Files
Joseph Doherty
780358c790
AbCip whole-UDT read optimization (#194) — declaration-driven member grouping collapses N per-member reads into one parent-UDT read + client-side decode. Closes task #194. On a batch that includes multiple members of the same hand-declared UDT tag, ReadAsync now issues one libplctag read on the parent + decodes each member from the runtime's buffer at its computed byte offset. A 6-member Motor UDT read goes from 6 libplctag round-trips to 1 — the Rockwell-suggested pattern for minimizing CIP request overhead on batch reads of UDT state (decision #11's follow-through on what the template decoder from task #179 was meant to enable). AbCipUdtMemberLayout is a pure-function helper that computes declared-member byte offsets under Logix natural-alignment rules (SInt 1-byte / Int 2-byte / DInt + Real + Dt 4-byte / LInt + ULInt + LReal 8-byte; alignment pad inserted before each member as needed). Opts out for BOOL / String / Structure members — BOOL storage in Logix UDTs packs into a hidden host byte whose position can't be computed from declaration-only info, and String members need length-prefix + STRING[82] fan-out which libplctag already handles via a per-tag DecodeValue path. The CIP Template Object shape from task #179 (when populated via FetchUdtShapeAsync) carries real offsets for those members — layering that richer path on top of the planner is a separate follow-up and does not change this PR's conservative behaviour. AbCipUdtReadPlanner is the scheduling function ReadAsync consults each batch — pure over (requests, tagsByName), emits Groups + Fallbacks. A group is formed when (a) the reference resolves to "parent.member"; (b) parent is a Structure tag with declared Members; (c) the layout helper succeeds on those members; (d) the specific member appears in the computed offset map; (e) at least two members of the same parent appear in the batch — single-member groups demote to the fallback path because one whole-UDT read vs one per-member read is equivalent cost but more client-side work. Original batch indices are preserved through the plan so out-of-order batches write decoded values back at the right output slot; the caller's result array order is invariant. IAbCipTagRuntime.DecodeValueAt(AbCipDataType, int offset, int? bitIndex) is the new hot-path method — LibplctagTagRuntime delegates to libplctag's offset-aware Get*(offset) calls (GetInt32, GetFloat32, etc.) that were always there; previously every call passed offset 0. DecodeValue(type, bitIndex) stays as the shorthand + forwards to DecodeValueAt with offset 0, preserving the existing single-tag read path + every test that exercises it. FakeAbCipTag gains a ValuesByOffset dictionary so tests can drive multi-member decoding by setting offset→value before the read fires; unmapped offsets fall back to the existing Value field so the 200+ existing tests that never set ValuesByOffset keep working unchanged. AbCipDriver.ReadAsync refactored: planner splits the batch, ReadGroupAsync handles each UDT group (one EnsureTagRuntimeAsync on the parent + one ReadAsync + N DecodeValueAt calls), ReadSingleAsync handles each fallback (the pre-#194 per-tag path, now extracted + threaded through). A per-group failure stamps the mapped libplctag status across every grouped member only — sibling groups + fallback refs are unaffected. Health-surface updates happen once per successful group rather than once per member to avoid ping-ponging the DriverState bookkeeping. Five AbCipUdtMemberLayoutTests: packed atomics get natural-alignment offsets including 8-byte pad before LInt; SInts pack without padding; BOOL/String/Structure opt out + return null; empty member list returns null. Six AbCipUdtReadPlannerTests: two members group; single-member demotes to fallback; unknown references fall back without poisoning groups; atomic top-level tags fall back untouched; UDTs containing BOOL don't group; original indices survive out-of-order batches. Five AbCipDriverWholeUdtReadTests (real driver + fake runtime): two grouped members trigger exactly one parent read + one fake runtime (proving the optimization engages); each member decodes at its own offset via ValuesByOffset; parent-read non-zero status stamps Bad across the group; mixed UDT-member + atomic top-level batch produces 2 runtimes + 2 reads (not 3); single-member-of-UDT still uses the member-level runtime (proving demotion works). Driver builds 0 errors; AbCip.Tests 227/227 (was 211, +16 new).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
AclsTab Probe-this-permission — first of three #196 slices. New /clusters/{ClusterId}/draft/{GenerationId} ACLs-tab gains a probe card above the grant table so operators can ask the trie "if cn=X asks for permission Y on node Z, would it be granted, and which rows contributed?" without shell-ing into the DB. Service thinly wraps the same PermissionTrieBuilder + PermissionTrie.CollectMatches call path the Server's dispatch layer uses at request time, so a probe answer is by construction identical to what the live server would decide. New PermissionProbeService.ProbeAsync(generationId, ldapGroup, NodeScope, requiredFlags) — loads the target generation's NodeAcl rows filtered to the cluster (critical: without the cluster filter, cross-cluster grants leak into the probe which tested false-positive in the unit suite), builds a trie, CollectMatches against the supplied scope + [ldapGroup], ORs the matched-grant flags into Effective, compares to Required. Returns PermissionProbeResult(Granted, Required, Effective, Matches) — Matches carries LdapGroup + Scope + PermissionFlags per matched row so the UI can render the contribution chain. Zero side effects + no audit rows — a failing probe is a question, not a denial. AclsTab.razor gains the probe card at the top (before the New-grant form + grant table): six inputs for ldap group + every NodeScope level (NamespaceId → UnsAreaId → UnsLineId → EquipmentId → TagId — blank fields become null so the trie walks only as deep as the operator specified), a NodePermissions dropdown filtered to skip None, Probe button, green Granted / red Denied badge + Required/Effective bitmask display, and (when matches exist) a small table showing which LdapGroup matched at which level with which flags. Admin csproj adds ProjectReference to Core — the trie + NodeScope live there + were previously Server-only. Five new PermissionProbeServiceTests covering: cluster-level row grants a namespace-level read; no-group-match denies with empty Effective; matching group but insufficient flags (Browse+Read vs WriteOperate required) denies with correct Effective bitmask; cross-cluster grants stay isolated (c2's WriteOperate does NOT leak into c1's probe); generation isolation (gen1's Read-only does NOT let gen2's WriteOperate-requiring probe pass). Admin.Tests 92/92 passing (was 87, +5). Admin builds 0 errors. Remaining #196 slices — SignalR invalidation + draft-diff ACL section — ship in follow-up PRs so the review surface per PR stays tight.
Roslyn analyzer — detect unwrapped driver-capability calls (OTOPCUA0001). Closes task #200. New netstandard2.0 analyzer project src/ZB.MOM.WW.OtOpcUa.Analyzers registered as an <Analyzer>-item ProjectReference from the Server csproj so the warning fires at every Server compile. First (and only so far) rule OTOPCUA0001 — "Driver capability call must be wrapped in CapabilityInvoker" — walks every InvocationOperation in the AST + trips when (a) the target method implements one of the seven guarded capability interfaces (IReadable / IWritable / ITagDiscovery / ISubscribable / IHostConnectivityProbe / IAlarmSource / IHistoryProvider) AND (b) the method's return type is Task, Task<T>, ValueTask, or ValueTask<T> — the async-wire-call constraint narrows the rule to the surfaces the Phase 6.1 pipeline actually wraps + sidesteps pure in-memory accessors like IHostConnectivityProbe.GetHostStatuses() which would trigger false positives AND (c) the call does NOT sit inside a lambda argument passed to CapabilityInvoker.ExecuteAsync / ExecuteWriteAsync / AlarmSurfaceInvoker.*. The wrapper detection walks up the syntax tree from the call site, finds any enclosing InvocationExpressionSyntax whose method's containing type is one of the wrapper classes, + verifies the call lives transitively inside that invocation's AnonymousFunctionExpressionSyntax argument — a sibling "result = await driver.ReadAsync(...)" followed by a separate invoker.ExecuteAsync(...) call does NOT satisfy the wrapping rule + the analyzer flags it (regression guard in the 5th test). Five xunit-v3 + Shouldly tests at tests/ZB.MOM.WW.OtOpcUa.Analyzers.Tests: direct ReadAsync in server namespace trips; wrapped ReadAsync inside CapabilityInvoker.ExecuteAsync lambda passes; direct WriteAsync trips; direct DiscoverAsync trips; sneaky pattern — read outside the lambda + ExecuteAsync with unrelated lambda nearby — still trips. Hand-rolled test harness compiles a stub-plus-user snippet via CSharpCompilation.WithAnalyzers + runs GetAnalyzerDiagnosticsAsync directly, deliberately avoiding Microsoft.CodeAnalysis.CSharp.Analyzer.Testing.XUnit because that package pins to xunit v2 + this repo is on xunit.v3 everywhere else. RS2008 release-tracking noise suppressed by adding AnalyzerReleases.Shipped.md + AnalyzerReleases.Unshipped.md as AdditionalFiles, which is the canonical Roslyn-analyzer hygiene path. Analyzer DLL referenced from Server.csproj via ProjectReference with OutputItemType=Analyzer + ReferenceOutputAssembly=false — the DLL ships as a compiler plugin, not a runtime dependency. Server build validates clean: the analyzer activates on every Server file but finds zero violations, which confirms the Phase 6.1 wrapping work done in prior PRs is complete + the analyzer is now the regression guard preventing the next new capability surface from being added raw. slnx updated with both the src + tests project entries. Full solution build clean, analyzer suite 5/5 passing.
Client rename residuals: lmxopcua-cli → otopcua-cli + LmxOpcUaClient → OtOpcUaClient with migration shim. Closes task #208 (the executable-name + LocalAppData-folder slice that was called out in Client.CLI.md / Client.UI.md as a deliberately-deferred residual of the Phase 0 rename). Six source references flipped to the canonical OtOpcUaClient spelling: Program.cs CliFx executable name + description (lmxopcua-cli → otopcua-cli), DefaultApplicationConfigurationFactory.cs ApplicationName + ApplicationUri (LmxOpcUaClient + urn:localhost:LmxOpcUaClient → OtOpcUaClient + urn:localhost:OtOpcUaClient), OpcUaClientService.CreateSessionAsync session-name arg, ConnectionSettings.CertificateStorePath default, MainWindowViewModel.CertificateStorePath default, JsonSettingsService.SettingsDir. Two consuming tests (ConnectionSettingsTests + MainWindowViewModelTests) updated to assert the new canonical name. New ClientStoragePaths static helper at src/ZB.MOM.WW.OtOpcUa.Client.Shared/ClientStoragePaths.cs is the migration shim — single entry point for the PKI root + pki subpath, runs a one-shot legacy-folder probe on first resolution: if {LocalAppData}/LmxOpcUaClient/ exists + {LocalAppData}/OtOpcUaClient/ does not, Directory.Move renames it in place (atomic on NTFS within the same volume) so trusted server certs + saved connection settings persist across the rename without operator action. Idempotent per-process via a Lock-guarded _migrationChecked flag so repeated CertificateStorePath getter calls on the hot path pay no IO cost beyond the first. Fresh-install path (neither folder exists) + already-migrated path (only canonical exists) + manual-override path (both exist — developer has set up something explicit) are all no-ops that leave state alone. IOException on the Directory.Move is swallowed + logged as a false return so a concurrent peer process losing the race doesn't crash the consumer; the losing process falls back to whatever state exists. Five new ClientStoragePathsTests assert: GetRoot ends with canonical name under LocalAppData, GetPkiPath nests pki under root, CanonicalFolderName is OtOpcUaClient, LegacyFolderName is LmxOpcUaClient (the migration contract — a typo here would leak the legacy folder past the shim), repeat invocation returns false after first-touch arms the in-process guard. Doc-side residual-explanation notes in docs/Client.CLI.md + docs/Client.UI.md are dropped now that the rename is real; replaced with a short "pre-#208 dev boxes migrate automatically on first launch" note that points at ClientStoragePaths. Sample CLI invocations in Client.CLI.md updated via sed from lmxopcua-cli to otopcua-cli across every command block (14 replacements). Pre-existing staleness in SubscribeCommandTests.Execute_PrintsSubscriptionMessage surfaced during the test run — the CLI's subscribe command has long since switched to an aggregate "Subscribed to {count}/{total} nodes (interval: ...)" output format but the test still asserted the original single-node form. Updated the assertion to match current output + added a comment explaining the change; this is unrelated to the rename but was blocking a green Client.CLI.Tests run. Full solution build 0 errors; Client.Shared.Tests 136/136 + 5 new shim tests passing; Client.UI.Tests 98/98; Client.CLI.Tests 52/52 (was 51/52 before the subscribe-test fix). No Admin/Core/Server changes — this touches only the client layer.
Client rename residuals: lmxopcua-cli → otopcua-cli + LmxOpcUaClient → OtOpcUaClient with migration shim. Closes task #208 (the executable-name + LocalAppData-folder slice that was called out in Client.CLI.md / Client.UI.md as a deliberately-deferred residual of the Phase 0 rename). Six source references flipped to the canonical OtOpcUaClient spelling: Program.cs CliFx executable name + description (lmxopcua-cli → otopcua-cli), DefaultApplicationConfigurationFactory.cs ApplicationName + ApplicationUri (LmxOpcUaClient + urn:localhost:LmxOpcUaClient → OtOpcUaClient + urn:localhost:OtOpcUaClient), OpcUaClientService.CreateSessionAsync session-name arg, ConnectionSettings.CertificateStorePath default, MainWindowViewModel.CertificateStorePath default, JsonSettingsService.SettingsDir. Two consuming tests (ConnectionSettingsTests + MainWindowViewModelTests) updated to assert the new canonical name. New ClientStoragePaths static helper at src/ZB.MOM.WW.OtOpcUa.Client.Shared/ClientStoragePaths.cs is the migration shim — single entry point for the PKI root + pki subpath, runs a one-shot legacy-folder probe on first resolution: if {LocalAppData}/LmxOpcUaClient/ exists + {LocalAppData}/OtOpcUaClient/ does not, Directory.Move renames it in place (atomic on NTFS within the same volume) so trusted server certs + saved connection settings persist across the rename without operator action. Idempotent per-process via a Lock-guarded _migrationChecked flag so repeated CertificateStorePath getter calls on the hot path pay no IO cost beyond the first. Fresh-install path (neither folder exists) + already-migrated path (only canonical exists) + manual-override path (both exist — developer has set up something explicit) are all no-ops that leave state alone. IOException on the Directory.Move is swallowed + logged as a false return so a concurrent peer process losing the race doesn't crash the consumer; the losing process falls back to whatever state exists. Five new ClientStoragePathsTests assert: GetRoot ends with canonical name under LocalAppData, GetPkiPath nests pki under root, CanonicalFolderName is OtOpcUaClient, LegacyFolderName is LmxOpcUaClient (the migration contract — a typo here would leak the legacy folder past the shim), repeat invocation returns false after first-touch arms the in-process guard. Doc-side residual-explanation notes in docs/Client.CLI.md + docs/Client.UI.md are dropped now that the rename is real; replaced with a short "pre-#208 dev boxes migrate automatically on first launch" note that points at ClientStoragePaths. Sample CLI invocations in Client.CLI.md updated via sed from lmxopcua-cli to otopcua-cli across every command block (14 replacements). Pre-existing staleness in SubscribeCommandTests.Execute_PrintsSubscriptionMessage surfaced during the test run — the CLI's subscribe command has long since switched to an aggregate "Subscribed to {count}/{total} nodes (interval: ...)" output format but the test still asserted the original single-node form. Updated the assertion to match current output + added a comment explaining the change; this is unrelated to the rename but was blocking a green Client.CLI.Tests run. Full solution build 0 errors; Client.Shared.Tests 136/136 + 5 new shim tests passing; Client.UI.Tests 98/98; Client.CLI.Tests 52/52 (was 51/52 before the subscribe-test fix). No Admin/Core/Server changes — this touches only the client layer.
Client rename residuals: lmxopcua-cli → otopcua-cli + LmxOpcUaClient → OtOpcUaClient with migration shim. Closes task #208 (the executable-name + LocalAppData-folder slice that was called out in Client.CLI.md / Client.UI.md as a deliberately-deferred residual of the Phase 0 rename). Six source references flipped to the canonical OtOpcUaClient spelling: Program.cs CliFx executable name + description (lmxopcua-cli → otopcua-cli), DefaultApplicationConfigurationFactory.cs ApplicationName + ApplicationUri (LmxOpcUaClient + urn:localhost:LmxOpcUaClient → OtOpcUaClient + urn:localhost:OtOpcUaClient), OpcUaClientService.CreateSessionAsync session-name arg, ConnectionSettings.CertificateStorePath default, MainWindowViewModel.CertificateStorePath default, JsonSettingsService.SettingsDir. Two consuming tests (ConnectionSettingsTests + MainWindowViewModelTests) updated to assert the new canonical name. New ClientStoragePaths static helper at src/ZB.MOM.WW.OtOpcUa.Client.Shared/ClientStoragePaths.cs is the migration shim — single entry point for the PKI root + pki subpath, runs a one-shot legacy-folder probe on first resolution: if {LocalAppData}/LmxOpcUaClient/ exists + {LocalAppData}/OtOpcUaClient/ does not, Directory.Move renames it in place (atomic on NTFS within the same volume) so trusted server certs + saved connection settings persist across the rename without operator action. Idempotent per-process via a Lock-guarded _migrationChecked flag so repeated CertificateStorePath getter calls on the hot path pay no IO cost beyond the first. Fresh-install path (neither folder exists) + already-migrated path (only canonical exists) + manual-override path (both exist — developer has set up something explicit) are all no-ops that leave state alone. IOException on the Directory.Move is swallowed + logged as a false return so a concurrent peer process losing the race doesn't crash the consumer; the losing process falls back to whatever state exists. Five new ClientStoragePathsTests assert: GetRoot ends with canonical name under LocalAppData, GetPkiPath nests pki under root, CanonicalFolderName is OtOpcUaClient, LegacyFolderName is LmxOpcUaClient (the migration contract — a typo here would leak the legacy folder past the shim), repeat invocation returns false after first-touch arms the in-process guard. Doc-side residual-explanation notes in docs/Client.CLI.md + docs/Client.UI.md are dropped now that the rename is real; replaced with a short "pre-#208 dev boxes migrate automatically on first launch" note that points at ClientStoragePaths. Sample CLI invocations in Client.CLI.md updated via sed from lmxopcua-cli to otopcua-cli across every command block (14 replacements). Pre-existing staleness in SubscribeCommandTests.Execute_PrintsSubscriptionMessage surfaced during the test run — the CLI's subscribe command has long since switched to an aggregate "Subscribed to {count}/{total} nodes (interval: ...)" output format but the test still asserted the original single-node form. Updated the assertion to match current output + added a comment explaining the change; this is unrelated to the rename but was blocking a green Client.CLI.Tests run. Full solution build 0 errors; Client.Shared.Tests 136/136 + 5 new shim tests passing; Client.UI.Tests 98/98; Client.CLI.Tests 52/52 (was 51/52 before the subscribe-test fix). No Admin/Core/Server changes — this touches only the client layer.
EquipmentNodeWalker — pure-function UNS tree materialization (ADR-001 Task A, task #210). The walker traverses the Config-DB snapshot for a single Equipment-kind namespace (Areas / Lines / Equipment / Tags) and streams IAddressSpaceBuilder.Folder + Variable + AddProperty calls to materialize the canonical 5-level Unified Namespace browse tree that decisions #116-#121 promise external consumers. Pure function: no OPC UA SDK dependency, no DB access, no state — consumes pre-loaded EF Core row collections + streams into the supplied builder. Server-side wiring (load snapshot → call walker → per-tag capability probe) is Task B's scope, alongside NodeScopeResolver's Config-DB join + the ACL integration test that closes task #195. This PR is the Core.OpcUa primitive the server will consume. Walk algorithm — content is grouped up-front (lines by area, equipment by line, tags by equipment) into OrdinalIgnoreCase dictionaries so the per-level nested foreach stays O(N+M) rather than O(N·M) at each UNS level; orderings are deterministic on Name with StringComparer.Ordinal so diffs across runs (e.g. integration-test assertions) are stable. Areas → Lines → Equipment emitted as Folder nodes with browse-name = Name per decision #120. Under each Equipment folder: five identifier properties per decision #121 (EquipmentId + EquipmentUuid always; MachineCode always — it's a required column on the entity; ZTag + SAPID skipped when null to avoid empty-string property noise); IdentificationFolderBuilder.Build materializes the OPC 40010 sub-folder when HasAnyFields(equipment) returns true, skipped otherwise to avoid a pointless empty folder; then one Variable node per Tag row bound to this Equipment (Tag.EquipmentId non-null matches Equipment.EquipmentId) emitted in Name order. Tags with null EquipmentId are walker-skipped — those are SystemPlatform-kind (Galaxy) tags that take the driver-native DiscoverAsync path per decision #120. DriverAttributeInfo construction: FullName = Tag.TagConfig (driver-specific wire-level address); DriverDataType parsed from Tag.DataType which stores the enum name string per decision #138; unparseable values fall back to DriverDataType.String so a one-off driver-specific type doesn't abort the whole walk (driver still sees the original address at runtime + can surface its own typed value via the variant). Address validation is deliberately NOT done at build time per ADR-001 Option A: unreachable addresses surface as OPC UA Bad status via the natural driver-read failure path at runtime, legible to operators through their Admin UI + OPC UA client inspection. Eight new EquipmentNodeWalkerTests: empty content emits nothing; Area/Line/Equipment folder emission order matches Name-sorted deterministic traversal; five identifier properties appear on Equipment nodes with correct values, ZTag + SAPID skipped when null + emitted when non-null; Identification sub-folder materialized when at least one OPC 40010 field is non-null + omitted when all are null; tags with matching EquipmentId emit as Variable nodes under the Equipment folder in Name order, tags with null EquipmentId walker-skipped; unparseable DataType falls back to String. RecordingBuilder test double captures Folder/Variable/Property calls into a tree structure tests can navigate. Core project builds 0 errors; Core.Tests 190/190 (was 182, +8 new walker tests). No Server/Admin changes — Task B lands the server-side wiring + consumes this walker from DriverNodeManager.
ab_server integration fixture — per-family profiles + documented CI-fetch contract. Closes task #180 (AB CIP follow-up — ab_server CI fixture). Replaces the prior hardcoded single-family fixture with a parametric AbServerProfile abstraction covering ControlLogix / CompactLogix / Micro800 / GuardLogix. Prebuilt-Windows-binary fetch is documented as a CI YAML step rather than fabricated C#-side, because SHA-pinned binary distribution is a CI workflow concern (libplctag owns releases, we pin a version + verify hash) not a test-framework concern. New AbServerProfile record + KnownProfiles static class at tests/.../AbServerProfile.cs. Four profiles: ControlLogix (widest coverage — DINT/REAL/BOOL/SINT/STRING atomic + DINT[16] array so the driver's @tags Symbol-Object decoder + array-bound path both get end-to-end coverage), CompactLogix (atomic subset — driver-side ConnectionSize quirk from PR 10 still applies since ab_server doesn't enforce the narrower limit), Micro800 (ab_server has no dedicated --plc micro800 mode — falls back to controllogix while driver-side path enforces empty routing + unconnected-only per PR 11; real Micro800 coverage requires a 2080 lab rig), GuardLogix (ab_server has no safety subsystem — profile emulates the _S-suffixed naming contract the driver's safety-ViewOnly classification reads in PR 12; real safety-lock behavior requires a 1756-L8xS physical rig). Each profile composes --plc + --tag args via BuildCliArgs(port) — pure string formatter so the composition logic is unit-testable without launching the simulator. AbServerFixture gains a ctor overload taking AbServerProfile + port (defaults back to ControlLogix on parameterless ctor so existing test suites keep compiling). Fixture's InitializeAsync hands the profile's CLI args to ProcessStartInfo.Arguments. New AbServerTheoryAttribute mirrors AbServerFactAttribute but extends TheoryAttribute so a single test can MemberData over KnownProfiles.All + cover all four families. AbCipReadSmokeTests converted from single-fact to theory parametrized over KnownProfiles.All — one row per family reads TestDINT + asserts Good status + Healthy driver state. Fixture lifecycle is explicit try/finally rather than await using because IAsyncLifetime.DisposeAsync returns ValueTask + xUnit's concrete IAsyncDisposable shim depends on xunit version; explicit beats implicit here. Eight new unit tests in AbServerProfileTests.cs (runs without the simulator so CI green even when the binary is absent): BuildCliArgs composes port + plc + tag flags in the documented order; empty seed-tag list still emits port + plc; SeedTag.ToCliSpec handles both 2-segment scalar + 3-segment array; KnownProfiles.ForFamily returns expected --plc arg for every family (verifies Micro800 + GuardLogix both fall back to controllogix); KnownProfiles.All covers every AbCipPlcFamily enum value (regression guard — adding a new family without a profile fails this test); ControlLogix seeds every atomic type the driver supports; GuardLogix seeds at least one _S-suffixed safety tag. Integration tests still skip cleanly when ab_server isn't on PATH. 11/11 unit tests passing in this project (8 new + 3 prior). Full Admin solution builds 0 errors. docs/v2/test-data-sources.md gets a new "CI fixture" subsection under §2.Gotchas with the exact GitHub Actions YAML step — fetch the pinned libplctag release, SHA256-verify against a pinned hash recorded in the repo's CI lockfile (drift = fail closed), extract, append to PATH. The C# harness stays PATH-driven so dev-box installs (cmake + make from source) work identically to CI.
AbCip whole-UDT read optimization (#194) — declaration-driven member grouping collapses N per-member reads into one parent-UDT read + client-side decode. Closes task #194. On a batch that includes multiple members of the same hand-declared UDT tag, ReadAsync now issues one libplctag read on the parent + decodes each member from the runtime's buffer at its computed byte offset. A 6-member Motor UDT read goes from 6 libplctag round-trips to 1 — the Rockwell-suggested pattern for minimizing CIP request overhead on batch reads of UDT state (decision #11's follow-through on what the template decoder from task #179 was meant to enable). AbCipUdtMemberLayout is a pure-function helper that computes declared-member byte offsets under Logix natural-alignment rules (SInt 1-byte / Int 2-byte / DInt + Real + Dt 4-byte / LInt + ULInt + LReal 8-byte; alignment pad inserted before each member as needed). Opts out for BOOL / String / Structure members — BOOL storage in Logix UDTs packs into a hidden host byte whose position can't be computed from declaration-only info, and String members need length-prefix + STRING[82] fan-out which libplctag already handles via a per-tag DecodeValue path. The CIP Template Object shape from task #179 (when populated via FetchUdtShapeAsync) carries real offsets for those members — layering that richer path on top of the planner is a separate follow-up and does not change this PR's conservative behaviour. AbCipUdtReadPlanner is the scheduling function ReadAsync consults each batch — pure over (requests, tagsByName), emits Groups + Fallbacks. A group is formed when (a) the reference resolves to "parent.member"; (b) parent is a Structure tag with declared Members; (c) the layout helper succeeds on those members; (d) the specific member appears in the computed offset map; (e) at least two members of the same parent appear in the batch — single-member groups demote to the fallback path because one whole-UDT read vs one per-member read is equivalent cost but more client-side work. Original batch indices are preserved through the plan so out-of-order batches write decoded values back at the right output slot; the caller's result array order is invariant. IAbCipTagRuntime.DecodeValueAt(AbCipDataType, int offset, int? bitIndex) is the new hot-path method — LibplctagTagRuntime delegates to libplctag's offset-aware Get*(offset) calls (GetInt32, GetFloat32, etc.) that were always there; previously every call passed offset 0. DecodeValue(type, bitIndex) stays as the shorthand + forwards to DecodeValueAt with offset 0, preserving the existing single-tag read path + every test that exercises it. FakeAbCipTag gains a ValuesByOffset dictionary so tests can drive multi-member decoding by setting offset→value before the read fires; unmapped offsets fall back to the existing Value field so the 200+ existing tests that never set ValuesByOffset keep working unchanged. AbCipDriver.ReadAsync refactored: planner splits the batch, ReadGroupAsync handles each UDT group (one EnsureTagRuntimeAsync on the parent + one ReadAsync + N DecodeValueAt calls), ReadSingleAsync handles each fallback (the pre-#194 per-tag path, now extracted + threaded through). A per-group failure stamps the mapped libplctag status across every grouped member only — sibling groups + fallback refs are unaffected. Health-surface updates happen once per successful group rather than once per member to avoid ping-ponging the DriverState bookkeeping. Five AbCipUdtMemberLayoutTests: packed atomics get natural-alignment offsets including 8-byte pad before LInt; SInts pack without padding; BOOL/String/Structure opt out + return null; empty member list returns null. Six AbCipUdtReadPlannerTests: two members group; single-member demotes to fallback; unknown references fall back without poisoning groups; atomic top-level tags fall back untouched; UDTs containing BOOL don't group; original indices survive out-of-order batches. Five AbCipDriverWholeUdtReadTests (real driver + fake runtime): two grouped members trigger exactly one parent read + one fake runtime (proving the optimization engages); each member decodes at its own offset via ValuesByOffset; parent-read non-zero status stamps Bad across the group; mixed UDT-member + atomic top-level batch produces 2 runtimes + 2 reads (not 3); single-member-of-UDT still uses the member-level runtime (proving demotion works). Driver builds 0 errors; AbCip.Tests 227/227 (was 211, +16 new).
RMW pass 2 — AbCip BOOL-within-DINT + AbLegacy bit-within-word. Closes task #181. AbCip — AbCipDriver.WriteAsync now detects BOOL writes with a bit index + routes them through WriteBitInDIntAsync: strip the .N suffix to form the parent DINT tag path (via AbCipTagPath with BitIndex=null + ToLibplctagName), get/create a cached parent IAbCipTagRuntime via EnsureParentRuntimeAsync (distinct from the bit-selector tag runtime so read + write target the DINT directly), acquire a per-parent-name SemaphoreSlim, Read → Convert.ToInt32 the current DINT → (current | 1<<bit) or (current & ~(1<<bit)) → Write via EncodeValue(DInt, updated). Per-parent lock prevents concurrent writers to the same DINT from losing updates — parallels Modbus + FOCAS pass 1. DeviceState gains ParentRuntimes dict + GetRmwLock helper + _rmwLocks ConcurrentDictionary. DisposeHandles now walks ParentRuntimes too. LibplctagTagRuntime.EncodeValue's BOOL-with-bitIndex branch stays as a defensive throw (message updated to point at the new driver-level dispatch) so an accidental bypass fails loudly rather than silently clobbering the whole DINT. AbLegacy — identical pattern for PCCC N-file bit writes. AbLegacyDriver.WriteAsync detects Bit with bitIndex + PMC letter not in {B, I, O} (B-file + I/O use their own bit-addressable semantics so don't RMW at N-file word level), routes through WriteBitInWordAsync which uses Int16 for the parent word, creates + caches a parent runtime with the suffix-stripped N7:0 address, acquires per-parent lock, RMW. DeviceState extended the same way as AbCip (ParentRuntimes + GetRmwLock). LibplctagLegacyTagRuntime.EncodeValue Bit-with-bitIndex branch points at the driver dispatch. Tests — 5 new AbCipBoolInDIntRmwTests (bit set ORs + preserves, bit clear ANDs + preserves, 8-way concurrent writes to same parent compose to 0xFF, different-parent writes get separate runtimes, repeat bit writes reuse the parent runtime init-count 1 + write-count 2), 4 new AbLegacyBitRmwTests (bit set preserves, bit clear preserves 0xFFF7, 8-way concurrent 0xFF, repeat writes reuse parent). Two pre-existing tests flipped — AbCipDriverWriteTests.Bit_in_dint_write_returns_BadNotSupported + AbLegacyReadWriteTests.Bit_within_word_write_rejected_as_BadNotSupported both now assert Good instead of BadNotSupported, renamed to _now_succeeds_via_RMW. Total tests — AbCip 166/166, AbLegacy 96/96, full solution builds 0 errors; Modbus + FOCAS + TwinCAT + other drivers untouched. Task #181 done across all four libplctag-backed + non-libplctag drivers (Modbus BitInRegister + AbCip BOOL-in-DINT + AbLegacy N-file bit + FOCAS PMC Bit — all with per-parent-word serialisation).
RMW pass 1 — Modbus BitInRegister + FOCAS PMC Bit write paths. First half of task #181 — the two drivers where read-modify-write is a clean protocol-level insertion (Modbus FC03/FC06 round-trip + FOCAS pmc_rdpmcrng / pmc_wrpmcrng round-trip). Per-driver SemaphoreSlim registry keyed on the parent word address serialises concurrent bit writes so two writers targeting different bits in the same word don't lose one another's update. Modbus — ModbusDriver gains WriteBitInRegisterAsync + _rmwLocks ConcurrentDictionary. WriteOneAsync routes BitInRegister (HoldingRegisters region only) through RMW ahead of the normal encode path. Read uses FC03 Read Holding Registers for 1 register at tag.Address, bit-op on the returned ushort via (current | 1<<bit) for set / (current & ~(1<<bit)) for clear, write back via FC06 Write Single Register. Per-address lock prevents concurrent bit writes to the same register from racing. Rejects out-of-range bits (0-15) with InvalidOperationException. EncodeRegister's BitInRegister branch repurposed as a defensive guard — if a non-RMW caller ever reaches it, throw so an unintended bypass stays loud rather than silently clobbering. FOCAS — FwlibFocasClient gains WritePmcBitAsync + _rmwLocks keyed on {addrType}:{byteAddr}. Driver-layer WriteAsync routes Bit writes with a bitIndex through the new path; other Pmc writes still hit the direct pmc_wrpmcrng path. RMW uses cnc_rdpmcrng + Byte dataType to grab the parent byte, bit-op with (current | 1<<bit) or (current & ~(1<<bit)), cnc_wrpmcrng to write back. Rejects out-of-range bits (0-7, FOCAS PMC bytes are 8-bit) with InvalidOperationException. EncodePmcValue's Bit branch now treats a no-bitIndex case as whole-byte boolean (non-zero / zero); bitIndex-present writes never hit this path because they dispatch to WritePmcBitAsync upstream. Tests — 5 new ModbusBitRmwTests + 4 new FocasPmcBitRmwTests + 1 renamed pre-existing test each covering — bit set preserves other bits, bit clear preserves other bits, concurrent bit writes to same word/byte compose correctly (8-parallel stress), bit writes on different parent words proceed without contention (4-parallel), sequential bit sets compose into 0xFF after all 8. Fake PmcRmwFake in FOCAS tests simulates the PMC byte storage + surfaces it through the IFocasClient contract so the test asserts driver-level behavior without needing Fwlib32.dll. FwlibNativeHelperTests.EncodePmcValue_Bit_throws_NotSupported_for_RMW_gap replaced with EncodePmcValue_Bit_without_bit_index_writes_byte_boolean reflecting the new behavior. ModbusDataTypeTests.BitInRegister_write_is_not_supported_in_PR24 renamed to BitInRegister_EncodeRegister_still_rejects_direct_calls; the message assertion updated to match the new defensive message. Modbus tests now 182/182, FOCAS tests now 119/119; full solution builds 0 errors; AbCip/AbLegacy/TwinCAT untouched (those get their RMW pass in a follow-up since libplctag bit access may need a parallel parent-word handle). Task #181 stays pending until that second pass lands.
Phase 3 PR 40 — LiveStackSmokeTests: write-roundtrip + subscribe-receives-OnDataChange against the live Galaxy. Finishes LMX #5 by exercising the IWritable + ISubscribable capability paths end-to-end through the Proxy → OtOpcUaGalaxyHost service → MXAccess → real Galaxy.
Phase 1 Streams B–E scaffold + Phase 2 Streams A–C scaffold — 8 new projects with ~70 new tests, all green alongside the 494 v1 IntegrationTests baseline (parity preserved: no v1 tests broken; legacy OtOpcUa.Host untouched). Phase 1 finish: Configuration project (16 entities + 10 enums + DbContext + DesignTimeDbContextFactory + InitialSchema/StoredProcedures/AuthorizationGrants migrations — 8 procs including sp_PublishGeneration with MERGE on ExternalIdReservation per decision #124, sp_RollbackToGeneration cloning rows into a new published generation, sp_ValidateDraft with cross-cluster-namespace + EquipmentUuid-immutability + ZTag/SAPID reservation pre-flight, sp_ComputeGenerationDiff with CHECKSUM-based row signature — plus OtOpcUaNode/OtOpcUaAdmin SQL roles with EXECUTE grants scoped to per-principal-class proc sets and DENY UPDATE/DELETE/INSERT/SELECT on dbo schema); managed DraftValidator covering UNS segment regex, path length, EquipmentUuid immutability across generations, same-cluster namespace binding (decision #122), reservation pre-flight, EquipmentId derivation (decision #125), driver↔namespace compatibility — returning every failing rule in one pass; LiteDB local cache with round-trip + ring pruning + corruption-fast-fail; GenerationApplier with per-entity Added/Removed/Modified diff and dependency-ordered callbacks (namespace → driver → device → equipment → poll-group → tag, Removed before Added); Core project with GenericDriverNodeManager (scaffold for the Phase 2 Galaxy port) and DriverHost lifecycle registry; Server project using Microsoft.Extensions.Hosting BackgroundService replacing TopShelf, with NodeBootstrap that falls back to LiteDB cache when the central DB is unreachable (decision #79); Admin project scaffolded as Blazor Server with Bootstrap 5 sidebar layout, cookie auth, three admin roles (ConfigViewer/ConfigEditor/FleetAdmin), Cluster + Generation services fronting the stored procs. Phase 2 scaffold: Driver.Galaxy.Shared (netstandard2.0) with full MessagePack IPC contract surface — Hello version negotiation, Open/CloseSession, Heartbeat, DiscoverHierarchy + GalaxyObjectInfo/GalaxyAttributeInfo, Read/WriteValues, Subscribe/Unsubscribe/OnDataChange, AlarmSubscribe/Event/Ack, HistoryRead, HostConnectivityStatus, Recycle — plus length-prefixed framing (decision #28) with a 16 MiB cap and thread-safe FrameWriter/FrameReader; Driver.Galaxy.Host (net48) implementing the Tier C cross-cutting protections from driver-stability.md — strict PipeAcl (allow configured server SID only, explicit deny on LocalSystem + Administrators), PipeServer with caller-SID verification via pipe.RunAsClient + WindowsIdentity.GetCurrent and per-process shared-secret Hello, Galaxy-specific MemoryWatchdog (warn at max(1.5×baseline, +200 MB), soft-recycle at max(2×baseline, +200 MB), hard ceiling 1.5 GB, slope ≥5 MB/min over 30-min rolling window), RecyclePolicy (1 soft recycle per hour cap + 03:00 local daily scheduled), PostMortemMmf (1000-entry ring buffer in %ProgramData%\OtOpcUa\driver-postmortem\galaxy.mmf, survives hard crash, readable cross-process), MxAccessHandle : SafeHandle (ReleaseHandle loops Marshal.ReleaseComObject until refcount=0 then calls optional unregister callback), StaPump with responsiveness probe (BlockingCollection dispatcher for Phase 1 — real Win32 GetMessage/DispatchMessage pump slots in with the same semantics when the Galaxy code lift happens), IsExternalInit shim for init setters on .NET 4.8; Driver.Galaxy.Proxy (net10) implementing IDriver + ITagDiscovery forwarding over the IPC channel with MX data-type and security-classification mapping, plus Supervisor pieces — Backoff (5s → 15s → 60s capped, reset-on-stable-run), CircuitBreaker (3 crashes per 5 min opens; 1h → 4h → manual cooldown escalation; sticky alert doesn't auto-clear), HeartbeatMonitor (2s cadence, 3 consecutive misses = host dead per driver-stability.md). Infrastructure: docker SQL Server remapped to host port 14330 to coexist with the native MSSQL14 Galaxy ZB DB instance on 1433; NuGetAuditSuppress applied per-project for two System.Security.Cryptography.Xml advisories that only reach via EF Core Design with PrivateAssets=all (fix ships in 11.0.0-preview); .slnx gains 14 project registrations. Deferred with explicit TODOs in docs/v2/implementation/phase-2-partial-exit-evidence.md: Phase 1 Stream E Admin UI pages (Generations listing + draft-diff-publish, Equipment CRUD with OPC 40010 fields, UNS Areas/Lines tabs, ACLs + permission simulator, Generic JSON config editor, SignalR real-time, Release-Reservation + Merge-Equipment workflows, LDAP login page, AppServer smoke test per decision #142), Phase 2 Stream D (Galaxy MXAccess code lift out of legacy OtOpcUa.Host, dual-service installer, appsettings → DriverConfig migration script, legacy Host deletion — blocked by parity), Phase 2 Stream E (v1 IntegrationTests against v2 topology, Client.CLI walkthrough diff, four 2026-04-13 stability findings regression tests, adversarial review — requires live MXAccess runtime).
RMW pass 1 — Modbus BitInRegister + FOCAS PMC Bit write paths. First half of task #181 — the two drivers where read-modify-write is a clean protocol-level insertion (Modbus FC03/FC06 round-trip + FOCAS pmc_rdpmcrng / pmc_wrpmcrng round-trip). Per-driver SemaphoreSlim registry keyed on the parent word address serialises concurrent bit writes so two writers targeting different bits in the same word don't lose one another's update. Modbus — ModbusDriver gains WriteBitInRegisterAsync + _rmwLocks ConcurrentDictionary. WriteOneAsync routes BitInRegister (HoldingRegisters region only) through RMW ahead of the normal encode path. Read uses FC03 Read Holding Registers for 1 register at tag.Address, bit-op on the returned ushort via (current | 1<<bit) for set / (current & ~(1<<bit)) for clear, write back via FC06 Write Single Register. Per-address lock prevents concurrent bit writes to the same register from racing. Rejects out-of-range bits (0-15) with InvalidOperationException. EncodeRegister's BitInRegister branch repurposed as a defensive guard — if a non-RMW caller ever reaches it, throw so an unintended bypass stays loud rather than silently clobbering. FOCAS — FwlibFocasClient gains WritePmcBitAsync + _rmwLocks keyed on {addrType}:{byteAddr}. Driver-layer WriteAsync routes Bit writes with a bitIndex through the new path; other Pmc writes still hit the direct pmc_wrpmcrng path. RMW uses cnc_rdpmcrng + Byte dataType to grab the parent byte, bit-op with (current | 1<<bit) or (current & ~(1<<bit)), cnc_wrpmcrng to write back. Rejects out-of-range bits (0-7, FOCAS PMC bytes are 8-bit) with InvalidOperationException. EncodePmcValue's Bit branch now treats a no-bitIndex case as whole-byte boolean (non-zero / zero); bitIndex-present writes never hit this path because they dispatch to WritePmcBitAsync upstream. Tests — 5 new ModbusBitRmwTests + 4 new FocasPmcBitRmwTests + 1 renamed pre-existing test each covering — bit set preserves other bits, bit clear preserves other bits, concurrent bit writes to same word/byte compose correctly (8-parallel stress), bit writes on different parent words proceed without contention (4-parallel), sequential bit sets compose into 0xFF after all 8. Fake PmcRmwFake in FOCAS tests simulates the PMC byte storage + surfaces it through the IFocasClient contract so the test asserts driver-level behavior without needing Fwlib32.dll. FwlibNativeHelperTests.EncodePmcValue_Bit_throws_NotSupported_for_RMW_gap replaced with EncodePmcValue_Bit_without_bit_index_writes_byte_boolean reflecting the new behavior. ModbusDataTypeTests.BitInRegister_write_is_not_supported_in_PR24 renamed to BitInRegister_EncodeRegister_still_rejects_direct_calls; the message assertion updated to match the new defensive message. Modbus tests now 182/182, FOCAS tests now 119/119; full solution builds 0 errors; AbCip/AbLegacy/TwinCAT untouched (those get their RMW pass in a follow-up since libplctag bit access may need a parallel parent-word handle). Task #181 stays pending until that second pass lands.
TwinCAT follow-up — Symbol browser via AdsClient + SymbolLoaderFactory. Closes task #188. Adds ITwinCATClient.BrowseSymbolsAsync — IAsyncEnumerable yielding TwinCATDiscoveredSymbol (InstancePath + mapped TwinCATDataType + ReadOnly flag) from the target's flat symbol table. AdsTwinCATClient implementation uses SymbolLoaderFactory.Create(_client, new SymbolLoaderSettings(SymbolsLoadMode.Flat)) + iterates loader.Symbols, maps IEC 61131-3 type names (BOOL/SINT/INT/DINT/LINT/REAL/LREAL/STRING/WSTRING/TIME/DATE/DT/TOD + BYTE/WORD/DWORD/LWORD unsigned-word aliases) through MapSymbolTypeName, checks SymbolAccessRights.Write bit for writable vs read-only. Unsupported types (UDTs / function blocks / arrays / pointers) surface with DataType=null so callers can skip or recurse. TwinCATDriverOptions.EnableControllerBrowse — new bool, default false to preserve the strict-config path. When true, DiscoverAsync iterates each device's BrowseSymbolsAsync, filters via TwinCATSystemSymbolFilter (rejects TwinCAT_*, Constants.*, Mc_*, __*, Global_Version* prefixes + anything empty), skips null-DataType symbols, emits surviving symbols under a per-device Discovered/ sub-folder with InstancePath as both FullName + BrowseName + ReadOnly→ViewOnly/writable→Operate. Pre-declared tags from TwinCATDriverOptions.Tags always emit regardless. Browse failure is non-fatal — exception caught + swallowed, pre-declared tags stay in the address space, operators see the failure in driver health on next read. TwinCATSystemSymbolFilter static class mirrors AbCipSystemTagFilter's shape with TwinCAT-specific prefixes. Fake client updated — BrowseResults list for test setup + FireNotification-style single-invocation on each subscribe, ThrowOnBrowse flag for failure testing. 8 new unit tests — strict path emits only pre-declared when EnableControllerBrowse=false, browse enabled adds Discovered/ folder, filter rejects system prefixes, null-DataType symbols skipped, ReadOnly symbols surface ViewOnly, browse failure leaves pre-declared intact, SystemSymbolFilter theory (10 cases). Total TwinCAT unit tests now 110/110 passing (+17 from the native-notification merge's 93); full solution builds 0 errors; other drivers untouched.
ADR-001 last-mile — Program.cs composes EquipmentNodeWalker into the production boot path. Closes task #214 + fully lands ADR-001 Option A as a live code path, not just a connected set of unit-tested primitives. After this PR a server booted against a real Config DB with Published Equipment rows materializes the UNS tree into the OPC UA address space on startup — the whole walker → wire-in → loader chain (PRs #153, #154, #155, #156) finally fires end-to-end in the production process. DriverEquipmentContentRegistry is the handoff between OpcUaServerService's bootstrap-time populate pass + OpcUaApplicationHost's StartAsync walker invocation. It's a singleton mutable holder with Get/Set/Count + Lock-guarded internal dictionary keyed OrdinalIgnoreCase to match the DriverInstanceId convention used by Equipment / Tag rows + walker grouping. Set-once-per-bootstrap semantics in practice though nothing enforces that at the type level — OpcUaServerService.PopulateEquipmentContentAsync is the only expected writer. Shared-mutable rather than immutable-passed-by-value because the DI graph builds OpcUaApplicationHost before NodeBootstrap has resolved the generation, so the registry must exist at compose time + fill at boot time. Program.cs now registers OpcUaApplicationHost via a factory lambda that threads registry.Get as the equipmentContentLookup delegate PR #155 added to the ctor seam — the one-line composition the earlier PR promised. EquipmentNamespaceContentLoader (from PR #156) is AddScoped since it takes the scoped OtOpcUaConfigDbContext; the populate pass in OpcUaServerService opens one IServiceScopeFactory scope + reuses the same loader + DbContext across every driver query rather than scoping-per-driver. OpcUaServerService.ExecuteAsync gets a new PopulateEquipmentContentAsync step between bootstrap + StartAsync: iterates DriverHost.RegisteredDriverIds, calls loader.LoadAsync per driver at the bootstrapped generationId, stashes non-null results in the registry. Null results are skipped — the wire-in's null-check treats absent registry entries as "this driver isn't Equipment-kind; let DiscoverAsync own the address space" which is the correct backward-compat path for Modbus / AB CIP / TwinCAT / FOCAS. Guarded on result.GenerationId being non-null — a fleet with no Published generation yet boots cleanly into a UNS-less address space and fills the registry on the next restart after first publish. Ctor on OpcUaServerService gained two new dependencies (DriverEquipmentContentRegistry + IServiceScopeFactory). No test file constructs OpcUaServerService directly so no downstream test breakage — the BackgroundService is only wired via DI in Program.cs. Four new DriverEquipmentContentRegistryTests: Get-null-for-unknown, Set-then-Get, case-insensitive driver-id lookup, Set-overwrites-existing. Server.Tests 190/190 (was 186, +4 new registry tests). Full ADR-001 Option A now lives at every layer: Core.OpcUa walker (#153) → ScopePathIndexBuilder (#154) → OpcUaApplicationHost wire-in (#155) → EquipmentNamespaceContentLoader (#156) → this PR's registry + Program.cs composition. The last pending loose end (full-integration smoke test that boots Program.cs against a seeded Config DB + verifies UNS tree via live OPC UA client) isn't strictly necessary because PR #155's OpcUaEquipmentWalkerIntegrationTests already proves the wire-in at the OPC UA client-browse level — the Program.cs composition added here is purely mechanical + well-covered by the four-file audit trail plus registry unit tests.