feat: Implement TimeType support - Infrastructure - sort, min/max

[YutaLin]

Which issue does this PR close?
Closes #4379

Rationale for this change
Support sort, min/max for TimeType, include native and JVM columnar

What changes are included in this PR?
support sort and add sql tests.

How are these changes tested?
run sql test in spark4.1, spark4.2

[YutaLin]

Hi @parthchandra, can you help me review this? Thanks!

[parthchandra]

Do we need to handle Literal Time64(Nanosecond) in the planner (create_expr)?
Maybe try a query with ORDER BY time_col LIMIT 10

[YutaLin]

Do we need to handle Literal Time64(Nanosecond) in the planner (create_expr)?
Maybe try a query with ORDER BY time_col LIMIT 10

Hi @parthchandra, thanks for review! CometLiteral already serializes a non-null TimeType literal as LongVal, but create_expr only handled LongVal for Int64 and timestamps. I added the missing conversion to ScalarValue::Time64Nanosecond(Some(value)) and a focused Rust unit test.
I also added more sql tests, please help me review again, appreicate!

[parthchandra]

@YutaLin could you rebase? I have one optional comment, otherwise this is looking good.

[parthchandra]

Optional: a test like -

query
  SELECT id, t
  FROM (
      SELECT id, make_time(hours, minutes, secs) AS t
      FROM test_make_time_sort
  )
  ORDER BY t ASC NULLS FIRST
  LIMIT 3

would be an additional check for supportedScalarSortElementType but reached via a different path.