test(live): capture OperationComplete trigger, subtag Clear and status rows on windev (#210, #211)
ci / nightly-windev (push) Has been skipped
ci / windows-x86 (push) Successful in 2m23s
ci / java (push) Failing after 7m49s
ci / portable (push) Failing after 16m36s

Live MXAccess captures against the windev DEV Galaxy, recorded in
docs/LiveCaptures-2026-11-13.md:

- OperationComplete follows a Suspend/Activate whose synchronous MxStatus is
  Pending (advised, resolved item); one Ok row from RespondingLmx. Pinned by
  the new OperationCompleteLiveProbeTests live fact, with an
  unresolved-reference negative control. (#211 §3.1)
- No public COM event carries completion-only status bytes. All 71
  completions were typed MXSTATUS_PROXY[], so they stay raw permanently.
  (#211 §3.3)
- Subtag-mode Clear is drivable: the TestAlarm UDAs are MxSecurityOperate,
  so AuthenticateUser + WriteSecured toggles them. The new self-driven
  LiveSubtag_SelfDrivenRaiseThenClear_SynthesizesClear fact asserts the
  UnackAlm -> UnackRtn Clear (mutation-checked). (#210 §3.7)
- The B8 buffered "NoData bootstrap" is a ConfigurationError resolution
  failure (detail 6 / 1002). The DEV Galaxy has no buffered attribute, so a
  multi-sample capture needs IDE work. provider_switches was not driven. (#210)
- Status-row catalogue for lmxopcua#872: no Warning row seen, and Pending
  only on the synchronous Suspend/Activate return.

Claude-Session: https://claude.ai/code/session_01GT67ACmpMR7b9LGRPLijMi
This commit is contained in:
Joseph Doherty
2026-11-12 23:34:46 -05:00
parent 4890a3d16e
commit 9e4be25807
7 changed files with 584 additions and 24 deletions
+4 -4
View File
@@ -75,13 +75,13 @@ These are documented, deliberate, and mostly enforced. Listed so the deferred su
Intentionally incomplete, awaiting a live capture or an AVEVA fix; raw payload/metadata is preserved in the meantime (no synthesis).
- 🟡 **3.1 `OperationComplete` native trigger condition unknown** → #211 — modeled and emitted only from the real event (no synthesis), but the runtime condition that fires it isn't captured. `docs/DesignDecisions.md:280-289`, `gateway.md:1094`, `docs/MxAccessWorkerInstanceDesign.md:341,366`.
- 🟡 **3.2 `OnBufferedDataChange` multi-sample conversion unvalidated — STILL OPEN (residual after B8)** → #210 — `AddBufferedItem`/`SetBufferedUpdateInterval` are now implemented and live-confirmed (§1.1), and the live B8 test (`f7ada90`) confirms the worker receives and cleanly converts the empty `NoData` bootstrap `OnBufferedDataChange` (no crash, no dropped payload). But the rig's object logic does not drive a multi-sample buffered batch on demand (same limitation as the alarm rig), so a real parallel quality/timestamp sample array (length > 1) has never been observed live — it is exercised only by the B-bundle unit tests against a fake `IMxAccessServer`. Re-run `GatewaySession_WithLiveWorker_BufferedItem_*` against a fast-changing simulated tag to close this. `docs/DesignDecisions.md:291-297`.
- 🟡 **3.3 Completion-only status → `MXSTATUS_PROXY[]` mapping unproven** → #211 — completion-only operation-status bytes are kept as raw diagnostic metadata until the analysis proves an exact mapping. `docs/DesignDecisions.md:299-306`.
- ✅ **3.1 `OperationComplete` native trigger — CAPTURED 2026-11-13** (#211): it follows a `Suspend`/`Activate` whose synchronous status is Pending (advised, resolved item); see `docs/LiveCaptures-2026-11-13.md` §1 and `OperationCompleteLiveProbeTests`. Original entry: modeled and emitted only from the real event (no synthesis), but the runtime condition that fires it isn't captured. `docs/DesignDecisions.md:280-289`, `gateway.md:1094`, `docs/MxAccessWorkerInstanceDesign.md:341,366`.
- 🟡 **3.2 `OnBufferedDataChange` multi-sample conversion unvalidated — STILL OPEN (residual after B8)** → #210 — **2026-11-13:** the B8 "empty NoData bootstrap" is a ConfigurationError resolution failure (detail 6 / 1002), and no deployed attribute in the DEV Galaxy has a buffer, so a capture needs a buffered field attribute deployed (`docs/LiveCaptures-2026-11-13.md` §5). Original entry: `AddBufferedItem`/`SetBufferedUpdateInterval` are now implemented and live-confirmed (§1.1), and the live B8 test (`f7ada90`) confirms the worker receives and cleanly converts the empty `NoData` bootstrap `OnBufferedDataChange` (no crash, no dropped payload). But the rig's object logic does not drive a multi-sample buffered batch on demand (same limitation as the alarm rig), so a real parallel quality/timestamp sample array (length > 1) has never been observed live — it is exercised only by the B-bundle unit tests against a fake `IMxAccessServer`. Re-run `GatewaySession_WithLiveWorker_BufferedItem_*` against a fast-changing simulated tag to close this. `docs/DesignDecisions.md:291-297`.
- ✅ **3.3 Completion-only status — SETTLED 2026-11-13** (#211): no public COM event delivers completion-only bytes. All 71 captured completions were typed `MXSTATUS_PROXY[]`, so the bytes stay raw permanently (`docs/LiveCaptures-2026-11-13.md` §2). Original entry: completion-only operation-status bytes are kept as raw diagnostic metadata until the analysis proves an exact mapping. `docs/DesignDecisions.md:299-306`.
- 🟡 **3.4 `AlarmAckByGUID` is `E_NOTIMPL` on this AVEVA build (historical; note corrected 2026-10-22)** — it throws `NotImplementedException`, but the worker no longer calls it. `MxCommandKind.AcknowledgeAlarm` has live dispatch (`Worker/MxAccess/MxAccessCommandExecutor.cs` `ExecuteAcknowledgeAlarm`, ~895-936): it parses `alarm_guid` and calls `alarmCommandHandler.Acknowledge(guid, comment, AlarmAckOperator)`, which acknowledges by writing `AckMsg` as the authenticated operator. A GUID ack therefore works; only the vendor `AlarmAckByGUID` call is dead. `docs/AlarmClientDiscovery.md:750-763`.
- 🟡 **3.5 8-arg `AlarmAckByName` v2 is a vendor stub (returns -55)** — no longer on the ack path (acks write `AckMsg`); `operator_full_name` is carried for forward-compat only and is not propagated (see 1.4). `docs/AlarmClientDiscovery.md:743-748`.
- 🟡 **3.6 Subtag degraded-mode fidelity limits** — `category`, `description`, `alarm_type_name`, operator fields, and `retrigger` are not populated/synthesized in subtag fallback (no subtag exists for them). Documented, by design. `docs/AlarmClientDiscovery.md:913-931`, `docs/plans/2026-06-13-alarm-subtag-fallback-design.md:292-298`.
- 🟡 **3.7 Subtag `Clear` transition unvalidated live** → #210 — Raise/Ack/AckMsg are live-confirmed; Clear is externally undrivable on the rig (object logic owns alarm state). Environmental, not code. (`project_alarm_subtag_fallback`, `project_rig_alarms_object_driven`.)
- ✅ **3.7 Subtag `Clear` transition — CAPTURED 2026-11-13** (#210): driven with `WriteSecured` and pinned by `AlarmSubtagLiveSmokeTests.LiveSubtag_SelfDrivenRaiseThenClear_SynthesizesClear` (`docs/LiveCaptures-2026-11-13.md` §4). Original entry: Raise/Ack/AckMsg are live-confirmed; Clear is externally undrivable on the rig (object logic owns alarm state). Environmental, not code. (`project_alarm_subtag_fallback`, `project_rig_alarms_object_driven`.)
---