fix(worker): handle AckMsg add/advise/write COM failures and server death

Claude-Session: https://claude.ai/code/session_016J6isfcgqy4SB2RK3ZvFvT
2026-10-19 19:25:43 -04:00
parent 7900df3dce
commit 8199553734
2 changed files with 135 additions and 10 deletions
@@ -111,17 +111,48 @@ public sealed class AlarmAckWriter : IAlarmAckWriter
                $"Operator authentication failed for '{ackOperator.VerifyUser}' (no user id returned).");
        }

-        if (!itemHandles.TryGetValue(itemAddress, out int itemHandle))
+        ulong baselineVersion = 0;
+        int itemHandle = 0;
+        bool cached = itemHandles.TryGetValue(itemAddress, out itemHandle);
+        string step = "AddItem";
+        try
        {
-            itemHandle = mx.AddItem(serverHandle, itemAddress);
-            // MXAccess rejects writes to added-but-unadvised items with E_INVALIDARG.
-            mx.Advise(serverHandle, itemHandle);
-            itemHandles[itemAddress] = itemHandle;
+            if (!cached)
+            {
+                itemHandle = mx.AddItem(serverHandle, itemAddress);
+                itemHandles[itemAddress] = itemHandle;
+                step = "Advise";
+                // MXAccess rejects writes to added-but-unadvised items with E_INVALIDARG.
+                mx.Advise(serverHandle, itemHandle);
+            }
+
+            pumpStep();
+            step = "Write";
+            baselineVersion = completionCache.CurrentVersion(serverHandle, itemHandle);
+            mx.Write(serverHandle, itemHandle, comment ?? string.Empty, userId);
+        }
+        catch (COMException ex)
+        {
+            int hresult = ex.HResult;
+            if (itemHandles.Remove(itemAddress))
+            {
+                try { mx.UnAdvise(serverHandle, itemHandle); } catch { /* best effort */ }
+                try { mx.RemoveItem(serverHandle, itemHandle); } catch { /* best effort */ }
+            }
+
+            if (IsServerGone(hresult))
+            {
+                TearDownServer();
+            }
+
+            // Diagnostic names the step and HRESULT only: exception text can echo the item or operator.
+            return AlarmAckResult.Create(
+                AlarmAckOutcome.WriteRejected,
+                hresult,
+                $"AckMsg {step} failed (HRESULT 0x{hresult:X8}).");
        }

-        pumpStep();
-        ulong baseline = completionCache.CurrentVersion(serverHandle, itemHandle);
-        mx.Write(serverHandle, itemHandle, comment ?? string.Empty, userId);
+        ulong baseline = baselineVersion;

        if (!completionCache.TryWaitForCompletion(
                serverHandle,
@@ -158,6 +189,17 @@ public sealed class AlarmAckWriter : IAlarmAckWriter
        if (disposed) return;
        disposed = true;

+        TearDownServer();
+    }
+
+    private static bool IsServerGone(int hresult) =>
+        hresult == unchecked((int)0x80010108) // RPC_E_DISCONNECTED
+        || hresult == unchecked((int)0x800706BA) // RPC_S_SERVER_UNAVAILABLE
+        || hresult == unchecked((int)0x80010105); // RPC_E_SERVERFAULT
+
+    /// <summary>Releases the MXAccess registration so the next call re-registers (also used by Dispose).</summary>
+    private void TearDownServer()
+    {
        if (comEventSource is not null)
        {
            try { comEventSource.OnWriteComplete -= OnWriteComplete; } catch { /* best effort */ }
@@ -190,7 +232,12 @@ public sealed class AlarmAckWriter : IAlarmAckWriter
    {
        if (server is not null) return server;

-        object created = factory!.Create() ?? throw new InvalidOperationException("MXAccess COM factory returned null.");
+        if (factory is null)
+        {
+            throw new InvalidOperationException("MXAccess server is gone and no COM factory is available to re-register.");
+        }
+
+        object created = factory.Create() ?? throw new InvalidOperationException("MXAccess COM factory returned null.");
        try
        {
            comObject = created;