refactor(#655): delete the dead driver-side history capability

HistoryRead is served server-side by IHistorianDataSource; nothing ever dispatched
to a driver's IHistoryProvider. Removes IHistoryProvider, the OpcUaClient history
passthrough, DriverCapability.HistoryRead (+ resilience tier rows and AdminUI row),
HistoryReadStatusException and its node-manager catches, and the tests that only
covered them. Shared result types and HistoryAggregateNotSupportedException stay
(moved to Historian/HistoryReadContracts.cs). A stored ResilienceConfig still
carrying a HistoryRead key parses (diagnostic only; new test).

Claude-Session: https://claude.ai/code/session_01Ef2ozxS7QPanGT8uoATbTe

tests/Drivers/ZB.MOM.WW.OtOpcUa.Driver.OpcUaClient.IntegrationTests/OpcPlcFixture.cs

@@ -16,9 +16,8 @@ namespace ZB.MOM.WW.OtOpcUa.Driver.OpcUaClient.IntegrationTests;
 ///     <para>
 ///         <b>Why opc-plc over loopback against our own server</b> — (1) independent
 ///         cert chain + user-token handling catches interop bugs loopback can't;
-///         (2) built-in alarm ConditionType + history simulation gives
-///         <see cref="Core.Abstractions.IAlarmSource"/> +
-///         <see cref="Core.Abstractions.IHistoryProvider"/> coverage without a custom
+///         (2) built-in alarm ConditionType simulation gives
+///         <see cref="Core.Abstractions.IAlarmSource"/> coverage without a custom
 ///         driver fake; (3) pinned image tag fixes the test surface in a way our own
 ///         evolving server wouldn't. Follow-up: add <c>open62541/open62541</c> as a
 ///         second image once this lands, for fully-independent-stack interop.