Merge fix/review-2026-10-21 to origin main and run the windev Worker gate #169

New Issue

2026-11-06T15:24:54-05:00

dohertj2 commented

2026-11-06 15:24:54 -05:00

The 2026-10-21 system-review remediation (streams A-M) is merged into local main (8c0a51d, 2026-10-27) but origin/main is still at 488649c — local main is 61 commits ahead. The integration branch origin/fix/review-2026-10-21 (b2bce9c) is pushed, so only the merge to main is missing.

Remaining work:

Merge fix/review-2026-10-21 into main on the remote (PR or push of the existing merge) and confirm CI is green.
Run windev-worker-ci.ps1 -Mode test (x86 Worker build + Worker.Tests) at the merged SHA, as the tracker requires for stream E worker changes, and record the result in the tracker.
After merge, close #167 and #168: both are fixed on this branch (R21-S6 unattached-session grace, R21-S2 sweep-close gauge decrement) but remain open on Gitea because the fix is not on origin/main.

Source: archreview/2026-10-21/remediation/00-tracking.md (Execution model + stream E note "After merge, run windev-worker-ci.ps1 -Mode test").
Evidence: git status -sb -> main...origin/main [ahead 61]; git log origin/fix/review-2026-10-21..main -> only the merge commit 8c0a51d.
Related: #167, #168, #150.

The 2026-10-21 system-review remediation (streams A-M) is merged into local `main` (`8c0a51d`, 2026-10-27) but **origin/main is still at `488649c`** — local main is 61 commits ahead. The integration branch `origin/fix/review-2026-10-21` (`b2bce9c`) is pushed, so only the merge to main is missing. Remaining work: - Merge `fix/review-2026-10-21` into `main` on the remote (PR or push of the existing merge) and confirm CI is green. - Run `windev-worker-ci.ps1 -Mode test` (x86 Worker build + Worker.Tests) at the merged SHA, as the tracker requires for stream E worker changes, and record the result in the tracker. - After merge, close #167 and #168: both are fixed on this branch (`R21-S6` unattached-session grace, `R21-S2` sweep-close gauge decrement) but remain open on Gitea because the fix is not on origin/main. Source: `archreview/2026-10-21/remediation/00-tracking.md` (Execution model + stream E note "After merge, run windev-worker-ci.ps1 -Mode test"). Evidence: `git status -sb` -> `main...origin/main [ahead 61]`; `git log origin/fix/review-2026-10-21..main` -> only the merge commit `8c0a51d`. Related: #167, #168, #150.

dohertj2 added the area:gateway area:worker priority:p1 type:deploy-gate migrated labels 2026-11-06 15:24:54 -05:00

dohertj2 commented

2026-11-06 15:36:53 -05:00

2026-11-06: local main (incl. the merged fix/review-2026-10-21 work) pushed to origin/main — now at 5a23577. Remaining: run the x86 Worker tests on windev against this tip, then close #167/#168 if green.

2026-11-06: local main (incl. the merged fix/review-2026-10-21 work) pushed to origin/main — now at 5a23577. Remaining: run the x86 Worker tests on windev against this tip, then close #167/#168 if green.

dohertj2 commented

2026-11-06 15:59:04 -05:00

Gate run, and it passed.

The merge part is done: the review remediation was already on origin/main (5a23577) when this started.
origin/main is now 777fef9. On top of 5a23577 it adds #193 (9354d69) and #187 (34a7453, 777fef9).

windev Worker gate at 777fef9:

scripts/ci/windev-worker-ci.ps1 -Mode test in C:\build\mxaccessgw-ci (the push-triggered run): SUCCESS mode=test sha=777fef9cadd1bb6e94d5639d79f49102ea8d02b2 (x86 Worker build + Worker.Tests). It also passed at 34a7453 and 5a23577.
Isolated clone C:\build\mxgw-issues-187 at the same SHA:
- dotnet build src/ZB.MOM.WW.MxGateway.Worker/ZB.MOM.WW.MxGateway.Worker.csproj -c Release -p:Platform=x86: 0 warnings, 0 errors.
- dotnet test src/ZB.MOM.WW.MxGateway.Worker.Tests/... -c Release -p:Platform=x86: 609 passed, 12 skipped, 0 failed (net48).
- dotnet test src/ZB.MOM.WW.MxGateway.Tests/... -c Release (gateway suite on Windows): 1349 passed, 0 failed.
- Live MXAccess smoke, GatewaySession_WithLiveWorker_RegistersAdvisesStreamsDataAndCloses only (read-only): passed.
macOS gateway suite: 1347 passed / 2 skipped (Windows-only).

The MxAccessGw NSSM service on windev was not touched; it was still Running afterwards.

I'm closing #167 and #168 at the same time. Both fixes are on origin/main and covered by tests (see the comments there).

Gate run, and it passed. - The merge part is done: the review remediation was already on origin/main (5a23577) when this started. - origin/main is now **777fef9**. On top of 5a23577 it adds #193 (9354d69) and #187 (34a7453, 777fef9). windev Worker gate at 777fef9: - `scripts/ci/windev-worker-ci.ps1 -Mode test` in `C:\build\mxaccessgw-ci` (the push-triggered run): `SUCCESS mode=test sha=777fef9cadd1bb6e94d5639d79f49102ea8d02b2` (x86 Worker build + Worker.Tests). It also passed at 34a7453 and 5a23577. - Isolated clone `C:\build\mxgw-issues-187` at the same SHA: - `dotnet build src/ZB.MOM.WW.MxGateway.Worker/ZB.MOM.WW.MxGateway.Worker.csproj -c Release -p:Platform=x86`: 0 warnings, 0 errors. - `dotnet test src/ZB.MOM.WW.MxGateway.Worker.Tests/... -c Release -p:Platform=x86`: **609 passed, 12 skipped, 0 failed** (net48). - `dotnet test src/ZB.MOM.WW.MxGateway.Tests/... -c Release` (gateway suite on Windows): **1349 passed, 0 failed**. - Live MXAccess smoke, `GatewaySession_WithLiveWorker_RegistersAdvisesStreamsDataAndCloses` only (read-only): passed. - macOS gateway suite: 1347 passed / 2 skipped (Windows-only). The MxAccessGw NSSM service on windev was not touched; it was still Running afterwards. I'm closing #167 and #168 at the same time. Both fixes are on origin/main and covered by tests (see the comments there).

dohertj2 closed this issue

2026-11-06 15:59:04 -05:00

dohertj2 referenced this issue

2026-11-06 15:59:04 -05:00

Sessions that never call StreamEvents hold a worker for the full lease (1800 s); no lease renewal #167

dohertj2 referenced this issue

2026-11-06 15:59:05 -05:00

mxgateway.sessions.open gauge never decremented for sessions closed by the timed sweep #168

Sign in to join this conversation.