[Spark] Use CatalogManaged commits in UniForm UC tests

[xzhseh]

Which Delta project/connector is this regarding?

• Spark
• Standalone
• Flink
• Kernel
• Other (fill in here)

Description

This PR updates UniFormE2EIcebergUCSuite so the UC-backed UniForm E2E tests use the current CatalogManaged table setup instead of injecting legacy coordinated commits coordinator table properties. The suite now enables CatalogOwnedTableFeature through the table feature property, bootstraps CatalogManaged commits with the in-memory coordinator used by existing tests, assigns the UC table id through a metadata update that simulates catalog integration, and then switches subsequent commits to the in-memory UCCommitCoordinatorClient. UCBackedDeltaCatalog now reads the table id from metadata configuration directly, which matches where the CatalogManaged path stores the catalog table id.

The shared UniForm E2E base keeps the REORG upgrade test enabled by default, but the UC-backed CatalogManaged suite opts out of that specific positive REORG path because CatalogManaged tables intentionally block REORG in production. The rest of the UC-backed UniForm E2E coverage still runs through CREATE, INSERT, UPDATE, DELETE, CREATE OR REPLACE SHALLOW CLONE, partitions, and nested schemas with the native Iceberg reader.

How was this patch tested?

Ran locally with Java 17 and Spark 4.0.1:

./build/sbt -DsparkVersion=4.0.1 "iceberg/testOnly org.apache.spark.sql.delta.uniform.UniFormE2EIcebergUCSuite"

Result: 10 tests passed.