fix(runtime): prune the native-alarm projector on deploy instead of clearing it (#641)

DriverHostActor cleared NativeAlarmProjector's per-condition prior state on EVERY apply, so after any deploy a driver Acknowledge of a still-active alarm projected Active=false (writing it inactive) and a Clear of an unacknowledged alarm projected Acked=true (silently acknowledging it). The projector is now pruned to the conditions the new composition still carries (RetainOnly). Claude-Session: https://claude.ai/code/session_01Ef2ozxS7QPanGT8uoATbTe
2026-10-08 17:05:42 -04:00
parent ddd7678269
commit 11f13f6326
4 changed files with 113 additions and 6 deletions
@@ -409,6 +409,66 @@ public sealed class DriverHostActorNativeAlarmTests : RuntimeActorTestBase
        publish.ExpectNoMsg(TimeSpan.FromMilliseconds(500));
    }

+    /// <summary>
+    ///     Gitea #641 — a redeploy that keeps the alarm must not reset the projector's prior state for it. It used
+    ///     to be cleared on EVERY apply, so a driver Acknowledge arriving after any deploy projected
+    ///     <c>(prev.Active = false, Acked)</c> and wrote an ACTIVE alarm inactive.
+    /// </summary>
+    [Fact]
+    public void An_acknowledge_after_a_redeploy_keeps_the_alarm_active()
+    {
+        var db = NewInMemoryDbFactory();
+        var deploymentId = SeedV3AlarmDeployment(db, RevA, Driver: "drv-1", Tag: "temp_hi");
+        var (actor, publish, coordinator) = SpawnHostAndApplyWithCoordinator(db, deploymentId);
+
+        actor.Tell(Transition(AlarmTransitionKind.Raise));
+        publish.ExpectMsg<OpcUaPublishActor.AlarmStateUpdate>(TimeSpan.FromSeconds(5)).State.Active.ShouldBeTrue();
+
+        Redeploy(db, actor, publish, coordinator);
+
+        actor.Tell(Transition(AlarmTransitionKind.Acknowledge));
+        var acked = publish.FishForMessage<OpcUaPublishActor.AlarmStateUpdate>(_ => true, TimeSpan.FromSeconds(5));
+        acked.State.Active.ShouldBeTrue("the alarm is still active at the device");
+        acked.State.Acknowledged.ShouldBeTrue();
+    }
+
+    /// <summary>Gitea #641 — likewise a Clear after a redeploy must not silently acknowledge an unacknowledged
+    /// alarm (the cleared prior state defaulted Acked to true).</summary>
+    [Fact]
+    public void A_clear_after_a_redeploy_keeps_the_alarm_unacknowledged()
+    {
+        var db = NewInMemoryDbFactory();
+        var deploymentId = SeedV3AlarmDeployment(db, RevA, Driver: "drv-1", Tag: "temp_hi");
+        var (actor, publish, coordinator) = SpawnHostAndApplyWithCoordinator(db, deploymentId);
+
+        actor.Tell(Transition(AlarmTransitionKind.Raise));
+        publish.ExpectMsg<OpcUaPublishActor.AlarmStateUpdate>(TimeSpan.FromSeconds(5)).State.Acknowledged.ShouldBeFalse();
+
+        Redeploy(db, actor, publish, coordinator);
+
+        actor.Tell(Transition(AlarmTransitionKind.Clear));
+        var cleared = publish.FishForMessage<OpcUaPublishActor.AlarmStateUpdate>(_ => true, TimeSpan.FromSeconds(5));
+        cleared.State.Active.ShouldBeFalse();
+        cleared.State.Acknowledged.ShouldBeFalse("nobody acknowledged it");
+    }
+
+    private static DriverInstanceActor.AttributeAlarmPublished Transition(AlarmTransitionKind kind) =>
+        new("drv-1", new AlarmEventArgs(
+            new StubAlarmHandle(), "Temp", AlarmRawPath, "OffNormalAlarm", kind.ToString(), AlarmSeverity.High, Ts,
+            Kind: kind));
+
+    /// <summary>A second deployment carrying the SAME alarm (a new revision of an identical artifact).</summary>
+    private void Redeploy(
+        IDbContextFactory<OtOpcUaConfigDbContext> db, IActorRef actor, TestProbe publish, TestProbe coordinator)
+    {
+        var revB = RevisionHash.Parse(new string('b', 64));
+        var second = SeedV3AlarmDeployment(db, revB, Driver: "drv-1", Tag: "temp_hi");
+        actor.Tell(new DispatchDeployment(second, revB, CorrelationId.NewId()));
+        coordinator.FishForMessage<ApplyAck>(a => a.DeploymentId == second, TimeSpan.FromSeconds(5))
+            .Outcome.ShouldBe(ApplyAckOutcome.Applied);
+        publish.FishForMessage<OpcUaPublishActor.RebuildAddressSpace>(_ => true, TimeSpan.FromSeconds(5));
+    }
+
    /// <summary>Subscribe <paramref name="probe"/> to the <c>alerts</c> DPS topic and wait for the ack.</summary>
    private void SubscribeToAlerts(TestProbe probe)
    {
@@ -439,6 +499,13 @@ public sealed class DriverHostActorNativeAlarmTests : RuntimeActorTestBase
    /// publishes an alarm. A VirtualTag-host probe is injected so the real host isn't spawned.</summary>
    private (IActorRef Actor, TestProbe Publish) SpawnHostAndApply(
        IDbContextFactory<OtOpcUaConfigDbContext> db, DeploymentId deploymentId)
+    {
+        var (actor, publish, _) = SpawnHostAndApplyWithCoordinator(db, deploymentId);
+        return (actor, publish);
+    }
+
+    private (IActorRef Actor, TestProbe Publish, TestProbe Coordinator) SpawnHostAndApplyWithCoordinator(
+        IDbContextFactory<OtOpcUaConfigDbContext> db, DeploymentId deploymentId)
    {
        var coordinator = CreateTestProbe();
        var publish = CreateTestProbe();
@@ -460,7 +527,7 @@ public sealed class DriverHostActorNativeAlarmTests : RuntimeActorTestBase
        // ExpectMsg<AlarmStateUpdate> assertions see only alarm updates.
        publish.ExpectMsg<OpcUaPublishActor.RebuildAddressSpace>(TimeSpan.FromSeconds(5));

-        return (actor, publish);
+        return (actor, publish, coordinator);
    }

    /// <summary>
@@ -24,6 +24,21 @@ public class NativeAlarmProjectorTests
        snap.Message.ShouldBe("boom");
    }

+    /// <summary>Gitea #641 — <c>RetainOnly</c> keeps a still-deployed condition's prior state and forgets a removed
+    /// one's (which then starts from the default inactive + acked).</summary>
+    [Fact]
+    public void RetainOnly_keeps_deployed_conditions_and_forgets_removed_ones()
+    {
+        var sut = new NativeAlarmProjector();
+        sut.Project("kept", Evt(AlarmTransitionKind.Raise));
+        sut.Project("removed", Evt(AlarmTransitionKind.Raise));
+
+        sut.RetainOnly(new HashSet<string>(StringComparer.Ordinal) { "kept" });
+
+        sut.Project("kept", Evt(AlarmTransitionKind.Acknowledge)).Active.ShouldBeTrue();
+        sut.Project("removed", Evt(AlarmTransitionKind.Acknowledge)).Active.ShouldBeFalse();
+    }
+
    [Fact]
    public void Retrigger_projects_active_unacked()
    {