fix(code-review): resolve Batch 2 open findings (AbCip, AbLegacy, Galaxy, FOCAS)

- Driver.AbCip.Contracts-001: parse 'writable' from TagConfig JSON (default true) instead of hardcoding
- Driver.AbCip.Contracts-002/-003: Dt type comment; drop dead [Display]/[Range] annotations
- Driver.AbCip.Contracts-004: dedicated AbCipEquipmentTagParser test class (+15)
- Driver.AbCip-017: document Tick severity Low-fallback on Bad severity read
- Driver.AbLegacy.Contracts-002/-003/-004: isArray-scalar remarks (+tests), MaxTagBytes/ForFamily docs
- Driver.Galaxy.Browser-003 + Driver.Galaxy.Contracts-003: extract ResolveApiKey -> GalaxySecretRef (dedup)
- Driver.Galaxy-019: cache buffered-interval only on Ok + ILogger warnings + ClassifyIntervalReply (+tests)
- Driver.FOCAS.Contracts-002: thread WriteIdempotent through DiscoverAsync (+test)
This commit is contained in:
Joseph Doherty
2026-06-20 22:43:36 -04:00
parent 3cc6a5f30d
commit ab57e53b92
26 changed files with 577 additions and 220 deletions
@@ -252,7 +252,7 @@ public sealed class GalaxyDriver
// listener (OTLP exporter, dotnet-trace, etc.) consumes these without the driver
// taking a dependency on the OpenTelemetry packages.
_subscriber = new TracedGalaxySubscriber(
new GatewayGalaxySubscriber(_ownedMxSession), _options.MxAccess.ClientName);
new GatewayGalaxySubscriber(_ownedMxSession, _logger), _options.MxAccess.ClientName);
_dataWriter = new TracedGalaxyDataWriter(
// Let the writer borrow live MXAccess item handles the subscription registry already
// holds, so the first write to an already-subscribed tag skips a redundant AddItem.
@@ -437,91 +437,14 @@ public sealed class GalaxyDriver
}
}
/// <summary>
/// Resolves <c>Gateway.ApiKeySecretRef</c> to the actual API-key bytes. Four
/// forms supported, evaluated in order:
/// <list type="number">
/// <item><c>env:NAME</c> — reads <c>Environment.GetEnvironmentVariable(NAME)</c>.
/// Throws when the variable is unset, so a misconfigured deployment fails
/// fast at InitializeAsync rather than silently sending an empty key.</item>
/// <item><c>file:PATH</c> — reads UTF-8 text from <c>PATH</c>, trimming
/// whitespace. Lets operators stash the key in an ACL'd file outside the
/// repo (the same pattern as the legacy <c>.local/galaxy-host-secret.txt</c>).</item>
/// <item><c>dev:KEY</c> — explicit cleartext literal. The <c>dev:</c> prefix
/// is a deliberate opt-in signal (dev box, parity rig) so the resolver
/// doesn't emit a warning; production should never use this arm.</item>
/// <item>Anything else — used as the literal API key for back-compat with
/// configs that pre-date this resolver. When a logger is supplied the
/// resolver emits a startup warning so an operator who accidentally
/// committed a cleartext key sees it (Driver.Galaxy-010).</item>
/// </list>
/// A future PR can swap any of these arms for a DPAPI-backed lookup without
/// changing the call site.
/// </summary>
/// <param name="secretRef">The secret reference string to resolve.</param>
internal static string ResolveApiKey(string secretRef) => ResolveApiKey(secretRef, logger: null);
/// <summary>
/// Logger-aware overload. Emits a <see cref="LogLevel.Warning"/> if the secret
/// ref falls through to the back-compat literal arm (an unprefixed cleartext
/// API key in <c>DriverConfig</c> JSON). The <c>dev:</c> prefix is the explicit
/// opt-in path that doesn't warn.
/// </summary>
/// <param name="secretRef">The secret reference string to resolve.</param>
/// <param name="logger">Optional logger for warning on cleartext keys.</param>
internal static string ResolveApiKey(string secretRef, ILogger? logger)
{
ArgumentException.ThrowIfNullOrEmpty(secretRef);
if (secretRef.StartsWith("env:", StringComparison.OrdinalIgnoreCase))
{
var name = secretRef[4..];
var value = Environment.GetEnvironmentVariable(name);
return !string.IsNullOrEmpty(value)
? value
: throw new InvalidOperationException(
$"Galaxy.Gateway.ApiKeySecretRef='{secretRef}' resolves to env var '{name}', but it is unset.");
}
if (secretRef.StartsWith("file:", StringComparison.OrdinalIgnoreCase))
{
var path = secretRef[5..];
if (!File.Exists(path))
{
throw new InvalidOperationException(
$"Galaxy.Gateway.ApiKeySecretRef='{secretRef}' points at '{path}', which doesn't exist.");
}
var contents = File.ReadAllText(path).Trim();
return !string.IsNullOrEmpty(contents)
? contents
: throw new InvalidOperationException(
$"Galaxy.Gateway.ApiKeySecretRef='{secretRef}' file '{path}' is empty.");
}
if (secretRef.StartsWith("dev:", StringComparison.OrdinalIgnoreCase))
{
// Explicit dev opt-in — no warning, the operator deliberately chose a
// cleartext literal (dev box, parity rig).
return secretRef[4..];
}
// Back-compat literal arm. An unprefixed string is treated as the literal
// API key — but emit a warning so an operator who accidentally committed a
// cleartext key into DriverConfig sees it at startup. Use the dev: prefix
// to suppress this warning when the literal is intentional.
logger?.LogWarning(
"Galaxy.Gateway.ApiKeySecretRef is being treated as a literal cleartext API key. " +
"Prefer env:NAME, file:PATH, or the explicit dev:KEY prefix for dev rigs — " +
"a literal key in DriverConfig JSON is stored in cleartext in the central config DB.");
return secretRef;
}
private MxGatewayClientOptions BuildClientOptions(GalaxyGatewayOptions gw) => new()
{
Endpoint = new Uri(gw.Endpoint, UriKind.Absolute),
// Driver.Galaxy-010: pass the logger so the literal-arm cleartext fallback
// surfaces a startup warning rather than silently shipping the key.
ApiKey = ResolveApiKey(gw.ApiKeySecretRef, _logger),
// surfaces a startup warning rather than silently shipping the key. The
// resolver lives in Driver.Galaxy.Contracts (GalaxySecretRef) so the runtime
// driver and the AdminUI browser share one implementation.
ApiKey = GalaxySecretRef.ResolveApiKey(gw.ApiKeySecretRef, _logger),
UseTls = gw.UseTls,
CaCertificatePath = gw.CaCertificatePath,
ConnectTimeout = TimeSpan.FromSeconds(gw.ConnectTimeoutSeconds),
@@ -1,3 +1,5 @@
using Microsoft.Extensions.Logging;
using Microsoft.Extensions.Logging.Abstractions;
using ZB.MOM.WW.MxGateway.Client;
using ZB.MOM.WW.MxGateway.Contracts.Proto;
// Use the generated nested status enum for the SetBufferedUpdateInterval reply check.
@@ -18,14 +20,21 @@ namespace ZB.MOM.WW.OtOpcUa.Driver.Galaxy.Runtime;
public sealed class GatewayGalaxySubscriber : IGalaxySubscriber
{
private readonly GalaxyMxSession _session;
private readonly ILogger _logger;
private readonly Lock _intervalLock = new();
private int _lastAppliedIntervalMs = -1; // -1 = never applied; 0 = explicit "use gw default"
/// <summary>Initializes a new instance of GatewayGalaxySubscriber.</summary>
/// <param name="session">The Galaxy MX session to use for subscription operations.</param>
public GatewayGalaxySubscriber(GalaxyMxSession session)
/// <param name="logger">
/// Optional logger; surfaces a warning when <c>SetBufferedUpdateInterval</c>
/// soft-fails so the cadence-not-applied condition isn't silent. Null is allowed
/// for unit-test construction and falls back to <see cref="NullLogger.Instance"/>.
/// </param>
public GatewayGalaxySubscriber(GalaxyMxSession session, ILogger? logger = null)
{
_session = session ?? throw new ArgumentNullException(nameof(session));
_logger = logger ?? NullLogger.Instance;
}
/// <summary>Subscribes to a bulk list of Galaxy references with optional buffered update interval.</summary>
@@ -86,11 +95,33 @@ public sealed class GatewayGalaxySubscriber : IGalaxySubscriber
},
cancellationToken).ConfigureAwait(false);
if (reply.ProtocolStatus?.Code is not (ProtocolStatusCode.Ok or ProtocolStatusCode.MxaccessFailure))
var code = reply.ProtocolStatus?.Code;
// MxaccessFailure means the COM-side SetBufferedUpdateInterval did NOT apply, so
// we must NOT cache the requested value — caching it would suppress the retry on
// the next subscribe at this interval. Only Ok records the value as applied.
if (!ClassifyIntervalReply(code))
{
// Don't throw on a soft failure — the SubscribeBulk will still succeed at the
// gw's default cadence, which is functional just not the requested cadence.
// The trace span (PR 6.1) plus the warning here gives ops the signal.
// The trace span (PR 6.1) plus this warning gives ops the signal, and leaving
// _lastAppliedIntervalMs unchanged lets the next subscribe re-attempt the set.
if (code == ProtocolStatusCode.MxaccessFailure)
{
_logger.LogWarning(
"Galaxy SetBufferedUpdateInterval({IntervalMs}ms) soft-failed (MxaccessFailure); " +
"buffered subscriptions on server handle {ServerHandle} will publish at the gateway's " +
"default cadence. The requested interval was not cached, so a later subscribe will retry it.",
intervalMs, serverHandle);
}
else
{
_logger.LogWarning(
"Galaxy SetBufferedUpdateInterval({IntervalMs}ms) returned an unexpected protocol status " +
"{Code} on server handle {ServerHandle}; treating it as not-applied and leaving the " +
"requested interval uncached so a later subscribe retries it.",
intervalMs, code, serverHandle);
}
return;
}
@@ -100,6 +131,17 @@ public sealed class GatewayGalaxySubscriber : IGalaxySubscriber
}
}
/// <summary>
/// Classifies a <c>SetBufferedUpdateInterval</c> reply: returns <c>true</c> only when
/// the requested interval was actually applied and may be cached as last-applied.
/// This is <see cref="ProtocolStatusCode.Ok"/> alone — <see cref="ProtocolStatusCode.MxaccessFailure"/>
/// means the COM-side set did NOT take effect (so caching it would prevent a retry),
/// and a <c>null</c> or any other unexpected code is treated as not-applied.
/// </summary>
/// <param name="code">The protocol status code from the gateway reply, or <c>null</c> when absent.</param>
/// <returns><c>true</c> if the interval should be recorded as applied; otherwise <c>false</c>.</returns>
internal static bool ClassifyIntervalReply(ProtocolStatusCode? code) => code == ProtocolStatusCode.Ok;
/// <summary>Unsubscribes from a bulk list of item handles.</summary>
/// <param name="itemHandles">The item handles to unsubscribe from.</param>
/// <param name="cancellationToken">The cancellation token.</param>