mirror of
https://github.com/OPCFoundation/UA-.NETStandard.git
synced 2026-10-22 09:37:12 +08:00
fix/encoder-stack-depth-json-flake
5 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 romanett
|
9b034f8291 |
Enable nullable reference types centrally; migrate all unset and annotations-only projects (#4700)
# Description Nullable reference types are now enabled once, in `common.props`, and every project in the repository builds warning-clean with them. **Configuration** - `<Nullable>enable</Nullable>` lives in `common.props`. All per-project `<Nullable>` settings are removed: 178 project files (enable, annotations, and the `disable` in the three Aot test projects). - 419 per-file `#nullable enable`/`restore` directives and the 4 `#nullable disable` directives are removed. - Kept on purpose: - The 5 `<auto-generated>` schema files (`UANodeSet.cs`, `UA Type Dictionary.cs`, `OPCBinarySchema.cs`, `SecuredApplication.cs`, `PInvokeHelper.cs`) keep `#nullable enable`. The compiler ignores the project setting in generated code, so the directive is required there. - `tests/Opc.Ua.Interop.LegacyPeer` keeps `disable`. It builds against the 1.5.378 packages and does not import the repository props. - `docs/DeveloperGuide.md` and `.github/copilot-instructions.md` now describe the central setting. **Migration** (the 43 unset and 13 annotations-only projects, roughly 10k diagnostics) - **Production code: annotated at the declarations.** Covers `tools/Opc.Ua.SourceGeneration.Core`, `src/Opc.Ua.Lds.Server` and the LDS sample. - Members, parameters and returns are nullable where null actually flows. - Null-check helpers carry `[NotNullWhen]` (`IsNull`, `IsValidXmlName`, `IsMethodTypeDesign`). - `CoreUtils.Format`, `Utils.Format`, `ITemplateWriter.WriteLine` take `params object?[]`. `Template.AddReplacement` takes `string?` (rendering already skips null replacements). - `IRegisteredServerStore.RegisterAsync` returns `Task<RegistrationEntry?>`: it returns null when a registration is removed. - Dereferences backed by an invariant (for example, the validated model) use `!`, so runtime behaviour is unchanged. - **Behavioural changes in production code** (small, deliberate): - `NodeSetToModelDesign` throws `InvalidOperationException` when the file is not a NodeSet. Before, it threw a NullReferenceException on the next line. - `ModelDesignValidator` starts its lookup tables empty instead of null. A redundant `if (m_nodes == null) return;` guard is gone. - **Test code: minimal annotations.** - `?` on values that can be null, `!` where the test asserts non-null. - `= null!` for fixture members assigned in SetUp. - `TryGetStructure<T>(out T? x)` where the inferred `T?` broke the `IEncodeable` constraint. - A few redundant `== null` checks on structs (`NodeId`, `ArrayOf<T>`) folded into the adjacent `IsNull`/`Count` checks. **Notes for reviewers** - The diff is large (~1,500 files) but almost entirely mechanical: `?`, `!` and directive removal. The non-annotation changes are listed above. - Known follow-up, not done here: `IModelDesign.Namespaces/TargetNamespace/NamespaceUris` stay non-nullable even though the validator returns null before it loads a dictionary. Making them nullable would ripple through every generator call site. ## Related Issues - Fixes #4439 ## Checklist - [ ] I have signed the [CLA](https://opcfoundation.org/license/cla/ContributorLicenseAgreementv1.0.pdf) and read the [CONTRIBUTING](https://github.com/OPCFoundation/UA-.NETStandard/blob/master/CONTRIBUTING.md) doc. - [ ] I have added tests that prove my fix is effective or that my feature works and increased code coverage. - [x] I have added all necessary documentation. - [x] I have verified that my changes do not introduce (new) build or analyzer warnings. - [ ] I ran **all** tests locally using the **UA.slnx** solution against at least .net **framework** and .net **10**, and all passed. - [ ] I fixed **all** failing and flaky tests in the CI pipelines and **all** CodeQL warnings. - [ ] I have addressed **all** PR feedback received. The "no new warnings" box covers net10.0 and net48: full non-incremental builds of `UA.slnx`, `tools/SourceGeneration.slnx`, `tools/MigrationAnalyzer.slnx` and the Aot test projects finished with 0 warnings and 0 errors under the repository's warnings-as-errors settings. net9.0 and net8.0 also had no nullable diagnostics, but they were not rebuilt after the final fixes. Tests have not been run locally yet. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
Marc Schier
|
4633aa7a34 |
Harden HA address-space hydration and replica-consistent NodeIds (#4445)
# Description Follow-up to #4390 for distributed address-space ownership, authoritative hydration, and replica-consistent wire NodeIds. Tracks #4444. Identity implementation: `a78d59067` (Server policy/lifecycle/adapter seam) and `07faf5cfd` (replica identity module, runtime wiring, examples and proofs). ## Replica identity - `UseReplicaNodeIdentity` / `ReplicaNodeIdFactory` reserve fixed shared namespace indexes before node-manager construction and reuse `DefaultNodeIdFactory`. ApplicationUri and built-in diagnostics/configuration remain local. - Replicated startup requires an explicit identity contract. Namespace layout, deterministic mode, replica-set identity and writer policy must agree. Standard factory views, replacement and registration preserve the policy. - Active/passive writer-assigned identities are replicated unchanged. Independent creation requires stable keys or explicit IDs. Live IDs and retained tombstones are reserved before writer allocation; no network/storage call runs in `New`. - Protected store descriptors use strong routing, including customized hybrid prefixes. Unknown legacy state and empty eventual scans cannot authorize automatic adoption. Verified-new hybrid provisioning is explicit. - Active/active frames compare the descriptor before CRDT merge. Rejected frames are not reported as applied; incompatible admission remains at ServiceLevel 0 through the standard service-level wiring. - Shared metadata validation covers type definitions, data types, references, permissions, NodeId/ExpandedNodeId/QualifiedName values, arrays/matrices, DataValues and encodeable structures/arguments, without reflection. - Runtime preparation hydrates retained IDs before replacement routes publish. Rebinding keeps the existing store/election or gossip transport/map and attaches to the new node-manager sources. Normal node-manager mutations, including AddNodes, reach the cached local-address-space adapter. - WoT materialization retains the stable configured factory. Samples expose genuinely factory-created nodes and a client `--identity` failover workflow. ## Preserved HA repairs The published `d99995003` and `79da74f14` fixes remain in place: - Shared linearizable sequence/lease coordination; no pretend CAS on CRDT and no private counter substituted for shared coordination. - Strict authoritative record decoding, protected pending reservations, safe snapshot publication and required delta retention. - Accepted topology state controls authoritative membership; stale deletion cannot remove a newer incarnation. - Independent subtree attachment tracking and correct disposal of the final snapshot task. - Values wait for their topology, and parent tombstone horizons prevent older descendant values crossing incarnations. A new deterministic regression also covers writer hydration preparation: temporary reconciliation removals must remain inbound application, not be recaptured as new shared deletion records. ## Validation ### Review feedback fixes `3a682155b` addresses the active-node notification, WoT namespace-index bounds, and retired-manager XML documentation feedback. The address-space adapter now notifies listeners with the registered active node, and both WoT managers reject missing/out-of-range namespace indexes before ownership is established. Targeted feedback validation passes on net10.0 and net48: 8 address-space adapter cases and 74 WoT cases per framework, plus 17 hydration/gossip/live-client integration cases on net10.0. The wrong-instance notification and missing namespace regressions were reproduced before the fixes. ### Identity implementation evidence | Selected scope | net10.0 | net48 | | --- | --- | --- | | HA repair, identity, metadata, peer admission and real-client identity/reload fixtures | 215 passed | 215 passed | | Factory, fluent authoring/import, NodeManagement and address-space adapter fixtures | 213 passed | 213 passed | | WoT node-manager fixtures | 68 passed | 68 passed | - Real UA-TCP clients compare full NodeIds/reference targets across independent servers in all four deterministic modes, with different manager/creation order and local allocations. Saved IDs work after endpoint failover and restart. - Real AddNodes-created IDs remain readable through successive live manager replacements in both replication modes, including nodes created after a reload. - Two process-level cases start three sample replicas, kill the serving process, verify a different serving ServerArray URI, and reuse cached factory-generated IDs for Read and CreateMonitoredItems without remapping or rebrowsing. - The identity/hydration scenario runs from the actual published win-x64 NativeAOT executable in the existing historian companion host. - The focused writer-recapture regression was observed failing without its fix, then passing. The real reload scenario was additionally repeated while diagnosing that race. - Changed-file formatting/style/analyzers were applied. Existing CA1508, CA1823, CA1861 and CA1850 warnings in untouched upstream files were not suppressed. The selected checks above do not replace full-solution CI, coverage gates or maintainer review, and do not claim that the entire solution or every native platform was run. #4444 remains the tracking issue. ## Contract and consistency limits Fixed identity does not make eventual payload reads linearizable. Hybrid views remain non-destructive, use explicit tombstones, and do not infer empty-store bootstrap or partition completion from absence. Compacted snapshots require authoritative storage. Failed/uncertain reservations require reconciliation. There is no automatic namespace-slot allocation, persistent logical-key mapping database, live renumbering, legacy-store wipe/adoption, or new hash algorithm. Opaque shared structures require a registered codec so their namespace-bearing fields can be validated. Application callbacks are still application-owned. ## Related issues - Tracks #4444. - Follow-up to #4390. - Builds on the merged NodeId factory (#4433), configure/seal (#4442) and lifecycle/import (#4427) integration at `aedd3f738`. ## Checklist - [ ] I have signed the CLA and read CONTRIBUTING. - [x] Added targeted regression and integration tests. - [x] Updated directly related API, HA, migration and sample documentation. - [x] Ran scoped formatting and the targeted framework checks listed above. - [ ] Full UA.slnx matrix and coverage gates completed. - [ ] All CI and maintainer review completed. --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 9dbd9527-d818-47a0-984c-fbb1472d3d6b |
||
|
Marc Schier
|
7411654c8d |
Complete OPC UA Part 11 Historical Access (#4390)
# Description Completes the OPC UA Part 11 Historical Access work requested by #4387 and the historian framework integration requested by #4400. The conformance/evidence catalog covers all 37 released UACore 1.05 Historical Access profiles: 15 Server facets and 22 Client facets, together with the standard aggregate functions and their conformance units. Server facets are advertised only when the resolved providers expose their required interfaces and capabilities. ## Server and provider support - Completes raw, modified, at-time, processed, annotation, structured, and event history. - Dispatches best-effort, bulk, and transactional updates with aligned statuses, old values, diagnostics, and rollback state. - Supports mixed history-update detail types and batched structured/annotation operations. - Uses composite identities and exclusive cursors for same-timestamp structured, modified, and event values. - Captures live data and events, installs historical configuration objects, and emits typed audit payloads. - Derives aggregate monitoring filters from per-node historian capabilities and rejects unsupported processed aggregates before provider dispatch. - Supports tri-state `Historizing`: set, clear, or preserve provider-owned state. ## Client support - Exposes modified values together with `ModificationInfo`. - Adds event-history reads and event insert/replace/update/delete operations. - Adds structured and annotation batching with strict response validation. - Shares continuation handling and release-on-abandon across read families. - Accepts inherited standard event fields rooted at event subtypes. ## Review hardening - Registers DI-selected historians before node-manager reconciliation and capability publication, while preserving provider ownership. - Uses committed Raft read barriers for linearizable `TryGetAsync` and `ScanAsync`. - Keeps claimed continuation state immutable until a fresh successor is saved, restoring the original after transient failures. - Uses exact-incarnation compare-and-delete so delayed cleanup cannot remove a restored same-ID continuation. - Cleans up indeterminate initial and replacement saves, and propagates shutdown cancellation through cleanup resolution. - Preserves legacy continuation envelopes/cursors and distinct annotation parent/property identities. - Deletes every structured entry at a requested timestamp and uses the complete modified-history ordering tuple. - Shares native/fallback at-time and AnnotationCount calculations, including reverse endpoints and cross-framework fractional intervals. - Primes modified aggregate items before buffered live delivery and deduplicates history/live overlap using provider-specific identities. - Keeps committed Modify results successful while reporting priming failures through the monitored item's notification queue. - Keeps protected-notification state live-only for queues larger than one. Queue size one retains the newest notification without pinning errors or transferring priority during resizing. Durable restore retains ordinary definitions, last values/errors, and raw queues without recreating priority. - Isolates aggregate calculation, modification preparation, and historical/live handoff in `AggregationFilterHandler`, retaining existing provider/factory wiring and supported past-start behavior. The larger-queue protection policy is unchanged pending the open review discussion. - Uses one `IMonitoringFilter` owner for protocol filters or the aggregation handler, eliminating parallel effective-filter storage and retaining the current definition during preparation. The original client filter stays distinct. - Integrates current master, retaining custom monitored-item creation decisions and per-node history callbacks. ## Redundant-server alignment and samples - Adds a protected shared historian with immutable segments, manifest CAS publication, writer fencing, and restart recovery. - Persists portable continuation state before exposing it to clients and transfers ownership during mirrored-session takeover. - Rejects eventual-consistency and active/active multi-writer historian configurations rather than silently degrading. - ReferenceServer and ConsoleReferenceClient demonstrate discovery, paging, raw/modified/at-time/processed reads, annotation/structured CRUD, event capture, and event CRUD. - RedundantServer and RedundantClient demonstrate active/passive history, continuation recovery after replica termination, and post-promotion writes. - Updates Historical Access, High Availability, profile, migration, and sample documentation. ## Compatibility - New 2.0 collection boundaries use `ArrayOf<T>` and public byte payloads use `ByteString`. - Removes the legacy public virtual `SamplingGroupManager.CreateMonitoredItem` / `ModifyMonitoredItem` entry points and protected creation factory, as agreed in review. Migration guidance points to the monitored-item manager and construction-decision APIs. - Retains established persisted formats and current write layouts. Retired interim notification-state formats (sample v2 / shared v4) are rejected, as agreed in review. - Keeps the historian integration compatible with .NET Framework and NativeAOT. ## Validation ### Latest CI source repair `73c619435` forwards the existing cancellation token at both predefined-node creation calls in the PR-added History event fixture. The merged lifecycle overload exposed these two CA2016 errors, which blocked History jobs, the solution build matrix, and CodeQL on GitHub and Azure. The exact compiler failure was reproduced locally; the repaired fixture passes 4/4 integration cases on net10.0 and 4/4 on net48 with clean scoped diagnostics. No assertions, tests, coverage thresholds, or CI configuration were weakened. The repaired History jobs on Linux/macOS and all four Azure solution builds now pass. The Windows all-TFM GitHub job instead lost its hosted runner's connection; no build log was recoverable, and GitHub recorded a runner-loss annotation rather than a compiler/test diagnostic. One unchanged retry was requested. Remaining checks are monitored with `gh pr checks --watch --fail-fast --interval 30`; CI is not yet reported as green. ### Latest base integration `bd52e5fea` merges upstream `master` at `da7412c42`. The lifecycle conflict retains upstream's shared cancellation-aware registration helper and the historian behavior. The hosted-historian fixture forwards its existing token to the new lifecycle overload. GDS project renames and all three AOT test hosts are retained, with their project references resolving after the merge. | Integrated scope | net10.0 | net48 | | --- | --- | --- | | Server lifecycle, fluent authoring/import, historian, monitored-item and hosting regressions | 732 passed | 732 passed | | NodeState lifecycle regressions | 216 passed | 216 passed | | Automatically merged Raft client fixtures | 28 passed | 28 passed | No selected tests failed or were skipped. The manually resolved files pass scoped formatting/analyzer checks. One CA1861 warning remains in the incoming `NodeSetImportIntegrationTests.cs`; that file is byte-identical to upstream master and was not rewritten or suppressed during integration. ### Latest feedback follow-up `102954eaa` replaces split filter references with the shared `IMonitoringFilter` contract implemented by protocol filters and `AggregationFilterHandler`. Targeted server regressions pass **103/103 on net10.0 and 103/103 on net48**; scoped formatting/analyzers are clean, and the updated XML-documentation inventory has zero gaps across 188 non-generated PR C# files. `93c5b5db8` renames `AggregationFilterHandler`, makes it the owner of the server-revised aggregate filter while preserving the original client filter, and removes the retired interim notification-state readers. Server regressions pass **101/101** and shared-store regressions **64/64**, on both net10.0 and net48, with no failures or skips. Runtime-file formatting/analyzers are clean. `5ccf45630` completes the public/internal XML-documentation sweep. The Roslyn inventory finds zero undocumented eligible declarations across all 186 non-generated PR C# files; the 110-file documentation slice changes only XML comments and spacing. `f17765475` applies the three requested Markdown edits. ### Earlier feedback follow-ups `a53950737` removes single-slot notification pinning and extracts aggregate implementation from MonitoredItem. Targeted sampling, aggregate modification, history/live handoff, queue/lifecycle, restore, and serialization cases pass **101/101 on net10.0 and 101/101 on net48**, without failures or skips. Scoped formatting/analyzers are clean; the focused review found no significant issues. The final formatted monitored-item fixtures also pass 45/45 on net48. The preceding `b21555568` removes the legacy sampling API bridge and durable notification-priority state. Its targeted Server cases passed 97/97 and shared subscription-store cases passed 64/64 on each framework. The broader evidence below remains pinned to its original tested commit, not attributed to these follow-ups. ### Earlier integrated-implementation validation Validated implementation commit: **`3efae597f`**. The following commit, `f6dbefb1c`, only adds two explanatory comment lines for the final standards review; it changes no executable code. | Scope | net10.0 | net48 | | --- | --- | --- | | Server | 5,187 passed; 56 skipped | Affected surface: 2,160 passed; 1 reproduced baseline failure; 56 skipped | | Client | 2,202 passed; 2 skipped | 2,204 passed; 5 skipped | | History integration | 519 passed; 21 skipped | 519 passed; 21 skipped | | Redundancy server | 641 passed | 641 passed | | Redundancy client | 124 passed | 124 passed | | Core shared-store regression suite | 23 passed | 23 passed | All six affected samples passed non-incremental Release/net10.0 builds with zero warnings and errors: ConsoleReferenceClient, ConsoleReferenceServer, RedundantServer, RedundantClient, MinimalBoilerServer, and PumpDeviceIntegrationServer. The published self-contained win-x64 NativeAOT executable is 116,924,928 bytes and has no CLR header or managed runtime dependency. Direct execution passed **HistoryAotTests 3/3** and **RaftAotTests 2/2**. Final standards and specification reviews found no confirmed outstanding high/medium functional gaps or hard standards violations. The remaining constructor-purpose clarification is addressed in `f6dbefb1c`. ### Reproduced baseline exceptions The net48 hosted-startup timeout also reproduces in a pristine worktree at the previous PR head, `24cf1416d`, in `RegisterPostStartRegistriesWiresHistorianAndAliasStoresOnPlainServerAsync`, with the same subsequent certificate-leak assertion. The corresponding renamed case fails before startup in the reviewed branch as well. This is reported separately from the net48 affected-surface results, not counted as a passing full Server suite. The net48 matrix selected the historian, aggregate, monitored-item, fluent, and continuation surfaces and excluded the Hosting namespace. `DurableDataValueQueueVerifyReferenceBatchingAsync` failed at item 1000 in the net48 affected-surface run and passed its single isolated rerun. The same failure at the same assertion reproduces on pristine `849e16acc`; the test, sample queue, queue factory, and batch persistor are unchanged. The original failure remains in the table rather than being erased by the rerun. ### CI source repairs - `4be7eb8f7` keeps diagnostics/configuration managers out of distributed address-space replication. Those managers inherit the source interface, but their live server state and redundancy metadata must remain replica-local. The ownership regression passes on net10.0 and net48, and all nine redundancy sample CI tests pass locally. - `d17f467b3` completes the sample startup repair: core managers are also replica-local; the startup banner no longer advertises readiness before initialization; the producer starts after distributed services; and native Raft startup no longer forces disruptive extra elections. The full sample suite passed 9/9, with initialized `RedundancySupport=Hot`, and focused Raft/address-space regressions passed 8/8 and 3/3. - `81b0b72be` moves the unchanged distributed-historian NativeAOT scenario into a smaller, always-run companion executable to bound Apple's linker input. All 137 AOT tests remain present; the moved test and helpers are unchanged. GitHub and Azure publish/run both native executables, with separate result directories where they share a job. The companion passed actual Windows NativeAOT execution. Linux and Intel macOS CI passed both executables, but this split alone did not fix ARM64 macOS linking. The remaining historian graph is rooted by ReferenceServer startup, not just the relocated test. No test exclusion or speculative linker setting was added to conceal it. - `8a5842913` isolates the independent MCP test graph into an always-run native host. The existing MCP source and assembly leak hooks are linked unchanged, and evaluated compiler inputs prove exactly one owner for every test source: 133 main + 3 MCP + 1 historian = 137 declarations. All three hosts run on every AOT platform. The main Windows native object is 9.35% smaller; both changed publishes had zero diagnostics. MCP passed 3/3; the identical main executable passed 133/133 on one unfiltered followup after two initial endpoint-discovery connection failures, which remain recorded in the local evidence. - The previous Intel macOS job successfully published and executed its native tests, then failed on artifact-upload DNS (`ENOTFOUND`). No upload gate, shared environment setting, or cache configuration was disabled. The first repair round also encountered the unrelated `MetricsAreEmittedForChannelLifetimeAsync` timing failure on macOS Core. Neither that test nor channel-manager code is changed by this PR, and its assertions remain intact. It passed unchanged on the final repair head. On `d17f467b3`, the Linux and macOS redundancy sample jobs pass; the remaining failure was ARM64 Apple's `too many large addends` assertion. The verified upstream relocation-anchor fix (dotnet/runtime#124721, backport dotnet/runtime#132171) is not present in the shipping ILCompiler 10.0.11. An experimental multi-module probe failed and was rejected, not added to CI. On `8a5842913`, [`aot-macos-latest`](https://github.com/OPCFoundation/UA-.NETStandard/actions/runs/34102158680/job/101678934057) successfully publishes and executes all three native hosts on ARM64: 136 main runtime cases, 1 historian case, and 3 MCP cases, with no skips. The previous Apple linker blocker is resolved. Linux NativeAOT also passes. No assertion, test, shared environment setting, cache, or required gate was weakened. The same CI run reports a separate macOS Sessions teardown failure: all 662 executed tests pass, but the global leak assertion detects one undisposed certificate. The MCP partition changes no runtime, Sessions, or test-fixture code. A single unchanged diagnostic rerun is queued once GitHub permits rerunning that job; the failure is not suppressed or counted as green. Other CI checks are still pending. ## Related issues - Fixes #4387 - Fixes #4400 ## Checklist - [ ] I have signed the CLA and read the contributing guide. - [x] I have added regression tests and directly related documentation. - [x] I have retained working reference and redundant-server sample wiring. - [x] Final validation and review evidence is complete, including the baseline exceptions above. - [ ] All required CI checks pass; the fail-fast rerun for `73c619435` is pending. --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 9dbd9527-d818-47a0-984c-fbb1472d3d6b |
||
|
Marc Schier
|
4e0b82af5f |
Add IServerContext and move server startup tasks onto it (#4193)
## Stacked on #4188
Base is `marcschier/server-seam-analysis`, not `master`.
`IServerInternal.cs` and
`ServerInternalData.cs` both diverge from master on that branch (chunk
4c removed the two
`object` locks), so branching off master would conflict on both files
*and* design
`IServerContext` against a stale interface. GitHub retargets the base to
`master`
automatically when #4188 merges. **Review #4188 first.**
Closes part of #4184.
## What this does
`IServerInternal` is a service locator: 56 members, 30-odd of which hand
out one of 20+
subsystems. A component that only needs to observe the server and act on
it has to take the
whole thing, and no signature tells a reader which parts a component
actually touches.
This adds `IServerContext` — the ambient view of a running server — and
moves
`IServerStartupTask` onto it.
```csharp
public interface IServerContext
{
ServerState CurrentState { get; }
ServerObjectState ServerObject { get; }
IServiceMessageContext MessageContext { get; }
ServerSystemContext DefaultSystemContext { get; }
ServerSystemContext CreateSystemContext(ISession session);
T? FindPredefinedNode<T>(NodeId nodeId) where T : NodeState;
IEnumerable<T> FindNodeManagers<T>() where T : class;
ValueTask ReportEventAsync(IFilterTarget e, CancellationToken ct = default);
ValueTask CloseSessionAsync(OperationContext context, NodeId sessionId,
bool deleteSubscriptions, CancellationToken ct = default);
ValueTask DeleteSubscriptionAsync(uint subscriptionId, CancellationToken ct = default);
void UpdateServerDiagnostics(Action<ServerDiagnosticsSummaryDataType> update);
}
```
Nothing on it returns a subsystem. A component that needs one takes it
as a constructor
dependency, which states the requirement instead of hiding it behind an
ambient handle.
`IServerInternal : IServerContext`, so every existing handle already
satisfies the new seam
and consumers migrate one at a time.
## The member set was measured, not guessed
Every member of `IServerInternal` was counted by read site across `src`,
`samples` and
`tests`, split by area. `Ext` is non-test production consumers, which is
what decides
whether something belongs on an ambient interface.
| Member | InServer | OtherSrc | Sample | Test | **Ext** | On
`ISystemContext`? |
|---|---:|---:|---:|---:|---:|---|
| `NamespaceUris` | 63 | 50 | 34 | 131 | **84** | yes |
| `Telemetry` | 58 | 23 | 35 | 5 | **58** | yes |
| `MessageContext` | 2 | 8 | 7 | 11 | **15** | no |
| `TypeTree` | 42 | 3 | 11 | 10 | **14** | yes, as `TypeTable` |
| `NodeManager` | 66 | 7 | 6 | 74 | **13** | no |
| `DefaultSystemContext` | 21 | 4 | 1 | 7 | **5** | no |
| `IdentityRegistry` | 1 | 4 | 1 | 1 | **5** | no |
| `SubscriptionManager` | 23 | 0 | 0 | 17 | **0** | no |
| `SessionManager` | 21 | 0 | 0 | 15 | **0** | no |
| `RequestManager` | 10 | 0 | 0 | 23 | **0** | no |
| `AggregateManager` | 12 | 0 | 0 | 2 | **0** | no |
| `ModellingRulesManager`, `ConformanceUnitsManager`,
`DiagnosticsEnabled` | 0 | 0 | 0 | 0 | **0** | no |
Two things fall out. The highest-traffic members are already carried by
`ISystemContext`,
which `ServerSystemContext` copies off the server at construction — so
most of those 278
`NamespaceUris` reads are callers reaching through the server for
something they already
hold. And the subsystems with zero external demand are exactly the ones
that should be
injected rather than published.
The check that matters: the eight `IServerStartupTask` implementations
in
`Opc.Ua.Redundancy.Server` are the heaviest real consumer of the ambient
handle, and
between them they touch **eight distinct members**. That is the shape
the interface is cut
to.
## Commits
### `e4a912eb` Remove seven unused members from `IServerInternal`
The deletion test, applied before anything is designed against the
interface.
| Removed | Evidence |
|---|---|
| `CloseSession(OperationContext, NodeId, bool)` | zero call sites;
sync-over-async; already obsolete |
| `ReadServerDiagnostics<TResult>` | zero production sites — **added one
commit earlier in #4188 for symmetry and never called** |
| `ModellingRulesManager` | property never read; `ServerInternalData`
owns the lifetime |
| `ConformanceUnitsManager` | same |
| `DiagnosticsEnabled` | façade over
`IDiagnosticsNodeManager.DiagnosticsEnabled`, which is what callers use
|
| `Status` | one site, itself inside an obsolete method |
| `ServerDiagnostics` | only reachable through the diagnostics methods
that replaced it |
`EndpointAddresses` was kept — `Opc.Ua.Di.Tests` uses it to resolve an
endpoint URL. Thin,
but real.
The only build fallout was three lines in one test assembly.
`Opc.Ua.Redundancy.Server`,
`Opc.Ua.Redundancy.Kubernetes`, `Opc.Ua.Gds.Server.Common`,
`Opc.Ua.WotCon.Server`,
`Opc.Ua.PubSub.Server`, `Opc.Ua.Di.Server`, `ConsoleReferenceServer` and
`Quickstarts.Servers` all built untouched. That is the evidence the
members were dead.
### `ffcc80b9` Add `IServerContext`
Additive. `ServerInternalData` implements it; no caller changes.
### `e8c4e15f` Hand startup tasks `IServerContext`
Nine implementations, `OpcUaServerHostedService`, and the test call
sites.
### `1b4de97a` Drop the `IServerStartupTask` migration-guide section
Review feedback. `IServerStartupTask` does not exist in `master378`, so
it is a 2.0-only
concept and nobody migrating from 1.5.378 can own an implementation of
it. The **Removed
members on `IServerInternal`** section stays - those members *are* in
master378 - and is now
self-contained.
### `2ab7d689` Merge `marcschier/server-seam-analysis`
Picks up the base branch including its merge of master. Textually clean,
no conflicts.
### `6a650f12` Migrate `RobotIntentHostStartupTask` to `IServerContext`
The merge was textually clean but **semantically broken**: master's
Robot Intent feature
(#4165) added a tenth `IServerStartupTask` implementation written
against the old
`IServerInternal` signature, so it no longer satisfied the interface.
Its body walked
`server.NodeManager.AsyncNodeManagers` and type-tested each entry for
`RobotIntentNodeManager` - exactly the pattern `FindNodeManagers<T>()`
replaces - so it
migrated the same way as the other nine. `FindNodeManagers` also
searches the synchronous
registry and de-dupes by reference, so a manager registered on both
faces still starts its
intent controller hosts exactly once.
## Three findings from migrating real consumers
**`MessageContext` is not derivable.** The design had callers use
`DefaultSystemContext.AsMessageContext()`, on the reasoning that a
message context can be
built from a system context. It can't: that conversion carries the
namespace tables and the
factory but **not** `MaxStringLength`, `MaxArrayLength` or
`MaxByteStringLength`. Four
replication paths in `Opc.Ua.Redundancy.Server` would have silently
fallen back to default
decoding limits — a component quietly accepting more than the server is
configured to. It
is a real member, and the interface documents why.
**`new`-shadowed members split the mock surface.** `IServerInternal`
first redeclared
`DefaultSystemContext` with a narrower type and `CurrentState` with a
setter. That
compiles, but a shadowed member is a *second interface slot*: production
code calling
through `IServerContext` read a slot that every existing
`Mock<IServerInternal>` leaves
unstubbed. Seventeen tests went from passing to `NullReferenceException`
**while still
compiling**. The narrower typing bought no encapsulation anyway — the
object really is a
`ServerSystemContext` and any caller can cast — so both shadows are
gone. `CurrentState` is
read-only on the ambient interface and the one site that assigns it uses
the concrete type,
as `GetStatus` already does.
**`IServiceProvider` is deliberately *not* on `OnServerStartedAsync`.**
The plan sketched
it. All nine implementations already receive their dependencies through
their constructors
from the builder extensions, so a container on the callback would have
replaced one service
locator with another.
## `FindNodeManagers<T>`
Two tasks walked `server.NodeManager.NodeManagers` and type-tested each
entry for
`ILocalAddressSpaceSource`; two more reached for a specific manager to
cast it to a
capability. Asking for the capability says what the caller wants; asking
for the registry
says how to find it.
```csharp
// before
foreach (INodeManager nodeManager in server.NodeManager.NodeManagers)
{
if (nodeManager is not ILocalAddressSpaceSource source) { continue; }
...
}
// after
foreach (ILocalAddressSpaceSource source in server.FindNodeManagers<ILocalAddressSpaceSource>())
{
...
}
```
It searches both the asynchronous and synchronous registries:
`ToSyncNodeManager()` returns
`this` when a manager already implements `INodeManager`, so the two
lists overlap and a
capability may sit on either face.
## Tests
Mock harnesses are **wired, not stubbed** — they return the node
managers and contexts the
tests already build, so the tests keep asserting against real behaviour
rather than against
a mock that returns whatever is convenient. Four new tests cover
`CreateSystemContext`
(identity/locale propagation, null guard), the explicit
`DefaultSystemContext` identity,
and `FindPredefinedNode` before the address space is bound.
| Suite | net10.0 | net48 |
|---|---|---|
| `Opc.Ua.Redundancy.Server.Tests` | 573 passed | 573 passed |
| `Opc.Ua.Redundancy.Kubernetes.Tests` | 97 passed | n/a (net8/net10
only) |
| `Opc.Ua.Server.Tests` | 4028 passed, 1 failed | 225 passed, 1 failed
(filtered) |
The single failure is
`ConfigureApplicationBuildsSharedClientAndServerConfigurationAsync`,
a local certificate-store problem. Verified it reproduces unchanged at
HEAD with these
changes stashed — it is not caused by this PR.
All twelve dependent assemblies and samples build clean on both TFMs,
zero warnings.
### Re-verified after the base merge (`6a650f12`)
`Opc.Ua.Server`, `Redundancy.Server`, `Redundancy.Kubernetes`,
`Robotics.Server`,
`WotCon.Server`, `Di.Server`, `PubSub.Server`, `Gds.Server.Common` and
`XRegistry.Server` all build clean on net10.0. `Opc.Ua.Robotics.Tests`
491 passed,
`Opc.Ua.Redundancy.Server.Tests` 573 passed,
`Opc.Ua.Redundancy.Kubernetes.Tests` 97 passed.
### CI status: 119 passed, 3 failed - none caused by this PR
| Failing job | Why it is not this PR |
|---|---|
| `SourceGeneration_Core_Tests_windows` (net48) |
`GenerateAndCompileDesignFileAsync(..., CSharp11, ...)` fails with
`CS9058: Feature 'Lock object' is not available in C# 11.0` because
generated `Opc.Ua.NodeStates.g.cs` emits `System.Threading.Lock`.
**Identical failure on base #4188 (build 16737)**, which contains none
of these commits. This PR touches no source-generation file. |
| `SourceGeneration_Core_Tests_windows` (net10.0) | Same root cause;
also failing on base #4188. |
| `PubSub_Eth_Tests_windows` (net48) | Agent DNS outage:
`NuGetToolInstaller` died with `ERR:getaddrinfo ENOTFOUND
dist.nuget.org`. The job aborted before restore/build, so **the test
step never executed**. Infra flake; needs a re-run, not a code change. |
Both `SourceGeneration.Core` failures belong to the base branch and
should be fixed there
rather than papered over here.
## Correction to the recorded plan
Chunk 7 (the frozen bind phase) was recorded as the largest mechanical
footprint remaining
and the one most likely to stall. It isn't: **each of the 12 `Set*`
mutators has exactly one
call site**, in the `StandardServer` wiring, plus one sample. It is a
~13 call-site change.
The 268-file figure that produced that assessment counted files
*referencing*
`IServerInternal`, which chunk 7 never has to touch, because the
interface survives and only
the mutators move off it.
## Docs
`docs/MigrationGuide.md` gains one section: the removed
`IServerInternal` members with their
replacements, plus the note that `MessageContext`,
`DefaultSystemContext`, `CurrentState`,
`ServerObject`, `ReportEventAsync`, `CloseSessionAsync`,
`DeleteSubscriptionAsync` and
`UpdateServerDiagnostics` moved *down* to `IServerContext` and stay
reachable through
`IServerInternal` unchanged, so no call site has to move.
An earlier revision also documented the `IServerStartupTask` signature
change there. That
was dropped after review (
|
||
|
Marc Schier
|
7cec2d614d | Move repository files to the consolidated layout |