chore(defi): remove deprecated ic-cdk imports in ic-ledger-suite-orchestrator

[gregorydemay]

Summary

Closes the last of the three remaining crates listed in DEFI-2304 — removing the #![allow(deprecated)] attributes introduced in #6264 when ic-cdk was upgraded to 0.18, by migrating the affected files off the deprecated ic_cdk surface.

This PR handles ic-ledger-suite-orchestrator:

• ic_cdk::api::management_canister::main::{CanisterIdRecord, CanisterStatusResponse, canister_status, start_canister, stop_canister} → ic_cdk::management_canister::{CanisterStatusArgs, CanisterStatusResult, StartCanisterArgs, StopCanisterArgs, canister_status, start_canister, stop_canister}.
• ic_cdk::api::call::{call, call_with_payment} and notify_with_payment128 in IcCanisterRuntime are rewritten on top of the ic_cdk::call::Call builder (Call::unbounded_wait(...).with_arg(...).with_cycles(...) and .oneway() for deposit_cycles).
• ic_cdk::api::call::RejectionCode → ic_cdk::call::{CallFailed, OnewayError, RejectCode}; the existing Reason::from_reject mapping is split into from_call_error / from_call_failed / from_oneway_error helpers that share a common from_reject_code_message core, preserving the same Reason variants.
• ic_cdk::id() → ic_cdk::api::canister_self().
• ic_cdk::api::canister_balance128() → ic_cdk::api::canister_cycle_balance().
• ic_cdk::api::stable::stable_size() → ic_cdk::stable::stable_size().
• get_canister_status now returns the modern CanisterStatusResult; ledger_suite_orchestrator.did is updated to mirror the new fields (the same shape already adopted by ckBTC minter in chore(XC): remove deprecation from ic-cdk in ic-ckbtc-minter #6761 and ckETH minter in chore(defi): remove deprecated ic-cdk imports in ic-cketh-minter #10290).

create_canister still pays the exact cycles_for_canister_creation amount the caller supplies (rather than letting the cdk auto-compute the cost via create_canister_with_extra_cycles), so the orchestrator's existing semantics around cycle accounting are unchanged.

Follows the pattern set by #6755 (ic-btc-checker) and #6761 (ic-ckbtc-minter). No behavior change.

[Copilot]

Pull request overview

Removes reliance on deprecated ic-cdk APIs in ic-ledger-suite-orchestrator by migrating management-canister calls and call patterns to the modern ic_cdk::management_canister surface and the ic_cdk::call::Call builder, and updates the public Candid interface accordingly.

Changes:

• Replaced deprecated management canister imports/types and updated get_canister_status to return the modern status shape.
• Rewrote canister-to-canister calls and cycle deposit logic to use ic_cdk::call::Call (including oneway notifications).
• Updated ledger_suite_orchestrator.did to reflect the modern CanisterStatusResult fields; removed #![allow(deprecated)] from affected test crates.

Reviewed changes

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

Show a summary per file

File	Description
rs/ethereum/ledger-suite-orchestrator/tests/tests.rs	Removes #![allow(deprecated)] after migrating off deprecated APIs.
rs/ethereum/ledger-suite-orchestrator/test_utils/src/lib.rs	Updates test utility decoding/types to use CanisterStatusResult.
rs/ethereum/ledger-suite-orchestrator/src/management/mod.rs	Migrates runtime management calls to ic_cdk::management_canister + Call builder; refactors rejection/error mapping.
rs/ethereum/ledger-suite-orchestrator/src/main.rs	Updates management-canister status endpoint, stable memory, and cycle balance calls to non-deprecated APIs.
rs/ethereum/ledger-suite-orchestrator/ledger_suite_orchestrator.did	Updates Candid types for the new canister status response shape.

---

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

[mbjorkqvist]

Thanks for the cleanup @gregorydemay!

[mbjorkqvist]

nit: I think it would be more accurate to call canister_liquid_cycle_balance here, but we would anyway return OutOfCycles from the create_canister call - the only difference would be that in case canister_cycle_balance is enough, but the liquid balance is too low, we'd end up making a call to create_canister that we could otherwise have avoided. Perhaps a test for this case could make sense (setting a high freezing threshold would reserve a lot of cycles)?

[gregorydemay]

🤖 Good catch — 824644f switches the pre-flight to ic_cdk::api::canister_liquid_cycle_balance() so it matches what Call::await will check internally (the variant that becomes CallFailed::InsufficientLiquidCycleBalance). Same Reason::OutOfCycles end result, but no longer wastes a Call::await when liquid balance is below payment and the raw cycle balance is above it.

I left the high-freezing-threshold integration test as a follow-up for now: it would need a new LedgerSuiteOrchestrator::with_high_freezing_threshold helper in test_utils and a dedicated test against create_canister, and the unit tests I added in src/management/tests.rs already lock in the InsufficientLiquidCycleBalance → OutOfCycles mapping. Happy to add the integration test in a follow-up PR if you think it would be valuable.

[mbjorkqvist]

Same comment as in the other PR - this seems like new behavior, and it could make sense to add a test.

[gregorydemay]

🤖 Same context as on #10290 (comment): the mapping is intentional new behaviour from the cdk 0.18 surface — Reason::OutOfCycles was a defined-but-never-produced variant before, and routing InsufficientLiquidCycleBalance to it yields a more accurate error message than the previous InternalError fallback.

824644f adds unit tests in src/management/tests.rs covering this for all three mapping helpers (from_call_error, from_call_failed, from_oneway_error) plus the reject-code and CallPerformFailed paths to lock in the existing shape.