[dborgards]

What does this change?

Two CANopen SDO block-client regression fixtures advanced protocol time while deferred RawCan echo confirmations still used host-time deadlines. Give them a separate VirtualClock confirmation actor and explicitly witness the actual confirmation timer before and after every step. The node uses the same time source, and actor round trips still settle its protocol timers.

The long-subblock case retains its 600 ms segment delays and one-second SDO/RawCan budgets. The late-confirmation case configures a 500 ms SDO timeout and confirms at 750 ms, before RawCan's unchanged one-second deadline: this isolates a protocol timeout followed by a transport-valid late confirmation. Previously its 1.5-second virtual wait passed the physical confirmation deadline without advancing it. All payload, end-request and timeout-abort assertions remain. No production code or public API changes.

Refs #362. Remaining reset and SDO block-server timing fixtures are still open in the inventory.

Type of change

• docs / test / refactor / chore / ci — no release

Validation

Both focused cases pass. Three real source mutation configurations were runtime-red and restored: restore host-time RawCan binding (both cases fail), omit per-segment SDO rearming, and ignore an already expired block-client timeout. Runner: /private/tmp/cankit-362-sdo-confirm-mutations.py.

Full local gates passed: 3043 net10 tests with coverage, net48 compilation, CI build, formatting, nine verified packages, 50 engineering tests and traceability. All 444 Cobertura classes/nested types were parsed; production diff is empty, with no changed executable missing/partial entries.

Checklist

• Build with -p:CI=true, test (net10.0), dotnet format --verify-no-changes, and pack + eng/verify-packages.py all pass locally
• New behaviour is covered by a test (existing regressions and positive timer witnesses)
• The requirement or ADR this relates to is referenced (Tests: wall-clock-gated tests after #92 — the 2026-10-21 inventory #362)

[cursor]

PR Summary

Low Risk
Test-only harness changes; no production or public API modifications.

Overview
Two SDO block-download regression tests (#362, C12) no longer advance protocol time with ManualTimeSource while deferred echo confirmations still used host-time deadlines.

They now run on a VirtualClock with a dedicated confirmation actor, ClockedCanBusService, and the node's timeSource aligned to that actor. Each segment step waits for the confirmation timer to arm, advances virtual time, then waits for the re-armed SDO segment timer before releasing echoes.

The late-confirmation case uses a 500 ms SdoTimeout (was 1 s) so the protocol timeout fires at 750 ms virtual time before RawCan's unchanged 1 s confirm deadline—isolating timeout-then-late-confirm without reviving the transfer. Assertions on payloads, end requests, and SdoProtocolTimedOut are unchanged.

^{Reviewed by Cursor Bugbot for commit 332e6fa. Bugbot is set up for automated code reviews on this repo. Configure here.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	🔄 Running since 2026-10-27T05:28:45.043408Z	332e6fa	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!