Replace the bare-TCP-only AbCipDriverProbe with a two-phase check:
Phase 1 keeps the existing TCP preflight; Phase 2 initialises a
LibplctagTagRuntime against the first device to open a real EIP session
and CIP Forward Open, so a live-but-rejecting CIP endpoint reads red
instead of a false-positive green.
Status mapping: ErrorNotFound / ErrorNoMatch / ErrorBadDevice → reachable
(controller answered CIP, probe tag absent); ErrorTimeout / ErrorBadConnection
/ ErrorBadGateway / ErrorWinsock / ErrorOpen / ErrorClose / ErrorRead /
ErrorWrite / ErrorBadReply / ErrorRemoteErr / ErrorPartial / ErrorAbort →
handshake failed. LibPlcTagException message text is used as a secondary
signal for the reachable-exception path. All other statuses default to
handshake-failed (conservative).
Add AbCipDriverProbeTests: invalid JSON, no devices, malformed host address,
closed-port TCP rejection, and black-hole timeout — all offline-determinable.
Happy path + CIP-error path covered live against the CIP sim.
Replace the bare TCP-connect return in OpcUaClientDriverProbe with a real
OPC UA GetEndpoints discovery handshake (mirroring SelectMatchingEndpointAsync
in the driver). TCP preflight still fast-fails closed ports; the handshake
confirms the remote is actually an OPC UA server, so a live-but-rejecting
non-OPC-UA process now reads RED instead of a false-healthy green.
Replace bare TCP-connect with a two-phase probe: Phase 1 keeps the
existing SocketException / timeout / generic preflight paths unchanged;
Phase 2 runs Plc.OpenAsync (COTP CR/CC + S7 setup-communication) so a
device that accepts TCP but is not an S7 PLC reads red instead of green.
A linked CTS distinguishes caller cancellation ("timed out") from the
S7netplus internal read-timeout OCE ("handshake failed: timed out").
Replace the bare TCP-connect probe in ModbusDriverProbe with a two-phase
check: TCP connect via ModbusTcpTransport (keeps the same SocketException /
timeout / generic error paths and messages), then a one-shot FC03 Read
Holding Registers (qty 1 @ addr 0). A normal response → Ok=true "Modbus
FC03 OK"; a Modbus exception PDU → Ok=true "Modbus FC03 OK (device
returned exception PDU)"; any other failure after TCP succeeds → Ok=false
"Reachable at host:port but Modbus FC03 handshake failed: …".
Add ModbusDriverProbeTests (6 tests) covering invalid JSON, missing
host/port, closed port, TCP-accept-then-close, canned MBAP happy path,
and Modbus exception PDU path. All 277 Modbus tests green.
Add IFocasClientFactory.EnsureUsable() — a config-time probe called by
FocasDriver.InitializeAsync before any background loops start. The
UnimplementedFocasClientFactory throws NotSupportedException immediately
(faulting the driver at init), eliminating the footgun where a driver on
the 'unimplemented' backend appeared Healthy then failed every read/write/
subscribe silently. WireFocasClientFactory and FakeFocasClientFactory are
no-ops. Backstop Create() throw remains in place.
Make MapDataType internal, split the combined Int64/UInt64 arm to return
DriverDataType.Int64 and DriverDataType.UInt64 respectively, and remove
the now-stale Driver.Modbus-007 caveat doc block and inline comment.
Add a Theory covering both cases; full suite 271/271 green.
The owning DriverInstanceActor re-subscribes alarms on every Connected
entry (DetachAlarmSource nulls its cached handle on Connected->Reconnecting
without calling UnsubscribeAlarmsAsync), and the driver object + its alarm
projection are reused across every in-place reconnect. Each SubscribeAsync
started a fresh, never-cancelled Task.Run poll loop and added it to _subs,
so N reconnects leaked N concurrent loops all polling the device and all
firing the same raise/clear transitions => duplicate alarm events + CPU/mem
growth.
Mirrors the Galaxy #399 fix (Clear-before-Add) but for live poll loops the
collapse must also CANCEL the superseded loops, not just drop references.
SubscribeAsync now snapshots existing subs under _subsLock, clears _subs,
adds the new sub, starts its loop, then retires each stale sub out-of-band
(RetireAsync: Cancel + await loop + Dispose CTS, fire-and-forget so the new
subscription's return isn't blocked on a poll interval). Snapshot+clear under
the same lock DisposeAsync uses guarantees no double-own / double-dispose.
There is exactly one consumer per driver instance (factory-per-actor), so
retiring all prior subscriptions before starting the new one is faithful.
Regression tests (TDD, fail->pass): subscribe twice then drive one device
raise; assert OnAlarmEvent fires exactly once (was twice with two leaked
loops).
GalaxyDriver's StreamAlarms feed is session-less and survives an in-place
reconnect, so DriverInstanceActor re-subscribed on every Connected re-entry
(after dropping its own cached handle without an Unsubscribe — sync teardown).
The re-subscribe was additive: _alarmSubscriptions.Add grew the list by one
untracked handle per reconnect cycle — a slow unbounded leak. Functionally
harmless (the gate is Count>0 and OnAlarmFeedTransition only reads [0], firing
once regardless), but it accumulated forever.
Fix: SubscribeAlarmsAsync clears the set before adding, collapsing to a single
live handle (under the existing _alarmHandlersLock, atomic w.r.t. the fan-out
reader). There is exactly one consumer per driver instance (factory-per-actor
lifecycle), so replacing the set with the latest handle is faithful. Chosen
over making the actor's sync DetachAlarmSource call UnsubscribeAlarmsAsync
async/fire-and-forget — disproportionate for a minor leak.
Regression test Re_subscribe_collapses_to_a_single_handle_no_accumulation
(TDD-verified: FAILS without the Clear — releasing the latest handle leaves
the feed open because stale handles remain; PASSES with the fix). Galaxy tests
263 pass / 3 skip; Runtime native-alarm 24 pass. Code-reviewed (approved).
Code-review refinement of the live-gw read-back helper: complete a
TaskCompletionSource<double?> from the pump instead of a captured local (explicit
cross-task visibility), pass bufferedUpdateIntervalMs:0 (Advise snapshot needs no
SetBufferedUpdateInterval), and document the Advise->OnDataChange filter. Live re-verified 2/2.
Skip-gated (MXGW_ENDPOINT + GALAXY_MXGW_API_KEY) like GatewayGalaxyAlarmFeedLiveTests.
Covers the two seams the unit suite can only fake in isolation:
- reopen: RecreateAsync + InvalidateHandleCaches re-establish write handles
- write: no-login supervisory write commits + persists (fresh-session read-back)
Live-verified 2/2 against 10.100.0.48:5120 (2026-06-14); skips cleanly in CI.
Add IGalaxyDataWriter.InvalidateHandleCaches() and call it in
GalaxyDriver.ReopenAsync after RecreateAsync succeeds. Prior to this
fix, GatewayGalaxyDataWriter's _itemHandles and _supervisedHandles
dictionaries survived across reconnects, causing the next write to
skip AddItem and AdviseSupervisory against already-dead handles.
Equipment tags resolved at runtime via FocasEquipmentTagParser were not
seeded in _parsedAddressesByTagName so both ReadAsync and WriteAsync
re-parsed the raw TagConfig JSON address string on every hot-path call.
Promoted the field to ConcurrentDictionary (read + write thread safety)
and introduced ResolveParsedAddress(GetOrAdd) so the first call stores
the parse result and all subsequent calls are a cache hit. Authored tags
seeded at InitializeAsync compile and work unchanged.
Live verification on a Windows VM surfaced a crash loop: TcpFrameServer.EnsureListening
assigned _listener = new TcpListener(...) BEFORE calling Start(). When Start() throws —
e.g. the port is in a Windows excluded/reserved range (WSAEACCES) or already in use — the
field was left non-null-but-unstarted, so the `if (_listener is not null) return` guard
permanently skipped re-Start() and every subsequent AcceptTcpClientAsync() threw the
misleading InvalidOperationException "Not listening" → 20 failures → exit 2 → NSSM restart
→ loop. Now _listener is assigned only after Start() succeeds, so a transient bind failure
is retried and a permanent one surfaces the real bind error each iteration. Adds a
regression test that forces a bind conflict and asserts the SocketException persists.
All five suppressed advisories are now resolved at baseline/resolved versions,
so every NuGetAuditSuppress is removed repo-wide:
- System.Security.Cryptography.Xml (GHSA-37gx-xxp4-5rgx / GHSA-w3x6-4m5h-cxqf)
-> fixed by the .NET 10 baseline (10.0.6)
- OPCFoundation Opc.Ua.Core (GHSA-h958-fxgg-g7w3) -> fixed at resolved 1.5.378.106
Two were still live and are now patched via direct security pins:
- OpenTelemetry.Api 1.9.0 -> 1.15.3 (GHSA-g94r-2vxg-569j) pinned in Cluster;
Runtime/ControlPlane/AdminUI + tests inherit via project reference
- Tmds.DBus.Protocol 0.20.0 -> 0.21.3 (GHSA-xrw6-gwf8-vvr9) pinned in Client.UI
Also correct the Historian sidecar runtime comments (x86 -> x64, matching the
csproj PlatformTarget). Solution audit: 0 vulnerable packages; full build clean.
The 2026-04-30 alarm plan banners claimed worker-side native alarm
subscription was blocked on a COM-bitness finding. That's stale: the
mxaccessgw .NET client now has true MxAccess alarm-event support, and a
live StreamAlarms check (+ new Skip-gated GatewayGalaxyAlarmFeedLiveTests
through the lmxopcua consumer) confirms native alarms — operator comment,
category, severity, timestamps — flow end-to-end. Reconcile both plan docs
to reality and add docs/plans/alarms-d1-smoke-artifact.md as the D.1
alarm-source deliverable. Historian-write live smoke + full server->A&C
round-trip remain (Windows parity rig only).
The driver/factory/seed use 'GalaxyMxGateway' (legacy 'Galaxy' was retired),
but the AdminUI editor router, GalaxyDriverPage, address picker, identity
dropdown, the Galaxy browser/probe, and DraftValidator still keyed on 'Galaxy'.
Result: the seeded GalaxyMxGateway driver couldn't be edited ('no editor
registered'), UI-created Galaxy drivers wrote a type with no factory, and a
SystemPlatform-bound GalaxyMxGateway driver failed publish validation.
Align all stragglers to GalaxyMxGateway (+ failing-test-first DraftValidator
coverage). ShouldStub's 'Galaxy' legacy safety-net left intact.
Completes the test side of the in-progress Galaxy-tag workstream:
- Phase7ApplierTests / Phase7ApplierHierarchyTests: supply the now-required
Galaxy-tag args to Phase7Plan / Phase7CompositionResult.
- Add genuine coverage for Phase7Applier.MaterialiseGalaxyTags (folder-per-distinct-path,
variable-per-tag node-id derivation, folder dedupe) + added-Galaxy-tags-trigger-rebuild.
- S7.Cli.Tests: use the project's S7CpuType (CLI option type) instead of S7.Net.CpuType.
Whole solution now builds 0/0; OpcUaServer.Tests 52, S7.Cli.Tests 36 green.
Imports the freshly-rebuilt ZB.MOM.WW.MxGateway.Client + ZB.MOM.WW.MxGateway.Contracts
nupkgs (0.1.0) from /tmp/mxgw-dist. Replaces the vendored libs/ DLLs and the
pre-restructure MxGateway.* namespaces across the runtime Galaxy driver,
Galaxy.Browser, and their tests.
Key changes:
- nuget-packages/ added as a local feed via NuGet.config; .gitignore exempts it
from the *.nupkg rule so the packages are tracked
- Directory.Packages.props pins both packages at 0.1.0
- 4 csprojs swap <Reference HintPath="libs/...dll"/> for <PackageReference/>
- 36 .cs files renamed `using MxGateway.*` -> `using ZB.MOM.WW.MxGateway.*`
- libs/ removed (vendored DLLs + README.md)
GalaxyBrowseSession rewritten around the new lazy API:
- RootAsync calls GalaxyRepositoryClient.BrowseAsync (returns LazyBrowseNodes)
and caches them by TagName instead of bulk-fetching the whole hierarchy
- ExpandAsync looks up the cached LazyBrowseNode and calls its ExpandAsync,
giving true one-wire-call-per-click instead of in-memory parent/child scan
- _byGobjectId + _hasChildrenSet dropped (LazyBrowseNode carries HasChildrenHint)
- AttributesAsync unchanged (already uses DiscoverHierarchyAsync MaxDepth=0)
Tests: Galaxy.Tests 245/245, Galaxy.Browser.Tests 10/10, AdminUI.Tests 66/66.
Pre-existing 12 solution errors unchanged (test sinks + Cli XML comments).
Introduces Driver.S7.Contracts (dependency-free POCO project) and moves
S7DriverOptions / S7ProbeOptions / S7TagDefinition / S7DataType into it.
Adds S7CpuType enum mirroring S7.Net.CpuType exactly (7 values with
explicit integer codes). Runtime S7CpuTypeMap bridges S7CpuType →
S7.Net.CpuType at the single Plc construction site in S7Driver.InitializeAsync.
S7DriverFactoryExtensions and S7CommandBase updated to use S7CpuType; test
files updated to match (S7_1500Profile, S7DriverScaffoldTests). AdminUI can
now reference Driver.S7.Contracts without pulling in S7netplus.
Adds <summary>, <param>, <typeparam>, and <inheritdoc/> tags to public
members surfaced by commentchecker — resolves 5,847 of 5,869 issues
(99.6%) across three /fixdocs passes.
Adds Directory.Packages.props (ManagePackageVersionsCentrally) and
Directory.Build.props (net10.0/nullable/implicit usings/LangVersion latest).
Strips Version attributes from every csproj PackageReference and consolidates
versions into the central file.
Side fixes (necessary to keep the build green on .NET SDK 10.0.105 on macOS):
- Microsoft.CodeAnalysis.CSharp{,.Workspaces}: 5.3.0 -> 5.0.0. The 5.3.0
analyzer DLL references compiler 5.3.0.0 and the local SDK ships compiler
5.0.0.0, producing CS9057 on every project that loaded the Analyzers
output. Master itself was broken on this machine pre-change.
- Server + Server.Tests pin OPCFoundation.NetStandard.Opc.Ua.{Configuration,
Client} to 1.5.374.126 via VersionOverride, matching Opc.Ua.Server's
pin. Mixing 1.5.378.106 Opc.Ua.Core transitively with 1.5.374.126
Opc.Ua.Server breaks CustomNodeManager2 override signatures
(CS0115 on LoadPredefinedNodes/Browse/HistoryRead*) and CS7069 in
the tests. The pin disappears when the legacy Server project is
deleted in Task 56.
- Client.UI + Client.UI.Tests: NuGetAuditSuppress for
GHSA-xrw6-gwf8-vvr9 (Tmds.DBus.Protocol 0.20.0 reaches both projects
transitively from Avalonia.Desktop on Linux/macOS only).
Deviation from the plan: TreatWarningsAsErrors=true is NOT set in
Directory.Build.props because the pre-v2 Admin/Server test projects carry
~240 xUnit1051 analyzer warnings that would fail the build. New v2 projects
opt in via their own csproj; the global flag can return once the legacy
projects are deleted in Task 56.
UnsTabDragDropE2ETests were timing out at the 'UNS Structure' nav-link
locator because AdminWebAppFactory never registered AdminHubConnectionFactory
/ HubTokenService / DataProtection — ClusterDetail.razor's @inject threw at
circuit boot, so the page never advanced past the Loading placeholder. 2 → 3
pass after the registrations land. Also documents the Modbus standard-vs-
exception_injection coverage matrix in the fixture README + cross-references
docs/drivers/AbServer-Test-Fixture.md from each Emulate test so a developer
landing on a skipped test has a direct doc pointer.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two real bugs uncovered by re-running with the new fixture defaults
pointing at the shared docker host. Both are test-side, not driver-side.
AbCip — Driver_reads_seeded_DInt_from_ab_server (4 parametrized rows):
Hardcoded 'ab://127.0.0.1:{port}/1,0' in the deviceUri instead of
the resolved fixture.Host. The new 10.100.0.35 default (and any
AB_SERVER_ENDPOINT override) silently couldn't reach this test —
the driver tried to connect to a non-existent localhost:44818 and
returned BadCommunicationError on all 4 profile rows. The sibling
Emulate tests already use the fixture's resolved endpoint; this
smoke test was missed in the original migration.
Fix: deviceUri = $"ab://{fixture.Host}:{fixture.Port}/1,0".
Modbus — Float32_With_CDAB_Roundtrips_Through_Wire:
Test wrote a Float32 to HR 100 (2 consecutive registers: 100+101).
standard.json's writable HR range declares [100,100] only — a
single-cell auto-incrementing register, not a 2-register pair. The
write to register 101 was rejected with Illegal Data Address
(BadOutOfRange).
Fix: moved the tag from HR 100 to HR 200 (in standard.json's
'[200, 209]' scratch range — 10 consecutive writable HRs). The
Float32+CDAB semantic the test exercises is unchanged.
Modbus — Block_Read_Coalescing_Reduces_PDU_Count_End_To_End:
Test read HR 300, 302, 304 — outside both the writable ranges and
the uint16 seed list. pymodbus rejects reads to unseeded HRs even
though 'hr size' is 2048. BadOutOfRange on every read.
Fix: moved the tags from 300/302/304 to 200/202/204 (within the
scratch range). The non-contiguous coalescing semantic (3 tags
inside a 5-register window with MaxReadGap=5) is preserved.
After this commit:
- Modbus.IntegrationTests: 6/38 pass / 32 skip / 0 fail
(was 4 pass / 32 skip / 2 fail; 32 skips are profile-gated
ExceptionInjectionTests — they need MODBUS_SIM_PROFILE=
exception_injection and a different container, intentional gating)
- AbCip.IntegrationTests: 10/12 pass / 2 skip / 0 fail
(was 6 pass / 2 skip / 4 fail; 2 skips are Emulate tests that
need the fixture for separate scenarios)
No driver code changed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Joseph Doherty