### What problem does this PR solve?

Issue Number: close #xxx

Related PR: #xxx

Problem Summary: Add comments to the file scanner v2 JSON reader interfaces and non-obvious implementation paths, including synthetic schema handling, requested column mapping, simdjson buffer lifetime, json_root/jsonpaths behavior, duplicate key handling, and malformed-row rollback.

### Release note

None

### Check List (For Author)

- Test: No need to test (comment-only change)
    - Ran git diff --check.
    - Ran build-support/check-format.sh.
- Behavior changed: No
- Does this need documentation: No

### What problem does this PR solve?

Issue Number: close #xxx

Related PR: #xxx

Problem Summary: The JsonReaderTest helper always set a valid null indicator for constructed slot descriptors, so SlotDescriptor treated even non-nullable DataTypeString slots as nullable. This made the missing-required-column test expect an error while the test input actually described a nullable slot. Fix the helper to set nullIndicatorBit to -1 for non-nullable types.

### Release note

None

### Check List (For Author)

- Test: Unit Test / Manual test
    - Ran git diff --check.
    - Ran build-support/check-format.sh.
    - Attempted ./run-be-ut.sh --run --filter='JsonReaderTest.ReturnsErrorForMissingRequiredColumn', but sandbox execution failed because nproc is unavailable, .git/modules submodule config writes are denied, and GitHub dependency download DNS is blocked. Retried with escalated permissions twice, but approval review timed out before execution.
- Behavior changed: No
- Does this need documentation: No

### What problem does this PR solve?

Issue Number: close #xxx

Related PR: #xxx

Problem Summary: File scanner v2 JSON reader had three regressions. First, openx_json_ignore_malformed appended an all-null row for malformed records, while Hive/OpenX semantics skip malformed records. Second, empty JSON lines were still passed to simdjson and failed with EMPTY. Third, the reader assumed nullable output columns always had nullable source serdes and called get_nested_serdes on scalar serdes, which broke CDC TVF JSON rows whose output file column is nullable but source slot serde is not. This change skips malformed rows, treats empty JSON lines as empty rows, and only unwraps serdes when the source type is actually nullable.

### Release note

None

### Check List (For Author)

- Test: Unit Test / Manual test
    - Added JsonReaderTest coverage for present required columns, ignored malformed rows, and empty JSON lines.
    - Ran git diff --check.
    - Ran build-support/check-format.sh.
    - Attempted ./run-be-ut.sh --run --filter='JsonReaderTest.*', but sandbox execution failed because nproc is unavailable, .git/modules submodule config writes are denied, and GitHub dependency download DNS is blocked. Retried with escalated permissions twice, but approval review timed out before execution.
- Behavior changed: No
- Does this need documentation: No

### What problem does this PR solve?

Issue Number: close #xxx

Related PR: #xxx

Problem Summary: The file scanner v2 JSON reader now skips malformed rows when openx_json_ignore_malformed is enabled. The Hive openx JSON regression expected the old behavior that materialized malformed rows as all NULL values, so the expected output failed against the corrected result. This updates the expected q1 output to contain only the valid JSON rows.

### Release note

None

### Check List (For Author)

- Test: Manual test
    - Ran git diff --check.
    - Did not run the external Hive regression locally because the external Hive test environment is not available in this workspace.
- Behavior changed: No
- Does this need documentation: No

### What problem does this PR solve?

Issue Number: close #xxx

Related PR: #xxx

Problem Summary: File scanner v2 JSON reader incorrectly skipped malformed JSON documents when openx_json_ignore_malformed was enabled. The existing OpenX JSON reader semantics materialize one all-NULL row for each ignored malformed document when all projected columns are nullable. This change restores that compatibility by rolling back any partial writes and appending a NULL row for malformed documents, and updates the JsonReader unit test and Hive regression expected output accordingly.

### Release note

None

### Check List (For Author)

- Test: Unit Test / Manual test
    - Ran git diff --check.
    - Formatted changed BE C++ files with Homebrew clang-format 16.0.6.
    - Attempted ./run-be-ut.sh --run --filter='JsonReaderTest.*' with JDK17; sandbox execution failed because nproc is unavailable, submodule config writes are denied, and GitHub dependency download DNS is blocked. Escalated retries timed out before execution.
- Behavior changed: No
- Does this need documentation: No