Spark: Support view schema binding

[manuzhang]

Depends on #14984.

This draft is stacked on the Spark 4.2 support branch because spark/v4.2 is not yet present on main. The schema-binding feature is isolated in the top commit and will be rebased onto main after #14984 merges.

Changes

• Route Spark 4.2 ALTER VIEW ... WITH SCHEMA <mode> commands for Iceberg views to a dedicated execution node.
• Persist schema-binding changes through Iceberg's property update API so the operation does not replace SQL representations or create a new view version.
• Invalidate cached view plans after changing the schema mode.
• Include the schema mode in SHOW CREATE TABLE output when Spark's schema-binding feature is enabled.
• Cover EVOLUTION and COMPENSATION behavior, query-column handling, SHOW CREATE output, and named/session catalog caches.

Verification

• Focused TestViews coverage for SHOW CREATE and ALTER VIEW ... WITH SCHEMA.
• Focused TestViewCatalogCache schema-mode coverage for named and session catalogs.
• Spark 4.2 extensions Scalastyle, Spotless, and Checkstyle.
• git diff --check.

---

AI Disclosure

• Model: GPT-5 Codex
• Platform/Tool: Codex
• Human Oversight: partially reviewed
• Prompt Summary: Split Spark 4.2 Iceberg view schema-binding execution, SHOW CREATE support, and tests into a dedicated follow-up pull request.