CAS: drop KeyStrictness, readers skip unknown ordinary keys

[filimonov]

Remove the KeyStrictness switch: every CAS on-store reader skips an unknown ordinary key, and an unknown !-prefixed key fails with UNKNOWN_FORMAT_VERSION.

Strict readers are a leftover of the codec rewrites. Since every object is JSON with the ! rule for decision-changing fields, strictness has no job left, and it made any optional field on cas_run, fold seal, cas_ref_ckpt, cas_ref_catalog or cas_gc_maintenance_state a compatibility-version bump. The switch was also declared twice (traits table and each reader call) and the two disagreed for cas_gc_maintenance_state.

Changes

• Enum, traits column and JsonObjectReader parameter removed; seven sites that rejected unknown keys with their own throw go through skipUnknown.
• Per object: a test that an unknown ordinary key is skipped and one that an unknown ! key fails.
• storage-layout.md and Formats/README.md state the rule once.

Risks / notes

• Reader-only: nothing is written differently, no format version changes.
• Readers of earlier releases stay strict for these five objects until every node runs this build.

Changelog category (leave one):

• Improvement

Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):

CAS on-store readers skip unknown optional fields in every object type, so later releases can add fields to cas_run, fold seal, cas_ref_ckpt, cas_ref_catalog and cas_gc_maintenance_state objects without a format version bump.

Documentation entry for user-facing changes

• Documentation is written (mandatory for new features)

CI/CD Options

Exclude tests:

• Fast test
• Integration Tests
• Stateless tests
• Stateful tests
• Unit tests
• Performance tests
• Aarch64 tests
• All with ASAN
• All with TSAN
• All with MSAN
• All with UBSAN
• All with Coverage
• All Regression
• Disable CI Cache

Regression jobs to run:

• Fast suites (mostly <1h)
• Aggregate Functions (2h)
• Alter (1.5h)
• Benchmark (30m)
• CAS (content-addressed storage; Antalya only)
• ClickHouse Keeper (1h)
• Iceberg (2h)
• LDAP (1h)
• OAuth (5m)
• Parquet (1.5h)
• RBAC (1.5h)
• SSL Server (1h)
• S3 (2h)
• S3 Export (2h)
• Swarms (30m)
• Tiered Storage (2h)

🤖 Generated with Claude Code

[github-actions]

Workflow [PR], commit [a61b9a1]

[k-morozov]

After this PR, we should discuss whether we need ! at all, or whether a format version bump would be a better choice.