[ams] Force ReachableFileCleanup when snapshots exist outside main ancestry

[j1wonpark]

Why are the changes needed?

When AMS expires snapshots on a single-ref Iceberg table, Iceberg auto-selects IncrementalFileCleanup. That strategy walks the current snapshot's ancestor chain via SnapshotUtil.ancestorIds, and if a parent snapshot is missing from metadata the walk stops silently. Snapshots below the break are then treated as non-ancestors, so the ADDED entries of the manifests they wrote get reverted during cleanup. If such a manifest was superseded by a merge but its data file is still carried over (as EXISTING) in a manifest the current snapshot references, cleanup deletes a file the live table still points at, and reads fail with NotFoundException.

This is the same root cause as apache/iceberg#13568, fixed by apache/iceberg#13614 in Iceberg 1.10.0, where the strategy selector now falls back to ReachableFileCleanup when snapshots exist outside the main ancestry. Iceberg 1.6.x ~ 1.9.x lack that guard and unconditionally pick the incremental strategy for a single ref. Since bumping Iceberg core isn't always possible for existing deployments, this PR backports an equivalent guard at the Amoro layer.

Brief change log

• IcebergTableMaintainer: before committing expireSnapshots, check hasSnapshotsOutsideMainAncestry() — the same SnapshotUtil.ancestorIds walk the cleanup uses — and force ReachableFileCleanup only in that at-risk state. Healthy tables keep the automatic incremental selection, so there is no cost impact.
• ReachableFileCleanupBridge (new, in package org.apache.iceberg): RemoveSnapshots#withIncrementalCleanup(boolean) is package-private, so the bridge calls it without reflection.

How was this patch tested?

• Added a regression test (TestExpireSnapshotsKeepReferencedFiles) that reproduces the carry-over + broken-chain state and asserts referenced data files survive expiration.
• Run test locally before making a pull request

Documentation

• Does this pull request introduce a new feature? (no)