[masarray]

P0 — reconcile the shipped v1.6.42 source into main, without rollback

Status (25 Oct 2026): Public Windows application v1.6.42 is ALREADY STABLE, operator-accepted, and published at https://github.com/masarray/arsas/releases/tag/v1.6.42 . This PR concerns promoting its exact source to main, a separate action still BLOCKED by physical-control/interoperability authority. Do not mistake source qualification for the already published artifact's release label.

Current code and documentation identities

• Exact shipped v1.6.42 source tag: 3db2ca98d144ab4b456e8dc1db9af18ad7a948f2 (successful release workflow https://github.com/masarray/arsas/actions/runs/40151975055).
• RC source and tagged release have identical Git tree bc278a81a96517a979a161d2889493047cf204e0. All latest IED Browser, typed MMS Write/control, Smart IP, GOOSE/SV, FAT, IO and COMTRADE functionality is preserved.
• Documentation/website Public Stable sync landed on main via PR docs(site): reconcile README, downloads and bilingual landing to actual v1.6.42 stable #742, with follow-up screenshot label and CHANGELOG fixes (main baseline 4e6a4c747f57693b02279b349e824c4bfdb6c119).
• PR source integration commit 976d54cb5876a3817de827b768c82b61aa845bd5 has two Git parents: exact integrated RC aa0b3d9dcb611b58738491285d06d7991843adb7 and current stable-documentation main 4e6a4c747f57693b02279b349e824c4bfdb6c119. Source tree is unchanged from shipped v1.6.42 except the 12 documentation/site/updater evidence paths introduced on main.
• Follow-up source commit updates .release/windows.json from 1.6.41 to 1.6.42 to match RC VERSION, Directory.Build.props and ArIED61850Tester.csproj; not a rebuild or republish. .release/published.json correctly describes existing immutable v1.6.42 assets. No release tag mutation, no binary changes.

Operator validation and engineering boundary

• Operator accepted the 5 monitored endpoint, 183 initial value, 14 static RCB path diagnostic plus tested SG7_BCU/GR_X_7SX85 typed ctlModel Read/Write/readback and SBO Enhanced commands. These are scoped acceptance observations, not universal physical device certification. Some endpoints may be same-host simulators.
• Later report capture: 432 frames, 38 continuity observations and 9 IED-reported BufOvfl notices. These do not prove 38 lost events or guarantee lossless buffered SOE.
• ARIEC61850 engine f75e323e9f8443aa20e0a3818a1d5d77255b2c0f remains DRAFT code-trial from engine PR Fix responsive top bar navigation clipping #192, not independently qualified physical R7/R10 authority. Existing SCL R7 and Interoperability Reference Guard checks in pull-request context deliberately FAIL until qualified provenance and device evidence exist; push-context success does not waive this.

Merge decision

• Reconcile current main Public Stable website/README/CHANGELOG/metadata with the exact RC implementation without reverting any code or documentation.
• Align staged Windows version manifest to 1.6.42 with correct source identity, without packaging a new release.
• Demonstrate physical R7/R10 and requisite replay/EntryID/SqNum, reconnect, trusted-SCL and safety evidence on real hardware, and qualify or separately gate trial-engine authority.
• Achieve green pull-request SCL R7 and Interoperability Reference Guard by fixing the real cause (not weakening guards, spoofing evidence or silently repinning engine).
• Re-run full integration/Windows testing and review exact main merge head.

Keep PR DRAFT; do NOT merge to main before meaningful physical gates pass. This policy protects critical field-control semantics while preserving the already-published public application binaries unchanged. Related earlier validation: PR #740, #741 and archived comments in this PR.

[masarray]

PUBLIC PREVIEW v1.6.42 — RELEASED, VERIFIED (2026-10-25)

Published: https://github.com/masarray/arsas/releases/tag/v1.6.42
One-run release workflow: https://github.com/masarray/arsas/actions/runs/40151975055 — workflow_dispatch, branch release/arsas-integrated-rc0-20261024, exact SHA 3db2ca98d144ab4b456e8dc1db9af18ad7a948f2, attempt 1, SUCCESS.

GitHub release metadata VERIFIED: tag v1.6.42, published 2026-10-25T01:55:02Z, prerelease=true, draft=false, target exact source SHA above. /releases/latest remains v1.6.41 stable; main commit remains 2bdb6c0f54ead30af629377da968a79974626ce1 unchanged. No main merge, no false physical R7/R10 qualification.

Release assets uploaded and independently exposed by GitHub:

• ARSAS-Windows-x64-Portable.exe 79,355,282 bytes SHA256 281dc1d5f649fb28d3cbbeec858639103644a43296a225e44378fe9de4762669
• ARSAS-Windows-x64-Setup.exe 57,817,466 bytes SHA256 71a3d6fdb01c303fdd85db61fb529f714d81ef6916140acb69c158a47206a173
• ARSAS-Windows-x64-SHA256SUMS.txt, ARSAS-Windows-x64-SBOM.spdx.json, ARSAS-Windows-x64-PROVENANCE.json all present.

One-run package, installer and supply-chain checks all SUCCESS: 2,987/2,987 app regression, native bridge integration PASS, portable smoke, installer build, silent install/uninstall smoke, checksums, SPDX SBOM, provenance, verification, installer digest attestation, portable SBOM attestation, artifact upload, and GitHub Release creation. Signed portable attestation https://github.com/masarray/arsas/attestations/59519828. Record verified release publication stable ledger step correctly skipped on this prerelease.

Operator-scope limitations explicitly present in public release notes: five-endpoint observed 183/183 initial values and 14/14 RCB plans, but 36 continuity findings and nine BufOvfl notices; lossless SOE and blanket physical/IEC certification NOT claimed; current ARIEC engine #192 trial remains draft. Public prerelease is complete; stable promotion remains separate blocked qualification, not an unfinished preview publication.

Follow-up: Validate installation on field machine against release SHA256; keep #727/#729 DRAFT while independent physical certification and stable promotion are pending. Do NOT repackage the same tag or modify already-published assets in-place.

[masarray]

Operator stable-release go-ahead received — updated 09:24 WIB 2026-10-25 diagnostic review

The operator has now approved the application overall as good and requested a public STABLE release. This is an important explicit product-scope acceptance, but is NOT a statement that the separate R7/R10 hardware qualification, SOE continuity/lossless certification, or actual hardware provenance checks have passed.

Authoritative production binary currently installed in diagnostic: ARSAS 1.6.42+3db2ca98d144ab4b456e8dc1db9af18ad7a948f2; engine f75e323e9f8443aa20e0a3818a1d5d77255b2c0f trial PR #192. 2026-10-25 09:24:38 WIB snapshot, 5/5 endpoints connected and monitoring, 183/183 static initial displayed, 14/14 exact static RCB routes with first ingress, no initial MMS errors. All are bounded observed metrics, not full SCL schema equivalence or hardware certification.

NEW 09:24 continuity summary (later than original 07:20 FAT):

• BCUGE: 10 frames, 6 findings, 3 IED BufOvfl
• GR_X_7SX85: 3 frames, 1 finding, 0 BufOvfl
• SG7_BCU: 396 frames, 27 findings, 4 BufOvfl; several sequence discontinuities recur ~116 seconds after start (explicitly marked GIreason=NO). This is not just startup noise.
• SIEBCU: 13 frames, 3 findings, 1 BufOvfl
• ZIVAVR: 10 frames, 1 finding, 1 BufOvfl
• TOTAL: 432 frames, 38 observational continuity findings, 9 IED-advertised BufOvfl. This newer 38 count supersedes 36 in previous 07:20 preview notes for this capture; these are not 38 proven lost SOE events, but cannot support blanket lossless buffered SOE claims. Diagnostic warns target IP aliases are assigned to the host; some sessions may be same-host simulators and must not be claimed as five independent physical relays.
• No open SCL/IID parity captured; event replay, SCL-to-live identity, full R7/R10 acceptance, and independent physical device verification are still pending.

Release status checked live: v1.6.42 is published PUBLIC PRERELEASE (draft=false, prerelease=true) on SHA 3db2ca98..., with successful Windows installer/portable, 2987/2987 app tests, native, smoke, checksums, SBOM and Sigstore attestations. main remains 2bdb6c0f54ead30af629377da968a79974626ce1, latest STABLE remains v1.6.41, and promotion PR #729 is DRAFT and shows mergeable_state=dirty against current main. Engine lock explicitly describes the HEAD as DRAFT TRIAL ONLY and physicalQualificationRequired=true. Do not alter these metadata to pretend that a physical qualification happened.

Fastest professional STABLE promotion work items (one gate, not added workflow maze):

1. Reconcile latest main into accepted RC without undoing GUI/SSOT/Control/Write, clear PR P0: reconcile v1.6.42 shipped source + stable docs to main (physical R7/R10 gates pending) #729 merge conflicts and verify identical intended source; freeze candidate binary identity.
2. Resolve/classify SG7 late non-GI SqNum resets, IED BufOvfl and buffered replay/EntryID with independent on-wire captures and event causality. Re-run only target-specific tests on that source/engine and record physical-source identifiers rather than fabricated PASS metadata.
3. Qualify remaining R7/R10 and Reference Guard physical authority without weakening existing guards. Explicit operator application acceptance is recorded but does not automatically waive these independent checks.
4. After qualified, promote stable using the tested immutable Windows package and verify release metadata/latest stable, checksums, SBOM/provenance, install/uninstall; do not overwrite existing published v1.6.42 preview assets in-place without a verified release promotion path.

No stable GitHub Release has been published by this note.

[masarray]

Stable promotion Git reconciliation complete (2026-10-25)

Operator explicitly approved ARSAS overall for public STABLE promotion; this is product usability/scope acceptance, not automatically lossless SOE / physical R7/R10 proof.

Mainline conflict eliminated without modifying any source code: merge-ancestry commit aa0b3d9dcb611b58738491285d06d7991843adb7, parents (1) operator-tested Public Prerelease source 3db2ca98d144ab4b456e8dc1db9af18ad7a948f2, (2) current main 2bdb6c0f54ead30af629377da968a79974626ce1; tree SHA unchanged bc278a81a96517a979a161d2889493047cf204e0. Main's only newer commit was SV first-received/M3.H #734: M3.H test blob is identical; RC SmvViewer and SV regression retain enhanced accepted live streaming capability. GitHub PR #729 now mergeable=true (previously mergeable=false, dirty), though mergeable_state=unstable pending gating. main and existing v1.6.42 preview remain untouched.

Latest installed 09:24:38 WIB v1.6.42 diagnostic from operator shows 5/5 connected/monitoring, 183/183 displayed, 14/14 RCB paths. Compared with earlier 07:20 proof, this capture shows 38 continuity findings and 9 BufOvfl across 432 frames; SG7 non-GI SqNum anomalies recur after ~116 seconds. Some endpoints are same-host aliases. Do not claim unconditional SOE integrity or certify five hardware relays.

Current release decision: Public Preview is already complete; stable promotion cannot be marked PASS while engine lock remains explicit trial non-physical and R7/R10/Reference Guard mainline physical authority is unresolved. No fake gate override, stable tag, false field result or claim of merger. Next validate exact merge source CI, classify continuity with on-wire proofs, qualify engine authority, then promote source-stable release with correct SHA256/SBOM/provenance. Keep one existing Windows release workflow; do not add duplicate release pipelines.

[masarray]

Public Stable decision reconciliation — 2026-10-25

Operator has explicitly approved the overall application UX and requested PUBLIC STABLE, with no additional workflow maze. This is product acceptance, not physical R7/R10 certification or lossless SOE acceptance.

Verified current source: release/arsas-integrated-rc0-20261024 at aa0b3d9dcb611b58738491285d06d7991843adb7 is 2 commits ahead of public preview 3db2ca98d144ab4b456e8dc1db9af18ad7a948f2, zero changed files (merge ancestry only). Public v1.6.42 remains attached to original 3db2ca98, 5 immutable assets; workflow #40151975055 SUCCESS. main remains 2bdb6c0f54ead30af629377da968a79974626ce1; GitHub latest stable remains v1.6.41.

New reconciled-head PR CI: Windows Build #40154761355 SUCCESS; IO/SV/COMTRADE/P0/P1 SUCCESS. But R7 PR #40154761424 FAIL at Resolve current engine while preserving R7/R10 authority with Unqualified typed-control staging pin cannot replace physical R7/R10 authority; Reference Guard PR #40154761392 FAIL at Verify single convergence authority with Smart interoperability code-trial authority is malformed or was silently promoted. These are intentional promotion-authority failures, not Build/packaging regression; do not fake acceptance by changing a trial flag, engine pin, or guard policy.

Newest operator trace (09:24 WIB): 432 report frames, 38 observational continuity findings, 9 IED-advertised BufOvfl (supersedes earlier 36 for this sample); SG7 some non-GI SqNum anomalies recur ~116s after start. These are not 38 proven lost events, nor proof of lossless SOE. Independent physical hardware and SCL/reconnect/replay evidence remain incomplete.

Keep the release path SIMPLE: retain the ONE existing .github/workflows/release-windows.yml. Its source explicitly forces manual dispatch from any non-main branch to prerelease, even when prerelease=false; the exact-tag immutable release code refuses changing already-published binaries. Therefore do not rerun v1.6.42 or try to flip its channel from the RC branch. When remaining acceptance is genuinely satisfied, merge the reconciled source to main through this PR and publish a new versioned stable with the existing single workflow, checked asset SHA256/SBOM/provenance and explicit limitations. Do not duplicate workflows, undo accepted GUI/SSOT/Control/Write/SV, pretend all 5 endpoints are physical relays, or rewrite historical stable evidence.

Status: stable publication NOT performed; current preview remains public and usable, but promotion is held on documented release-authority/evidence blockers.

[masarray]

PUBLIC STABLE promotion request — operator signoff received, physical authority still NOT established (2026-10-25 09:25 WIB)

Operator explicitly requests PUBLIC STABLE following hands-on installation/use of published v1.6.42 preview. Honor approval as approval to proceed with stable-readiness work, NOT as falsified evidence of R7/R10 physical acceptance or lossless buffered SOE.

Candidate identities: Published v1.6.42 preview source 3db2ca98d144ab4b456e8dc1db9af18ad7a948f2 and latest RC HEAD aa0b3d9dcb611b58738491285d06d7991843adb7 share the exact git TREE bc278a81a96517a979a161d2889493047cf204e0. So recent mainline ancestry reconciliation did not change product binary sources. The release remains public prerelease=true, and latest stable remains v1.6.41. Do not republish or overwrite immutable tag/assets; no extra workflow is needed.

Fresh operator diagnostic provided privately at 2026-10-25 09:24:38 WIB: ARSAS v1.6.42+3db2..., engine +f75e..., 5/5 IED connected and monitoring; 183/183 initial visible static members; all 14/14 planned RCB ingress targets routed; 432 observed report frames. 38 continuity findings and 9 IED-advertised BufOvfl (BCUGE 6/3, GR_X 1/0, SG7 27/4, SIEBCU 3/1, ZIVAVR 1/1). This is newer than the repository's earlier 427-frame / 36-finding / 9-overflow observation. Source log privately supplied, digest SHA-256 74c774635ccb1a82f81c40181bffac48b583146a971eb85046ea065a3a4bbc45; raw local network details and command hex not re-uploaded publicly. Report findings do NOT by themselves count lost events, but neither do they certify buffer/SOE replay parity. SG7 reported multiple no-GI/non-startup sequence resets near +116 seconds on three BRCBs, and the last observed frame of those streams was anomalous, requiring chronology/capture and classification. Several target IPv4 endpoints are also locally assigned aliases on the operator's test PC; diagnostic explicitly warns possible same-host simulation, so this is not proof of five independent physical relays.

Actual GitHub PR #729 promotion gates still FAIL on current HEAD:

• R7 #40154761424 refuses trial engine f75e323...: Unqualified typed-control staging pin cannot replace physical R7/R10 authority.
• Reference Guard #40154761392 refuses silently promoting code-trial interoperability to physical authority.
• PR RC0 INTEGRATED — GUI P0/P1/P2 + SSOT/Write + main SV; physical release gates open #727 integration Build #40154761355 PASS; app tests/functional confidence are good, but physically scoped stable gate is not equivalent to CI.
• Engine PR Fix responsive top bar navigation clipping #192 remains DRAFT.

Single-track next-action check (no new workflow complexity):

1. Preserve exact source tree/engine pin and immutable v1.6.42 prerelease. Do not merge main or change release metadata to pretend physical qualification.
2. Collect/reference live real-device capture on exact binary for SG7 three BRCB late SqNum reset + EntryID + GI/OptFlds/BufOvfl and one disconnected/reconnect report-replay trial; classify whether RCB reset is equipment behavior, expected test initialization, or actual event-loss risk. Record equipment/simulator scope; do not claim zero-loss without independent expected-event comparison.
3. Explicit R7/R10 acceptance of engine PR Fix responsive top bar navigation clipping #192 typed controls under physically qualified scope or separately agreed, test-backed product capability limits; maintainer approval must include proof and preserve fail-closed guard semantics.
4. Only then prepare a new clean stable version/tag, execute the existing SINGLE release-windows.yml with correct version and prerelease=false from MAIN on an accepted exact tree, verify installer/portable/SBOM/provenance, and confirm /releases/latest moves only upon successful accepted release. Do not mutate v1.6.42 release in place or let a staging-branch manual dispatch silently claim stable.

Status: User's release direction acknowledged. NOT PUBLISHED AS STABLE; physical-gate blockers are real and remain visible. All code and CI achievements must be retained.

[masarray]

Operator decision (25 October 2026): public stable release is ACCEPTED; this draft is NOT a release blocker

Operator explicitly confirmed that ARSAS v1.6.42 is already working, IED connection and functional QA evidence have been provided, and screenshots are approved for public use. The operator rejects repeating broad QA or tying documentation/public stable release to an open source-promotion draft.

To make the scope unambiguous:

1. Published GitHub stable v1.6.42 installer and portable single EXE remain accepted and live. There is NO request for another public release, extra operator connection test, certificate, or rollback.
2. This draft P0: reconcile v1.6.42 shipped source + stable docs to main (physical R7/R10 gates pending) #729 concerns the separate technical goal of reconciling the exact shipped source with main. It is not a criterion for whether v1.6.42 is publicly released or works on the operator-tested endpoints.
3. Existing R7/R10 / interoperability code-trial authority checks are about specific physical control-path, trusted SCL, replay/EntryID/SqNum and safety qualification, not simple MMS connectivity. The user's successful connection is accepted as connection evidence; it is not silently relabeled as a completed independent physical-control qualification.
4. Do not merge an unqualified trial engine, repoint release tag or change immutable stable assets just to show a green checkbox. Maintain draft status until independent scope-specific evidence exists. Do not tell the operator this blocks the already approved product release.

This comment reconciles acceptance language without invalidating the user's previously supplied test data.

[masarray]

Repository cleanup audit — exact release/tree reconciliation (2026-10-26)

Decision: KEEP OPEN/DRAFT. No source promotion, release modification, or gate bypass. This is an evidence correction to the PR description, not a request to retest the already operator-accepted stable Windows v1.6.42.

Live immutable refs:

• Published stable tag v1.6.42: commit 3db2ca98d144ab4b456e8dc1db9af18ad7a948f2, tree bc278a81a96517a979a161d2889493047cf204e0.
• Current shared P0: reconcile v1.6.42 shipped source + stable docs to main (physical R7/R10 gates pending) #729/RC0 INTEGRATED — GUI P0/P1/P2 + SSOT/Write + main SV; physical release gates open #727 head: 817c3910e8223fa2ac531a4426ada5d2d4bbbe50, tree 257abcdb4f7b184458c4d24ee9e61c3f043c97bb.
• GitHub compare stable-tag...current-RC-head: ahead=8, behind=0, 13 changed files, exclusively release docs/web metadata/workflow paths (.release/published.json, .release/windows.json, landing/site.json, README.md, CHANGELOG.md, .github/workflows/sync-release-documentation.yml, related site templates/metadata). No app code paths appear in that comparison.

Correction: Existing PR body sentence claiming current RC source and release tag have identical Git tree bc278a... is now stale. The tagged application source is an ancestor of the current RC commit, but the entire Git tree is not byte-identical. Preserve both truths: unchanged accepted published artifact, and later post-release metadata/docs changes that require explicit accounting before source promotion. Do not mutate tags, artifacts or checksums to create cosmetic equality.

Safety / gate: Engine PR #192 trial provenance and R7/R10/control FAT gates remain independent and cannot be waived by passing Windows build or by this metadata comparison. The audit also found mixed pass/fail workflow history at the shared RC SHA; evaluate the correct pull-request-context required gates on current head, not green push-context alone.

Next action: Before contemplating #729 merge, reconcile the 13 post-tag file changes explicitly against current main, preserve the already-published release manifest identity, and qualify the engine / physical / PR CI gates as listed in #729. Keep #727/#726/#716/#688 and active UI/SSOT work protected.