diff --git a/tests/Server/ZB.MOM.WW.OtOpcUa.AdminUI.Tests/Uns/RawTagConfigGateTests.cs b/tests/Server/ZB.MOM.WW.OtOpcUa.AdminUI.Tests/Uns/RawTagConfigGateTests.cs
index cf8ea634..5305df94 100644
--- a/tests/Server/ZB.MOM.WW.OtOpcUa.AdminUI.Tests/Uns/RawTagConfigGateTests.cs
+++ b/tests/Server/ZB.MOM.WW.OtOpcUa.AdminUI.Tests/Uns/RawTagConfigGateTests.cs
@@ -318,9 +318,18 @@ public sealed class RawTagConfigGateTests
}
///
- /// Every blob emits for a browsable driver must
- /// survive the gate the commit now passes through — otherwise browse-commit would author rows the
- /// import refuses.
+ /// Browse-commit goes through , so every blob
+ /// emits must survive the gate — otherwise the
+ /// commit would author rows the import refuses.
+ ///
+ /// Sql is deliberately absent, and that is a FINDING, not an oversight. Its branch
+ /// emits {"model":"WideRow","table":…,"columnName":…} with no rowSelector, which
+ /// SqlTagConfigModel.Validate refuses and SqlEquipmentTagParser refuses too — so a
+ /// browse-committed Sql tag has never been readable, it merely failed later (at the deploy gate,
+ /// or silently at the driver) instead of at commit. Pinning today's behaviour either way would
+ /// encode a defect; it is reported for its own fix (the browse session should state a row
+ /// selector, or Sql should commit through the typed editor).
+ ///
///
/// The browsable driver type.
[Theory]
@@ -329,6 +338,8 @@ public sealed class RawTagConfigGateTests
[InlineData(DriverTypeNames.AbLegacy)]
[InlineData(DriverTypeNames.TwinCAT)]
[InlineData(DriverTypeNames.OpcUaClient)]
+ [InlineData(DriverTypeNames.MTConnect)]
+ [InlineData(DriverTypeNames.Mqtt)]
[InlineData(DriverTypeNames.Galaxy)]
public void A_browse_committed_config_passes_the_gate(string driverType)
{