Repository navigation
Conversation
Cross-repo engine synchronizationLinked ARIEC61850 engine issue #158 for typed whole-FCD The uploaded C264 IID/diagnostic remain user-provided trial data and were not committed to public GitHub. This PR contains only two new Markdown documents, stacked on #501; user-installed merge 1a364f and #501 source trees are identical. The local simulator could support MMS without emitting Ethernet GOOSE; real L2 evidence must precede any adapter auto-start claim. Handoff prompt file: |
|
P8 implementation checkpoint: consumer GOOSE same-host NIC identity fix is now isolated in draft PR #505 (#505), one atomic commit |
|
Reconciliation finding (2026-10-26): KEEP OPEN — distinct documentation/evidence not in RC. This docs-only PR HEAD |
Reconciliation D3 — sanitized archival disposition (2026-10-26)Disposition: B — the two distinct historical C264 planning/handoff documents are now archived as engineering evidence, not code-merged or declared physically accepted. Closing this administrative historical PR unmerged; original head/branch remain accessible. Canonical replacement: isolated audit archive C264 sanitized engineering evidence, immutable archive commit Unique delta preserved: Both original #502 files were read in full and individually mapped. Safety: No field endpoints, MAC, raw user IID, privileged capture/logs copied into new archive. No claim that planned source was deployed, that a particular NIC transmits GOOSE, or that the historic simulator result extends to physical R7/R10. Closing only the planning document PR; no branch deletion, no unsafe cherry-pick/merge, and no Windows CI rerun. Follow-up: Use sanitized archival P0–P5 findings only after verifying current RC behavior. Active #606/#753 GOOSE raw admission, #615 MMS late-command ownership, #493/#495 F650 report and #729 physical gates remain independently unresolved; re-open if a concrete unique requirement in the original files was omitted from the archive. |
Documentation-only, baseline-locked strategy (no functional code changes)
Derived from user-supplied C264 same-host simulator diagnostic + IID and verified source review. New plan and full new-thread handoff are saved under
docs/field-trials/C264_SAME_HOST_REPAIR_PLAN_2026-10-10.mdanddocs/field-trials/C264_HANDOFF_PROMPT_2026-10-10.md.Protected baselines: ARSAS draft #494/#497/#498/#499/#500/#501 source HEAD 34d8f55; user-installed C264 executable merge commit 1a364f is tree-identical; engine ARIEC61850 #157 HEAD 8b0c0f. ARSAS draft chain has diverged from main, so no merge/rebase/reset, engine update or public release.
Source-grounded causes: C264 local MMS simulator uses KM_TEST virtual loopback, while GOOSE card resolver rejects loopback and fails to select an actual Npcap observer; local MMS success is not proof of Ethernet GOOSE publisher emission. Static reporting 207 points, 73 pending and 4/29 proven RCB ingress; some received structured whole-FCD values are rejected during semantic scalar projection. UX fixes: center editable IED-IP template, align fault-record header select-all checkbox with row checkboxes.
Phased implementation: P0 sanitized synthetic replay and regression locks; P1 focused WPF alignment; P2 engine-owned exact typed FCD schema projection; P3 context-aware same-host GOOSE capture with bounded read-only evidence probe and explicit no-publisher state; P4 honest Pending/Rejected/No-quality UI; P5 exact-head Windows CI and local simulator/physical acceptance. No cyclic MMS process fallback, no dynamic DataSet writes, no untyped regex guesses, no additional native code.
PR intentionally DRAFT. Next implementation work requires discrete code PRs with tests and field evidence; this PR itself must not be misrepresented as a bug fix.