fix(bqjdbc): fix Long to java.sql.Time coercion

[Neenu1995]

b/510853631

chore(bqjdbc): suppress deprecation warnings

• Silenced Deprecation Warnings: Added @deprecated and @SuppressWarnings("deprecation") annotations to all mandated overridden methods in BigQueryBaseResultSet, BigQueryPreparedStatement, and BigQueryCallableStatement to satisfy the standard JDBC interface contract warning-free.
• Unified Arrow Coercion (LongToTime): Refactored the Arrow microsecond parser to use new Time(value / 1000) to achieve a zero-overhead, warning-free conversion that fully preserves sub-second millisecond precision.
• Fixed JSON Coercion Wall-Clock Shifts (FieldValueToTime): Introduced local timezone offset compensation arithmetic (millisOfDay - TimeZone.getDefault().getOffset()) to ensure the retrieved timezone-agnostic time-of-day matches the stored database value exactly, regardless of the client machine's timezone.
• Fixed Timestamp Shifting (FieldValueToTimestamp & LongToTimestamp): Standardized both numeric timestamp parsers to use the timezone-agnostic Timestamp.from(instant) standard, preserving the exact UTC point in time and eliminating dual timezone shifts in non-UTC system timezones.
• Modernized the Test Suite: Restructured and updated the expected assertions across multiple unit test classes (ArrowFormatType, FieldValueType, BigQueryArrowArrayOfPrimitives, BigQueryArrowStruct) to enforce these timezone-robust, high-precision, and compliant results.

[gemini-code-assist]

Code Review

This pull request updates several JDBC methods in BigQueryBaseResultSet, BigQueryCallableStatement, and BigQueryPreparedStatement with @deprecated and @SuppressWarnings annotations to align with the JDBC specification. Additionally, it replaces a deprecated java.sql.Time constructor in BigQueryTypeCoercionUtility with a Calendar-based implementation. A review comment suggests using java.time.LocalTime and java.sql.Time.valueOf() instead of Calendar to avoid performance overhead during result set iteration, as Calendar is relatively expensive for high-frequency operations.