fix(serde): reject invalid schema and partition IDs in JSON

[kamcheungting-db]

Summary

Require schema and partition IDs to be actual, in-range JSON integers.

Generic JSON conversion could previously coerce invalid values. For an ID field, this could turn:

1.5

into:

1

This PR accepts values such as:

"schema-id": 10

and rejects:

"schema-id": 1.5
"schema-id": true
"schema-id": 2147483648

The validation covers schema IDs, field IDs, identifier field IDs, nested list/map IDs, partition source/field IDs, partition spec IDs, and the table-metadata current-schema-id / default-spec-id references.

Missing partition field IDs remain supported for legacy v1 metadata; an explicitly present invalid value, including null, is rejected.

Stack

This is 4 of 4 and depends on #939:

1. Move DataFile serde into core (refactor(serde): move DataFile JSON handling from REST to core #974)
2. Java-compatible DataFile JSON (fix(serde): make DataFile JSON compatible with Java Iceberg #975)
3. Core FileScanTask JSON serde (feat(serde): support standalone FileScanTask JSON in core #939)
4. This PR: strict schema and partition ID parsing

Each PR targets main; this PR contains all preceding commits and should be reviewed last.

Testing

• Added schema and partition ID validation tests covering fractional, boolean, and out-of-range values, plus legacy v1 null and fractional table-metadata references.
• Full stack: 145 JSON serde tests passed (1 disabled).
• Full stack: 332 REST tests passed; 1 existing test skipped.

[kamcheungting-db]

This PR is now the final layer in the serde stack:

1. refactor(serde): move DataFile JSON handling from REST to core #974 — move DataFile serde into core
2. fix(serde): make DataFile JSON compatible with Java Iceberg #975 — Java-compatible DataFile JSON
3. feat(serde): support standalone FileScanTask JSON in core #939 — core FileScanTask serde
4. fix(serde): reject invalid schema and partition IDs in JSON #940 — strict schema and partition ID parsing

For the incremental diff that belongs only to this PR, use:
kamcheungting-db/iceberg-cpp@filescan-task-serde...strict-schema-spec-json-ids

In short, an ID like 10 remains valid, while 1.5, true, and values outside int32_t are rejected. Missing v1 partition field IDs remain supported.

This PR contains the earlier stack commits because all Apache PRs target main; it should be reviewed last.