a risk for the stack's lifetime.
• Unused derives: none — no types are introduced.
• Primitive-obsession parameters: none — no signatures are introduced.
• Divergent invariant handling: none — no invariant checks are introduced or moved.
• Silent fallbacks: none — no failure paths are introduced.
• Dead code (G9): the three pre-existing dead deps above; nothing newly orphaned by this deletion.
• Component boundary: clean. The delta touches only rs/ethereum/ledger-suite-orchestrator/test_utils/** plus Cargo.lock.
• Cargo/Bazel mirroring: exact. All four removals are mirrored 1:1 — assert_matches↔@crate_index//:assert_matches, ic-state-machine-tests↔//rs/state_machine_tests, ic-types↔//rs/types/types, ic-management-canister-types-private↔//rs/types/management_canister_types — and Cargo.lock drops exactly those four entries. ic-types-cycles is correctly retained: pocket_ic/universal_canister.rs:7,43,62 uses Cycles.
• Docs: the PR description and commit message list three dropped deps but four were dropped — assert_matches is missing from both. Trivial; mention it if you push another commit.

Testing

Pure deletion of unreferenced code, so no behaviour changes and no new tests are warranted; the guarantee needed is that the surviving suites are unaffected, which they are.

What I ran

• cargo check --all-targets --all-features -p ic-ledger-suite-orchestrator-test-utils -p ic-ledger-suite-orchestrator -p ic-cketh-test-utils -p ic-cketh-minter → clean.
• bazel test //rs/ethereum/ledger-suite-orchestrator:integration_tests //rs/ethereum/cketh/minter:integration_tests → 3/3 PASSED (integration_tests 15.8s, cketh_test 75.9s, ckerc20_test 147.7s).
• bazel build //rs/ethereum/... → 48 targets, success (nothing else in the component referenced the deleted fixture).
• The four Bazel queries quoted above.