[BUG] Decimal Division Mismatch On Some Versions Of Spark

[razajafri]

Describe the bug
When resultDecimalType was introduced in Spark 3.4.0+ and Databricks 330+, we started doing the right thing when dividing DecimalTypes while Spark still gave the wrong results according to the bug

Steps/Code to reproduce bug

scala> val df = Seq(("-0.172787979", "533704665545018957788294905796.5"), ("1", "2")).toDF("_1", "_2")
scala> df.repartition(3).selectExpr("cast(_1 as decimal(9,9)) / cast(_2 as decimal(31,1))").show(false)
+------------------------------------------------------+
|(CAST(_1 AS DECIMAL(9,9)) / CAST(_2 AS DECIMAL(31,1)))|
+------------------------------------------------------+
|-0.0000000000000000000000000000003237520              |
|NULL                                                  |
+------------------------------------------------------+

scala> spark.conf.set("spark.rapids.sql.enabled", false)

scala> df.repartition(3).selectExpr("cast(_1 as decimal(9,9)) / cast(_2 as decimal(31,1))").show(false) 
+------------------------------------------------------+
|(CAST(_1 AS DECIMAL(9,9)) / CAST(_2 AS DECIMAL(31,1)))|
+------------------------------------------------------+
|-0.0000000000000000000000000000003237521              |
|NULL                                                  |
+------------------------------------------------------+

Expected behavior
We should match Spark bug for bug for versions of Databricks 330+ and Spark 340+

Additional context
The original bug was created as part of the audit process. In resolving that as part of the spark-rapids-jni PR, it was discovered that the division got broken on Spark versions listed above after the release of Spark 330db.