sql: add skip_rbr_unique_rowid_checks storage parameter for RBR tables

[DrewKimball]

Previously, REGIONAL BY ROW tables performed expensive cross-region uniqueness checks for all unique indexes and constraints, including those on columns with unique_rowid() or unordered_unique_rowid() defaults. These functions generate globally unique values with near-zero collision probability, making cross-region checks unnecessary overhead.

This commit adds a new storage parameter skip_rbr_unique_rowid_checks that can be set on REGIONAL BY ROW tables to skip cross-region uniqueness checks for columns using unique_rowid() or unordered_unique_rowid(). The optimization applies to:

• Columns with DEFAULT unique_rowid() or unordered_unique_rowid().
• Both INSERT and UPDATE operations.
• Primary indexes, unique secondary indexes, and UNIQUE WITHOUT INDEX.

The parameter can only be set on REGIONAL BY ROW tables and is automatically unset when changing to REGIONAL BY TABLE or GLOBAL locality. User-specified values continue to receive full uniqueness checks across regions, ensuring correctness. A notice is displayed upon setting the parameter warning not to mix user-supplied values in a column with a unique_rowid()-implied uniqueness constraint.

Epic: CRDB-39721

Release note (performance improvement): Added storage parameter skip_rbr_unique_rowid_checks for REGIONAL BY ROW tables to skip expensive cross-region uniqueness checks on columns using unique_rowid() or unordered_unique_rowid() as the default expression. This improves performance for multi-region tables while maintaining correctness guarantees.

[blathers-crl]

Your pull request contains more than 1000 changes. It is strongly encouraged to split big PRs into smaller chunks.

_{🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf.}

[cockroach-teamcity]

This change is