fix(historian): address code review on Raw HistoryRead paging

C1 (critical): a boundary tie cluster larger than NumValuesPerNode could
silently truncate a resumed read to GoodNoData, permanently dropping the
un-emitted ties — the (timestamp, skip) cursor cannot advance past a single
timestamp the fixed-(start,end,cap) backend keeps re-returning. Now detected
and failed LOUDLY per node with BadHistoryOperationUnsupported + a log naming
the tag/timestamp/cap; documented in Historian.md with the larger-cap remedy.
Regression test Raw_tie_cluster_larger_than_page_fails_loudly_not_silently.

I3: build HistoryData before Save() so a projection failure can never orphan a
stored continuation cursor.

N1 (YAGNI): drop the never-produced HistoryReadKind enum + Processed-only
Aggregate/IntervalTicks fields from HistoryContinuationState — only Raw pages.

N3: ComputeResumeCursor guards its documented non-empty precondition.

I1: document InMemoryHistoryContinuationStore's eventual-consistency (test double).

Build clean, 182/182 OpcUaServer tests pass.

tests/Server/ZB.MOM.WW.OtOpcUa.OpcUaServer.Tests/HistoryPagingTests.cs

@@ -153,9 +153,9 @@ public sealed class HistoryPagingTests
     {
         var store = new InMemoryHistoryContinuationStore();
         var state = new HistoryContinuationState(
-            HistoryReadKind.Raw, "WW.Tag", new DateTime(2026, 1, 1, 0, 0, 0, DateTimeKind.Utc),
+            "WW.Tag", new DateTime(2026, 1, 1, 0, 0, 0, DateTimeKind.Utc),
             new DateTime(2026, 1, 2, 0, 0, 0, DateTimeKind.Utc), BoundarySkipCount: 2,
-            NumValuesPerNode: 100, Aggregate: default, IntervalTicks: 0);
+            NumValuesPerNode: 100);
 
         var cp = store.Save(session: null, state);
 
@@ -209,6 +209,6 @@ public sealed class HistoryPagingTests
     }
 
     private static HistoryContinuationState RawState() => new(
-        HistoryReadKind.Raw, "WW.Tag", DateTime.UtcNow, DateTime.UtcNow.AddHours(1),
-        BoundarySkipCount: 1, NumValuesPerNode: 50, Aggregate: default, IntervalTicks: 0);
+        "WW.Tag", DateTime.UtcNow, DateTime.UtcNow.AddHours(1),
+        BoundarySkipCount: 1, NumValuesPerNode: 50);
 }

tests/Server/ZB.MOM.WW.OtOpcUa.OpcUaServer.Tests/NodeManagerHistoryReadPagingTests.cs

@@ -146,6 +146,45 @@ public sealed class NodeManagerHistoryReadPagingTests : IDisposable
         await host.DisposeAsync();
     }
 
+    /// <summary>Degenerate tie cluster larger than the page cap: a single timestamp carrying MORE ties
+    /// than <c>NumValuesPerNode</c> cannot be paged past by a (timestamp, skip) cursor — the fixed-(start,
+    /// end,cap) backend keeps returning the same first <c>cap</c> ties. Rather than silently truncate to
+    /// GoodNoData (permanently dropping the un-emitted ties), the resume read FAILS LOUDLY for that node
+    /// with <c>BadHistoryOperationUnsupported</c>. (Regression for the data-loss path the carry-offset
+    /// cursor cannot resolve; the operator's remedy is a larger NumValuesPerNode.)</summary>
+    [Fact]
+    public async Task Raw_tie_cluster_larger_than_page_fails_loudly_not_silently()
+    {
+        var (host, server) = await BootAsync();
+        var nm = server.NodeManager!;
+        nm.HistoryContinuationStore = new InMemoryHistoryContinuationStore();
+
+        // 6 samples ALL sharing one timestamp (Epoch+2s) — a tie cluster of 6 with a page cap of 4.
+        var t = Epoch.AddSeconds(2);
+        var series = Enumerable.Range(0, 6)
+            .Select(i => new DataValueSnapshot((double)i, StatusCodes.Good, t, t)).ToArray();
+        nm.HistorianDataSource = new SeriesHistorianDataSource(series);
+
+        nm.EnsureVariable("eq-1/burst", parentFolderNodeId: null, displayName: "Burst", dataType: "Double",
+            writable: false, historianTagname: "WW.Burst");
+        var nodeId = nm.TryGetVariable("eq-1/burst")!.NodeId;
+
+        // Page 1: a full page of the first 4 ties, with a continuation point.
+        var (r1, e1, cp1) = ReadRaw(nm, nodeId, Epoch, Epoch.AddHours(1), max: 4, inboundCp: null);
+        e1.StatusCode.Code.ShouldBe(StatusCodes.Good);
+        r1.ShouldBe(new[] { 0.0, 1.0, 2.0, 3.0 });
+        cp1.ShouldNotBeNull();
+
+        // Page 2: the cursor cannot advance past the oversized cluster ⇒ a clear error, NOT a silent
+        // GoodNoData that would drop samples 4 and 5.
+        var (r2, e2, cp2) = ReadRaw(nm, nodeId, Epoch, Epoch.AddHours(1), max: 4, inboundCp: cp1);
+        e2.StatusCode.Code.ShouldBe(StatusCodes.BadHistoryOperationUnsupported);
+        r2.ShouldBeEmpty();
+        cp2.ShouldBeNull();
+
+        await host.DisposeAsync();
+    }
+
     /// <summary>NumValuesPerNode == 0 ("all values") never pages — the whole series returns in one shot
     /// with a null continuation point.</summary>
     [Fact]