Skip to content

DeveloperGuide.md

Latest commit

 

History

History
560 lines (401 loc) · 57.4 KB

DeveloperGuide.md

File metadata and controls

560 lines (401 loc) · 57.4 KB
Raw

Developer Guide

This guide is the starting point for contributing to the OPC UA .NET Standard stack. It covers what to install, how to build and test, the coding standards ("dos and don'ts"), and task-oriented "how to" recipes (starting with how to add logging). It links out to the topic-specific documents in docs/README.md rather than repeating them.

If you are new here, read the sections in order: Prerequisites → Repository layout → Building → Running tests → Coding standards. The How-to guides and Packages, platform support, and versioning sections are reference material you can jump to as needed.

Contents

Prerequisites

  • .NET SDK 10.0 — the whole repository builds and restores with the .NET 10 SDK. Older SDKs are not supported for building main. The class libraries still target older frameworks (see Packages, platform support, and versioning), but you build them with the .NET 10 SDK.
  • An IDE (optional but recommended) — Visual Studio 2026, Visual Studio Code with the C# Dev Kit, or JetBrains Rider. Everything can also be done from the command line with dotnet.
  • git — to clone and to create feature branches.
  • Docker Desktop (optional) — only needed to run the containerized reference server; see ContainerReferenceServer.md.

The C# language version is pinned (LangVersion 14) and analyzer/style rules are enforced by the build, so no extra tooling install is required to get the same diagnostics locally that CI produces.

Repository layout

Path Contents
src/ The core stack and higher-level libraries: Opc.Ua.Types, Opc.Ua.Core*, Opc.Ua.Client, Opc.Ua.Server, Opc.Ua.Configuration, Opc.Ua.PubSub (+ transports), the GDS / DI / LDS / WoT libraries, and the Opc.Ua.Redundancy* family.
samples/ Reference and sample apps: ConsoleReferenceServer, ConsoleReferenceClient, Quickstarts.Servers, the Minimal* / PumpDeviceIntegrationServer NativeAOT samples, Redundant*, etc.
tests/ Unit and integration test projects, mirroring the library structure, plus shared test frameworks.
tools/ Source generators, migration analyzers, and the installable Opc.Ua.Mcp tool. Each analyzer and generator has a build project and — for the source generators — a *.Pack project that packages it under a Roslyn-versioned analyzer folder.
docs/ This documentation set (indexed by docs/README.md).
fuzzing/ SharpFuzz / libFuzzer fuzz targets (see Fuzzing.md).

Central build configuration lives at the repository root and is imported by every project:

  • UA.slnx — the solution containing all projects.
  • Directory.Build.props / Directory.Build.targets — global MSBuild properties and targets.
  • Directory.Packages.props — Central Package Management: every NuGet version is declared here.
  • common.props / targets.props — shared properties, analyzer settings, and the target-framework matrix.
  • .editorconfig — the authoritative code-style and analyzer-severity rules (enforced at build time).

Building

From the repository root:

dotnet restore UA.slnx
dotnet build UA.slnx

Notes:

  • Warnings are errors. TreatWarningsAsErrors is enabled, so compiler (CSxxxx) and Roslynator (RCSxxxx) diagnostics fail the build. Microsoft Code Analysis (CAxxxx) diagnostics are emitted as non-fatal warnings unless a rule is promoted to error in .editorconfig. Fix all of them before opening a pull request.

  • Building a single target framework. By default the libraries multi-target the whole matrix (see Packages, platform support, and versioning). To restrict a local build to one framework, pass -p:CustomTargetFrameworks, for example:

    dotnet build src/Opc.Ua.Core/Opc.Ua.Core.csproj -f net10.0 -p:CustomTargetFrameworks=net10.0

  • Offline / restricted networks. NuGetAudit is enabled and fails the build with NU1900 when it cannot reach the audit service. If you build offline, pass -p:NuGetAudit=false.

  • Source generators are consumed as project references. Projects that use the in-repo generators reference tools/Opc.Ua.SourceGeneration[.Stack] with OutputItemType=Analyzer. MSBuild only hands the compiler the generator assembly itself, so Directory.Build.targets adds the generator's runtime closure (its output directory, minus the Roslyn host assemblies) as Analyzer items — the same payload the generator NuGet packages ship under analyzers/dotnet/<roslyn>/cs. Without it the generators cannot resolve their dependencies and fail to initialise with CS8784.

  • Analyzers and generators are shipped under a Roslyn-versioned analyzer folder. roslyn.props pins the Roslyn API version, and each generator has a *.Pack project that ships it under analyzers/dotnet/roslyn<major>.<minor>/cs. See Repository layout and the support matrix. Because the repository's own projects consume that same build, building this repository requires a Roslyn 5.x host (the .NET 10 SDK or Visual Studio 2026).

Running tests

Run the whole suite from the solution:

dotnet test UA.slnx

Conventions and requirements:

  • Frameworks. Test projects use either NUnit (with Assert.That assertions and Moq for mocking) or TUnit (with its own assertions and mock helpers). Do not mix the two in one project, and do not use the classic NUnit asserts (Assert.AreEqual, …).

  • Coverage. Coverage is measured with Coverlet and must not regress; every non-application, non-test project should stay at or above 80 %. Two gates enforce this in CI — see Continuous integration.

  • Integration tests. Client/server and pub/sub features need integration tests as well as unit tests. A feature library's integration tests normally live with its unit tests in <Component>.Tests, for example Opc.Ua.Robotics.Tests, and every test project name ends in .Tests. Split integration tests into a separate project only when they run long, destabilise the unit tests, or the suite needs further division. Keep them deterministic: allocate a free port per fixture rather than hard-coding one, wait on the actual signal instead of using Thread.Sleep as a synchronisation primitive, and dispose every session, subscription and server in teardown including on failure. A flaky integration test is worse than none.

  • Test output. Keep successful-test output compact. NUnit and the test runner retain captured output and create additional copies when forwarding and serializing it to TRX; repeated exhaustive dumps can consume gigabytes of memory, inflate results artifacts, and stall the runner's shared socket (issue #4213). For small failure-only diagnostics, buffer the dump and emit it only when the test does not pass; see EncoderCommon.TestOutput in tests/Opc.Ua.Core.TestFramework/EncoderCommon.cs. Stream large dumps to files and register them with TestContext.AddTestAttachment instead of building a large string or writing every line to TestContext.Out. CommonTestWorkers.BrowseFullAddressSpaceWorkerAsync(..., outputResult: true) follows this pattern: the full listing is a test-result attachment, while inline output contains only the reference count. Keep attachment files until the runner has collected them.

  • Certificate leak diagnostics. Set OPCUA_CERTIFICATE_LEAK_TRACKING=1 before starting a test process to capture allocation stacks for Certificate handles that are not explicitly disposed. The assembly-level leak detector includes live and unreachable outstanding handles, their allocation stacks, and the NUnit fixture where available. Tracking is enabled by default in Debug builds and opt-in in Release because stack capture adds per-certificate overhead. The equivalent AppContext switch is Opc.Ua.Security.Certificates.CertificateLeakTracking; set it before the first Certificate is created. For example, in PowerShell:

    $env:OPCUA_CERTIFICATE_LEAK_TRACKING = '1'
dotnet test `
    tests\Opc.Ua.Sessions.Tests\Opc.Ua.Sessions.Tests.csproj `
    -f net10.0 `
    -p:CustomTestTarget=net10.0

  • Before a pull request the UA.slnx suite must pass on at least .NET Framework 4.8 and .NET 10.0.

  • Testing a specific target framework. The libraries multi-target, but the test executables run on one framework at a time. To run the suite against a non-default framework, set CustomTestTarget (supported values: net48, net8.0, net9.0, net10.0). The batch file tests/customtest.bat cleans, restores, and runs the tests for a chosen target; in Visual Studio, uncomment and set the CustomTestTarget property in targets.props. A clean build for the target is recommended when switching.

  • CI matrix. The pull-request gate runs the test suite on net48 and net10.0, and compiles the solution for every supported target framework; the remaining test matrices (Debug, .NET 9/8) run in scheduled or manual CI. Fix all failing, flaky, and CodeQL findings in the pipelines. See Continuous integration.

Wire compatibility with the released 1.5 stack is covered by tests/Opc.Ua.Interop.Tests, in both directions (2.0 client to 1.5 server and 1.5 client to 2.0 server):

  • sessions over None, the RSA policies and the ECC policies, with anonymous and user name logon, and the read, write, browse, call and subscription services;
  • the existing 2.0 ClientTest workers that do not depend on the in-process server, run against the 1.5 Quickstarts reference server (ClientTestFramework.ExternalServerUrl);
  • custom data types (structures, unions, optional fields, enumerations) loaded, decoded and written back with the other side's complex type system;
  • certificate trust with auto-accept off: untrusted peers, CA-issued certificates, revoked certificates;
  • message sizes and service limits: multi-chunk messages, values above the peer's encoding limits, responses above the client's MaxMessageSize, operation limits, browse continuation points and security token renewal.

The 1.5 side is tests/Opc.Ua.Interop.LegacyPeer, a console application built from the OPCFoundation.NetStandard.Opc.Ua.* NuGet packages and started as a child process, because the two stacks share assembly names and cannot be loaded into one process. It deliberately does not import the repository build settings. In server mode it hosts either a small interop server with deliberately tight limits or the 1.5 Quickstarts reference server; in client mode it runs named checks and prints one JSON result line per check, which the tests report as separate results. The test project builds it and copies it to legacy-peer/ next to the tests; pass -p:LegacyStackVersion=<version> to test another 1.5 release, or set OPCUA_INTEROP_LEGACY_PEER to the path of a prebuilt Opc.Ua.Interop.LegacyPeer.dll. 1.5.378 cannot reload a certificate it has just created in a long store path, so the tests keep their PKI folders directly below the temp folder. The project is a *.Tests.csproj, so CI runs it on every pull-request test leg like any other test project.

Channel recovery regressions use ManagedSessionReconnectTests in the client test project and ClientChannelManagerManagedTests / ReconnectDeadlineTests in the core test project. The composed fixture scripts an in-process transport while retaining the real session, V2 publishing workers and channel ready gate. Use phase signals and the injected clock rather than a server-process kill or sleeps. Keep every phase wait bounded. When checking the recreate/drain wiring, disable the managed-session channel deadline so outer takeover cannot mask a broken drain, and verify the parked Publish attempt has unwound before allowing replacement-session creation to finish.

Coding standards (dos and don'ts)

All rules apply to new code and to existing code you touch. The .editorconfig is authoritative and enforced at build time; the highlights below are the ones most often missed.

Formatting and style

  • Add the OPC Foundation MIT license header to every new source file.
  • 4-space indentation, max line length 120, CRLF line endings, UTF-8, final newline, no trailing whitespace.
  • Allman braces; always specify access modifiers explicitly; member order is constructors → properties/events → methods → fields, each public → protected → internal → private.
  • Do not use #region/#endregion or comment-only section dividers. Do not add #nullable enable to a file when the project already sets <Nullable>enable</Nullable>.
  • Put every XML-doc <summary> text on its own line (never a single-line /// <summary> … </summary>).
  • Follow standard C# naming; no underscores in method or test-method names (tests use PascalCase).

API and language

  • Async only. New code uses async/await (TAP). Do not add APM or sync-over-async (.Result, .Wait(), GetAwaiter().GetResult()) unless explicitly requested.
  • No object in public API (except when overriding Equals). For OPC UA values use Variant.
  • INullable types must not be wrapped in System.Nullable<T> (T?); use .IsNull / .Null instead. On struct types prefer TryGet/TryGetValue over casting; never use Variant.AsBoxedValue or IUnion.Value.
  • Prefer ArrayOf<T> over read-only collection types / IReadOnlyList<T> / arrays in new public API; prefer ByteString over byte[]; prefer Span<byte>/ReadOnlySpan<byte> over byte[].
  • Do not use [Obsolete] API (outside test code) and do not add API that is not NativeAOT-compatible.
  • Maintain backward compatibility with 1.5.378; mark replaced API [Obsolete] rather than removing it.

Concurrency

  • Never expose locks in any API surface. For a synchronous lock use System.Threading.Lock (a polyfill is provided for older TFMs) — never private readonly object m_lock = new(). Prefer SemaphoreSlim where async coordination is needed.

Architecture

  • Make non-abstract public classes sealed by default; prefer a provider model with injectable providers over inheritance.
  • Wire new functionality into the dependency-injection infrastructure (with a direct "construct it yourself" fallback) and expose it through the fluent API where possible.
  • Reuse the existing base services (telemetry, file system, certificate/secret stores, state machines, sessions, source generators, …) instead of re-implementing them.

Security

  • Never hardcode credentials, certificates, or secrets. Manage certificates through the certificate store system and secrets through the secret store (see CertificateManager.md and Certificates.md).
  • Use only SHA-2 or stronger hash algorithms; use the audit and redaction APIs for sensitive data.

Logging — use source-generated logging; never call ILogger.LogInformation/LogError/… directly. See Add a log message (source-generated).

How-to guides

Add a log message (source-generated)

The stack uses LoggerMessageAttribute source-generated logging everywhere. It avoids boxing value-type arguments, caches the message formatter, and emits an IsEnabled check so a disabled level costs nothing. Direct ILogger.LogInformation/LogError/… calls are not allowed. The runtime/observability side (how the ILogger is created from ITelemetryContext) is documented in Diagnostics.md; this section is the authoring recipe.

Recipe

  1. Get a logger. Obtain an ILogger from the ambient ITelemetryContext (telemetry.CreateLogger<T>()); most types already hold one in an m_logger field.
  2. Find or create the log class. Each file that logs has, at its end, an internal static partial class <PrimaryClass>Log holding [LoggerMessage] extension methods on ILogger. Add your message there. If several closely-related files emit the same messages, use one shared <Area>Log class instead of duplicating (for example the encoders/decoders in Opc.Ua.Types share EncodingLog).
  3. Reserve an event id. Each project has one internal static class <AssemblyToken>EventIds at its root (see Event-id convention). Use the existing per-class offset.
  4. Declare the message. Add a partial method with [LoggerMessage(EventId = <AssemblyToken>EventIds.<Class> + <index>, Level = LogLevel.<Level>, Message = "…")] (see Log class convention).
  5. Call it. Replace the old logger.LogXxx(...) call with logger.<MethodName>(args).

Event-id convention

Each project owns exactly one event-id class, named <AssemblyToken>EventIds, in namespace Opc.Ua, in a file EventIds.cs at the project root. <AssemblyToken> is the assembly name with the Opc.Ua. prefix removed and dots dropped — for example Opc.Ua.Core → CoreEventIds, Opc.Ua.Core.Types → CoreTypesEventIds, Opc.Ua.Client → ClientEventIds.

The token prefix is required because the stack uses InternalsVisibleTo: two internal classes with the same name in the same namespace collide across an IVT boundary (CS0436). The class holds one public const int offset per log class. Offsets are assigned in class-alphabetical order starting at 0; each block reserves at least five spare slots for future messages and is then rounded up to the next multiple of ten, so ids stay documented and managed in one place. Every log method sets EventId = <AssemblyToken>EventIds.<Class> + <zero-based message index within that class>.

Narrow exception: retained EventSource-compatibility ids

The four legacy System.Diagnostics.Tracing.EventSource providers (OPC-UA-Core, OPC-UA-Client, OPC-UA-Server, Opc.Ua.ChannelManager) were removed and replaced with [LoggerMessage] equivalents. Their compatibility log methods are a deliberate, narrow exception to the convention above: each keeps the exact numeric id, event name, level, message template, and structured fields the corresponding ETW event had, so consumers can preserve event identity when they migrate from ETW to ILogger. Concretely:

  • The compatibility log class uses the old provider name as its ILogger category (e.g. "OPC-UA-Core", "Opc.Ua.ChannelManager") instead of the typed, per-class category used elsewhere in the project.
  • EventId resolves to the literal legacy numeric id (e.g. 10 for OPC-UA-Core's former ServiceCallStart) rather than a normal per-class offset. Keep these values in the affected project's EventIds.cs; compatibility ids are scoped to their own logger category, so they may intentionally overlap ordinary per-assembly values.
  • Every compatibility method sets EventName explicitly ([LoggerMessage(EventId = 10, EventName = "ServiceCallStart", Level = LogLevel.Trace, Message = "...")]) so EventId.Name matches the original ETW event name exactly.
  • ETW-only metadata (provider GUID, Task, Keywords, manifest) is not retained because there is no ILogger equivalent.
  • Do not use this pattern for new log messages. It exists only to preserve the event identities that previously shipped through the four EventSource providers; see Diagnostics.md and migrate/2.0.x/telemetry.md for the full removal/compatibility mapping.

Log class convention

  • One log class per file, named <PrimaryClass>Log, internal static partial, appended at the end of the file inside the same namespace.
  • Methods are extension methods on ILogger (public static partial void <Name>(this ILogger logger, …)) so call sites read naturally as logger.<Name>(…).
  • Identical this ILogger overloads (same name and parameter types) declared in more than one class of the same namespace collide (CS0121) — deduplicate them into a single shared <Area>Log class. Overloads that differ by name or by parameter type are fine.

Message, level, and parameter rules

  • Message text is exact and static; use named placeholders ({ChannelId}) that match a parameter of the same name. Never interpolate ($"…"). An Exception argument is detected by its type and does not need a placeholder.
  • Parameter types must match the real argument type. Do not use object/object?, and do not call .ToString() on an argument (type the parameter instead, e.g. an enum or int); an unnecessary .ToString() trips RCS1097/CA1305. Declare a parameter nullable (string?, Uri?, Exception?) only when the argument can actually be null, otherwise the compiler reports CS8604.
  • Guard only expensive arguments. If a call passes an expensive computed argument (string.Join(...), a LINQ projection, .ToString() on a complex object) wrap it in if (logger.IsEnabled(<level>)); source generation does not suppress eager evaluation of the arguments, and CA1873 flags it. Do not guard cheap arguments (locals, fields, ids) — over-guarding trips RCS1006/RCS1061. A guard must never gate an expression that has an observable side effect.
  • Dynamic levels stay hand-written. [LoggerMessage] needs a compile-time Level. A call whose level is only known at runtime keeps the structured logger.Log(logLevel, "{Template}", args) form wrapped in if (logger.IsEnabled(logLevel)). These are the only remaining direct ILogger.Log calls.
  • Shared/linked source files that are <Compile Include>-d into more than one project (for example a sample file linked into a test project) cannot reference another assembly's <AssemblyToken>EventIds class — give their log class literal EventId integers in a high, dedicated range instead.
  • Duplicate generator on netstandard. A project that also references an R9 package (Microsoft.Extensions.Http.Resilience, .Compliance, .Telemetry, …) gets the Microsoft.Gen.Logging generator in addition to the in-box one; on netstandard both implement every partial method (CS0757). The repo's Directory.Build.targets removes the R9 analyzer on netstandard only — no per-project action is needed.

Worked example

// EventIds.cs (project root) — the assembly-token prefix avoids CS0436 across
// InternalsVisibleTo boundaries.
namespace Opc.Ua
{
    internal static class TypesEventIds
    {
        public const int Encoding = 20;   // shared codec block (reserves 20)
        public const int Matrix = 50;     // per-file block (reserves 10)
    }
}

// end of Matrix.cs
internal static partial class MatrixLog
{
    [LoggerMessage(EventId = TypesEventIds.Matrix + 0, Level = LogLevel.Debug,
        Message = "ReadArray read dimensions[{Index}] = {Dimensions}. Matrix will have 0 elements.")]
    public static partial void ReadArrayZeroDimension(this ILogger logger, int index, int[] dimensions);
}

// call site
logger.ReadArrayZeroDimension(index, dimensions);

Checklist

  • Message text and level are unchanged from the original call (behavior-preserving).
  • Placeholders are named and match parameter names; no interpolation.
  • Parameter types match the arguments; nullable only where needed; no object.
  • Expensive arguments are guarded with IsEnabled; cheap ones are not.
  • EventId uses the project's <AssemblyToken>EventIds offset (or a literal range for a shared/linked file, or a literal legacy id with an explicit EventName for a retained EventSource-compatibility message — see the narrow exception).
  • When testing with a mocked ILogger, stub IsEnabled(...) => true and match on EventId.Name, not the (empty) source-generated state ToString().

Other common tasks

  • Add a new feature — implement it in the right library, add unit and (for client/server/pubsub) integration tests, update or add a doc under docs/, and keep backward compatibility (see Coding standards).
  • Add a document — put it in docs/ and link it from docs/README.md.
  • Add a dependency — declare the version in Directory.Packages.props (Central Package Management), prefer AOT/trimmable and permissively licensed packages, and get maintainer approval first.
  • Certificates and secrets — see Certificates.md and CertificateManager.md.
  • Source-generated node managers / data types — see NodeManagers.md and SourceGeneratedDataTypes.md.
  • Server namespace metadata / history advertisement — see NodeManagers.md.
  • Dependency injection — see DependencyInjection.md.
  • NativeAOT — see NativeAoT.md.

Packages, platform support, and versioning

Released packages

The following NuGet packages are released on a monthly cadence (with hot fixes for security issues). The OPCFoundation prefix is reserved, and the assemblies and packages are signed by the OPC Foundation.

For improved source-level debugging, symbol packages for non-.Debug package IDs are published on nuget.org in snupkg format. Debug-compiled packages are retained with a .Debug suffix on GitHub Packages but are not published to nuget.org.

In-development builds from master (2.0.0-preview.<N>.g<commit>) are published only to the GitHub Packages feed. nuget.org receives a version only through a manually dispatched release.yml run from a release/<major>.<minor> branch: an approved promotion of a stable release, or a numbered preview release such as 2.0.0-preview.6. 2.0.0-preview.2 to .5 came from the earlier, retired release/2.0.0 line and are the last previews that also put .Debug packages on nuget.org; see .Debug packages are no longer published to nuget.org. To consume the nuget.org previews, use 2.0.0-preview.* to float to the latest published 2.0.0-preview.N release, pass --prerelease to dotnet add package, or select Include prerelease in Visual Studio. No additional package source or credentials are required for nuget.org; the GitHub Packages feed needs a classic PAT with read:packages.

The full set of packages nuget-publish.yml produces is pinned in .azurepipelines/expected-packages.txt. .azurepipelines/validate-source-generator-packages.ps1 fails the build when the packed output does not match it, so adding, removing or renaming a shipped package has to be done deliberately in the same pull request. That script also validates the analyzer packages: their analyzers/dotnet/roslyn<major>.<minor>/cs layout, that they carry their runtime closure privately, that the model generator's auto-imported build/<PackageId>.props is named after the package id, and — end to end — that a standalone project consuming the packed generator with a NodeSet actually gets code generated.

Supported target frameworks

The class libraries currently target:

  1. .NET Standard 2.0 (Opc.Ua.Types and Opc.Ua.SourceGeneration.Core only, because the source generators load them)
  2. .NET Framework 4.8
  3. .NET 8.0
  4. .NET 9.0
  5. .NET 10.0

The pull-request gate compiles every one of these targets, but only runs the test suite on (2) and (5) to keep the feedback loop short; the remaining test matrices are covered by scheduled or manual CI. See Running tests for how to build and test a specific framework locally with CustomTestTarget / tests/customtest.bat, and Continuous integration for how the matrices are split.

Supported analyzer and source generator hosts

The analyzer and source generator packages ship under analyzers/dotnet/roslyn<major>.<minor>/cs. The .NET SDK loads the highest folder its compiler supports and ignores folders above it, so an older host cleanly skips the analyzer instead of loading it and failing at generator-initialization time.

Roslyn API Package folder Minimum host
4.14 analyzers/dotnet/roslyn4.14/cs Visual Studio 2022 17.14 / .NET 9 SDK
5.0 analyzers/dotnet/roslyn5.0/cs Visual Studio 2026 18.0 / .NET 10 SDK

The version is declared once in roslyn.props.

Adding a band below 4.14 is not just another entry in that file. The analyzer closure — the generator, Opc.Ua.SourceGeneration.Core and Opc.Ua.Types — must bind against the Roslyn host's own System.Collections.Immutable and System.Reflection.Metadata. .NET satisfies a reference from a higher assembly version but never from a lower one, and those assemblies are supplied by the compiler, so the closure must reference the lowest version across every supported band and must never ship a copy of its own. Roslyn 4.14 and 5.0 both depend on 9.0.0, which is why $(RoslynRuntimeVersion) in roslyn.props drives the central pin and one build of the non-Roslyn closure serves both bands. Going lower — Roslyn 4.8 wants 7.x — would mean building that whole closure, Opc.Ua.Types included, a second time.

Get it wrong and the failure is silent: the generator is skipped (CS9057), fails to load (CS8032) or throws MissingMethodException while initializing (CS8784) — all warnings, so the consumer just gets no generated code. validate-source-generator-packages.ps1 therefore refuses any package that ships Microsoft.CodeAnalysis*, System.Collections.Immutable or System.Reflection.Metadata, and runs the packed down-level payload through a real compiler of that band.

Versioning

From 2.0 onward, package versions are produced by Nerdbank.GitVersioning (nbgv) from the version.json file at the repository root. That file holds the base version (currently 2.0.0-preview.{height} while master is developing the next release) and requests SemVer 2.0 package versions (nugetPackageVersion.semVer: 2); nbgv derives the version height, prerelease tag, and build metadata from the git history, and version.props maps the computed values onto the assembly and package version properties.

Stable (public-release) versions are produced only from a canonical release/<major>.<minor> branch (e.g. release/2.0, release/2.1) — never from master, a tag, or any other branch — and only at the exact commit whose version.json carries the plain <major>.<minor>.<patch> version with no prerelease label (e.g. 2.0.0). Patch numbers increase by exactly one per release on their line (2.0.0 → 2.0.1 → 2.0.2); a new minor line resets the patch to zero (2.1.0). See Release process for the full branch/version model and the step-by-step procedure for cutting a release, shipping a patch or minor version, backporting a fix, and promoting a stable candidate.

The XRegistry, WoT Connectivity, Vision, Robotics, Redundancy, Positioning, OpenUSD, ISA95, AI, and DI package families remain preview packages even when the root version is stable. Their numeric version follows the root version: for example, a stable 2.0.0 root produces 2.0.0-preview.N for these families (N a committed, manually curated number in preview-version.props that always sorts above every already-published preview — see Release process) and 2.0.0 for the other packages. The same policy applies to the Robotics and Vision MCP extensions and the OpenUSD connector tools. The package validation manifest (.azurepipelines/validate-nuget-package-set.ps1) records the root package version, a preview/stable channel, and the distinct family versions so the signed release workflow can promote an intentional mixed-version set.

The earlier 1.x packages used a different, spec-derived scheme in which the first two digits encoded the embedded NodeSet spec version (for example 1.5.378.x corresponds to OPC UA spec V1.05, mapped to release branches such as release/1.4.372). That scheme no longer applies from 2.0 onward.

Continuous integration

GitHub Actions runs all CI for master. .github/workflows/buildandtest.yml runs the complete build and test workload on GitHub-hosted runners for every triggering branch, and .github/workflows/nightly.yml runs the full-scope workload on the weekly schedule and on demand. .github/workflows/nuget-publish.yml builds, signs and publishes the packages. The other workflows in .github/workflows/ cover CodeQL, container images and the opt-in stress, stability and long-haul suites.

The only Azure Pipelines definition left on master is the manually queued, opt-in .azurepipelines/onefuzz.yml, kept because submission to the internal OneFuzz service requires the Azure DevOps-only onefuzz-task; it has no push, PR or schedule trigger. The other PowerShell helpers that remain under .azurepipelines/ are used by the GitHub Actions workflows; the directory keeps its historical name so those paths stay stable. Branches that still carry their own azure-pipelines.yml (master378, the 1.x release/* lines and the frozen release/2.0.0) keep their Azure validation; see Azure Pipelines on other branches.

What runs where

One table describes the entire migrated workload: $Profiles and $BuildProfiles in .github/scripts/get-ci-matrix.ps1. Both workflows call that script, so a profile added there appears in both without editing any YAML.

Pull request (buildandtest.yml) Full scope (nightly.yml)
Test profiles Windows net48, Windows/Linux/macOS net10.0 the above plus net9.0, net8.0, the Debug legs, and the tiers that lift the category filter
Solution builds every .slnx on Windows for net48/net10.0 × Debug/Release, UA.slnx on Linux for net8.0/net9.0/net10.0 every .slnx on Windows for all four TFMs × Debug/Release, plus the Linux legs
Native AoT linux-x64, osx-x64, osx-arm64 the above plus win-x64
Coverage project floors and the graduated patch gate project floors only
Trigger every push and pull request weekly schedule and workflow_dispatch

A profile pins its target framework through CustomTestTarget, not --framework, because that is the mechanism targets.props uses.

GitHub refuses to start a run whose matrices expand past 256 jobs, and the limit cannot be raised. The project fan-out is therefore batched: each matrix entry carries several projects that .github/scripts/run-dotnet-tests.ps1 runs in sequence, keeping per-project results and per-project timeouts. The batch size is the smallest one that fits the budget, so it grows by itself as test projects are added instead of silently truncating the matrix. CiMatrixScriptTests asserts that both scopes fit, that every *.Tests.csproj on disk is either scheduled or explicitly excluded, and that a narrowed matrix can never expand to nothing.

The executor fails a project when its build fails, when the TRX counters report any failure, when it produces no TRX at all, when nothing in it actually ran, or when it outlives its per-project ceiling. A "no tests ran" outcome is a failure, not a pass — and that includes a suite whose tests were all skipped, which records a non-zero total while verifying nothing.

That per-project ceiling is a single combined budget covering the project's build and its test run, measured by one stopwatch. The matrix derives each job's timeout-minutes from it as 20 + projectCount × perProjectTimeout, so spending it twice per project would let a batch outlive its job: GitHub would cancel the run, and a cancelled job produces neither the executor's per-project annotation nor its results. A batch whose budgets add up past the job ceiling is an error in get-ci-matrix.ps1 rather than a clamped timeout-minutes, for the same reason. CiMatrixScriptTests pins both halves of that arithmetic.

The verdict comes from the emitted TRX rather than from the dotnet test exit code, and lives in .github/scripts/get-test-verdict.ps1 so it can be tested on its own — see CiTestVerdictTests. A non-zero exit is tolerated when, and only when, the results record at least one passing test and no failure, error, timeout, abort or passedButRunAborted. That combination means the host died during process exit, after the last test and every teardown had already run; failing it would report a false red. It is not a macOS quirk — Windows hosts do it too (run 35714133848, test-windows-net48 (5/30): Opc.Ua.Client.Tests reported 256 passed, 0 failed, host exit 1). A host that dies mid-run leaves a non-zero counter and is still rejected, and a run in which every test was skipped is rejected whatever the exit code says. Every tolerated run raises a warning annotation and is labelled in the job summary, so a host that keeps dying stays visible.

Why a project can be skipped

RestrictForLegacyTfm in targets.props turns a project that does not support the requested CustomTestTarget into an empty shell with IsTestProject=false. Running dotnet test against one of those produces no TRX, which the executor treats as a failure — so it first probes dotnet msbuild -getProperty:IsTestProject and records the project as not applicable instead. Those rows appear in the job summary, so a project that quietly stops being applicable everywhere is visible rather than invisible.

Test tiers

The pull-request profiles filter out TestCategory=LongRunning and TestCategory=Stress. The tiers those filters leave out run elsewhere:

Tier Where it runs
LongRunning categories in mainline projects linux-long-running profile, nightly.yml only
Opc.Ua.Subscriptions.Durable.Tests windows-durable / linux-durable profiles, nightly.yml only
Opc.Ua.Stress.Tests .github/workflows/stress-test.yml, opt-in
Opc.Ua.Aot.Tests and the .Historian / .Mcp companions aot-test job (both workflows) — published and run as native executables, not through dotnet test
Opc.Ua.OneFuzz.Validator.Tests fuzz-drop job — it pins net10.0 to match the drop it validates

Running the full scope

nightly.yml runs weekly at Sunday 02:00 UTC. You can also start it from the Actions tab (or gh workflow run nightly.yml) with:

Input Effect
include_macos Include the macOS profiles (default true)

It uploads a workload-manifest artifact recording every (project × profile) tuple the run intended to cover, so a run's scope can be compared against another inventory instead of inferred from job names.

The fuzz crash corpus

The checked-in corpus under fuzzing/ runs on every pull request. The full crash corpus — about 22k inputs from earlier fuzzing campaigns, formerly the Azure secure file FuzzingArtifacts.zip — lives on the orphan branch fuzz-corpus and takes too long to replay per pull request. The crash-corpus job in nightly.yml checks out a commit of that branch pinned in the workflow (FUZZ_CORPUS_COMMIT), overlays it onto the matching fuzzing/ projects and runs Opc.Ua.Encoders.Fuzz.Tests on every non-macOS profile. It needs no secrets, and the nightly summary fails if it was skipped. See fuzzing/CrashCorpus.md for how to add inputs and update the pin.

Required checks and coverage

Three concerns are deliberately kept apart:

Concern Check In the branch ruleset?
Every build and test passed build-and-test summary Yes — required
Every sample container image built images summary Eligible — require it
Coverage meets the thresholds code coverage No — advisory

build-and-test summary is a single rollup job on purpose. The jobs underneath it are matrix-generated, so their names change whenever a test project or a profile is added; requiring a generated name would break as soon as the matrix changed. The job runs on always() and inspects needs.*.result itself, calling exit 1 on anything that is neither success nor skipped — a failing dependency therefore shows as a red X.

always() is not optional here: a job skipped because a dependency failed surfaces to GitHub as skipped, and a required check reporting skipped is treated as satisfied. Without always() the rollup would wave a red build straight through.

For the same reason the workflow carries no paths: filter. A workflow filtered out by paths never reports its checks at all, and a required check that never reports leaves a pull request permanently "Expected — waiting for status to be reported". The decision is applied inside the discover job instead: a docs-only pull request skips the expensive jobs and still gets a legitimate green summary.

That decision is a deny-list, not an allow-list of build inputs, and it lives in .github/scripts/get-path-relevance.ps1 so it can be tested — see CiPathRelevanceTests. It fails dangerously in one direction only: calling a real change irrelevant skips the build, test and AoT jobs while the required summary still reports success, so a broken change merges behind a green check. An allow-list of extensions cannot be kept complete — the source generators consume .xml and .csv design files (the AdditionalFiles items in src/Opc.Ua.WotCon/Opc.Ua.WotCon.csproj), test projects carry XML and JSON fixtures, and .editorconfig is enforced at build time. Only Markdown and the docs/ tree, which holds nothing but Markdown and images, are skippable; everything else builds.

.github/workflows/docker-image.yml (Images CI) follows the identical pattern for the sample container images. Its build legs are named build-and-push-image (refserver), (boilerserver) and so on, so they cannot be pinned in a ruleset either; images summary is the fixed name that rolls all of them up. One broken image fails it, because a matrix job aggregates to success only when every leg succeeded. Its pull_request trigger carries the same branch list as the CI workflow and no paths-ignore, so both gates report on exactly the same set of pull requests; the docs/tests exclusion moved into its own discover job. Add images summary to the ruleset alongside build-and-test summary.

The coverage check reports a clean failure when the thresholds are missed, so a miss is visible on the pull request, but it never blocks the merge. Do not add it to the ruleset — that would make a coverage dip unmergeable, which is not the intent.

build-and-test summary is the only required check on master.

How coverage is measured

Every test matrix entry collects coverage while it runs and publishes its raw Cobertura fragment as an artifact. The coverage job then downloads every fragment the run produced, merges them once with ReportGenerator, and evaluates the merged report. It never re-runs the tests — doing so serialises a suite that was deliberately fanned out across matrix jobs and blows the job timeout.

Coverage is collected only on .NET 8.0 and newer hosts. coverlet.collector 10.x ships build assets for net8.0+ only, so a .NET Framework test host cannot load the XPlat Code Coverage collector at all — VSTest merely warns and writes nothing. The expander therefore never requests coverage on a net4* profile, and CiMatrixScriptTests asserts that. The long-running and durable tiers also opt out: they re-run projects the filtered legs already covered, so folding their numbers in would double-count them.

The evaluation is .azurepipelines/check-coverage.ps1, driven by coverage-thresholds.json:

Check Behaviour
Project floor Total line and branch rates must meet the absolute floors in coverage-thresholds.json. The ignore globs are applied here too, so samples, tests and generated code do not count.
Patch coverage On pull requests, lines you added or modified must reach a floor that is graduated by how much changed — see below. The uncovered changed lines are listed by file.
Baseline delta Reports how total coverage compares with the recorded baselineLineRate. Warning only, even within this advisory check.

Ratchet minimumLineRate, minimumBranchRate and baselineLineRate upward as coverage improves; never lower them to turn a red check green.

The script lives under .azurepipelines/ for historical reasons only; it is not Azure-specific.

Two things about the ignore globs regularly catch people out. samples/** is ignored, so a sample can carry tests for its own sake — a wrong kinematics solver would make a sample lie — without those lines counting toward the patch gate. tools/** is not ignored, so anything you change under tools/ is measured like product code, and an assembly no test project references contributes changed lines that are counted as uncovered because no report mentions them. If you add code there, make sure a test project loads the assembly, or the patch gate will read far lower than the per-file numbers suggest.

Patch coverage is graduated by patch size

A coverage percentage over a handful of lines carries almost no information. One uncovered line in a two-line fix reads as 50%, and a flat floor would fail it — which teaches authors to ignore the check rather than act on it. So the requirement scales with how much actually changed:

Coverable changed lines Floor Below the floor
1 – 10 50 % ⚠️ warning, check still passes
11 – 100 60 % ⚠️ warning, check still passes
more than 100 patch.target − patch.threshold (75 %) ❌ failure

Only changes larger than the last band can fail the patch check. At that size the percentage is meaningful, and a large untested change is exactly what the check exists to catch. Below it you still get a warning naming the uncovered lines, so the signal is never silent — it just does not block.

The bands live in patch.bands in coverage-thresholds.json. They are consulted in order and the first band whose maxChangedLines covers the patch wins; set enforced: true on a band to make it blocking. Anything larger than the last band falls through to patch.target − patch.threshold and is always enforced.

Remember that the coverage check as a whole is advisory and stays out of the branch ruleset — an enforced band produces a red Code coverage check, not a blocked merge.

Codecov

The merged report is also uploaded to codecov.io, which is where the pull-request comment, the file-by-file diff view and the coverage trend live. Codecov does not gate. Both of its status checks are informational: true in codecov.yml, because two gates with two sets of thresholds would eventually disagree about the same pull request and the easier one to silence would win. The rules that actually decide are the ones above.

The upload carries the actions flag. It is optional and never fails a build: turn it off with the ENABLE_CODECOV workflow env (default 'true'); it is also skipped when the CODECOV_TOKEN secret is unavailable, as on fork pull requests.

Keep the ignore list in codecov.yml in step with the one in coverage-thresholds.json, or the two will report on different code.

Where the numbers appear

The script renders a markdown summary, so you never have to open a raw log to see why coverage moved. It is appended to the run's job summary and posted as a single sticky pull-request comment that is updated in place on each run. Threshold misses additionally appear as run annotations. On a pull request from a fork the token is read-only, so the comment is skipped and only the job summary is written. The merged HTML report is published as a coverage-report artifact.

A full-scope nightly.yml run reports different numbers from a pull-request run, and that is expected: it merges a different set of profiles.

To reproduce a coverage failure locally, generate the same report with tests/codecoverage.cmd (or tests/codecoverage.sh) and run the script against it:

./.azurepipelines/check-coverage.ps1 -CoberturaPath ./CodeCoverage/Cobertura.xml -BaseRef master -SummaryPath ./coverage-summary.md

Omit -BaseRef to check only the project floor, and -SummaryPath to skip the markdown summary.

Reproducing a CI leg locally

Both workflows delegate to the same two scripts, so any leg can be reproduced without a runner. Expand the matrix to see what a scope covers:

./.github/scripts/get-ci-matrix.ps1 -Scope pr -ManifestPath ./pr-manifest.json

Then run one entry's projects exactly as CI would:

./.github/scripts/run-dotnet-tests.ps1 `
    -Projects 'tests/Opc.Ua.Core.Tests/Opc.Ua.Core.Tests.csproj' `
    -CustomTestTarget net10.0 -Framework net10.0 -Configuration Release `
    -Filter 'TestCategory!=LongRunning&TestCategory!=Stress' `
    -ResultsDirectory ./TestResults

Add -Coverage to collect Cobertura fragments, and -QuietOutput to redirect child output to a log file instead of the console.

Pull requests from outside contributors

GitHub Actions applies its "Approve and run workflows" gate to outside contributors and to the GitHub Copilot coding agent: a maintainer reviews the change and then approves the workflow run.

Azure Pipelines on other branches

master no longer carries azure-pipelines.yml, azure-pipelines-preview.yml or the Azure CI templates, so a push to master (or to a branch cut from it) gives those Azure definitions nothing to run. The OneFuzz definition, which points at .azurepipelines/onefuzz.yml, is unaffected and stays manual. The Azure DevOps definitions themselves live in the Azure DevOps portal, not in this repository. Disabling their master triggers there (definition 14's push trigger and service-side schedule, definition 16's build-completion trigger, and the stale definition 13) is an administrator task and stops any residual "file not found" runs. Preserve the definitions, their artifacts, feeds, secure files, pools and service connections — master378, the 1.x release/* lines and the frozen release/2.0.0 keep their own copy of azure-pipelines.yml and still use them.

Before cutting a canonical release/2.<minor> branch: the Release ruleset requires OPCFoundation.UA-.NETStandard for every refs/heads/release/*. A new 2.x branch is cut from master and has no Azure pipeline, so that Azure context would never report and would block every pull request into the new line. Exclude the new ref from that ruleset — or add a 2.x ruleset requiring build-and-test summary — as part of creating the branch. The existing 1.x lines and the frozen release/2.0.0 keep their own copy of azure-pipelines.yml and must keep the Azure requirement.

Contributing and pull requests

  • Fork the repository (or, if you have write access, push a branch prefixed with your username) and open a pull request. You must agree to the Contributor License Agreement; the "I AGREE" prompt appears automatically on your first PR. See CONTRIBUTING.md.
  • Before submitting: all tests pass, code analysis is clean (no new warnings), the change keeps backward compatibility, and security implications are reviewed.
  • The pull-request template asks you to confirm the CLA, added tests/coverage, documentation, a warning-free build, that the UA.slnx suite passed on .NET Framework 4.8 and .NET 10.0, and that CI and CodeQL are green.
  • You can run the opc-ua-codestyle-enforcer agent to drive analyzer warnings to zero before opening the PR.

Related documentation

Fuzz replay has a dedicated GitHub Actions matrix and a local fuzzing/Scripts/test-fuzzing.ps1 entry point. It covers the modern framework matrix and applicable .NET Framework 4.8 projects, including seed-, dictionary- and script-only changes. Target inventory and fork behavior mappings belong in fuzzing/fuzz-targets.json, not hardcoded test-count floors. The dedicated fuzz-parity.runsettings includes generated protocol methods in fuzz-only coverage; it does not alter the general coverage policy. Fuzz scripts also set FuzzCoverage=true to omit generated data-type coverage-exclusion attributes only in those builds; an include filter alone cannot override a compiled exclusion attribute. Internal OneFuzz drops must remain uninstrumented locally and require owner-supplied configuration and actual worker execution evidence.