[SPARK-8777] [SQL] Add random data generator test utilities to Spark SQL

This commit adds a set of random data generation utilities to Spark SQL, for use in its own unit tests.

- `RandomDataGenerator.forType(DataType)` returns an `Option[() => Any]` that, if defined, contains a function for generating random values for the given DataType.  The random values use the external representations for the given DataType (for example, for DateType we return `java.sql.Date` instances instead of longs).
- `DateTypeTestUtilities` defines some convenience fields for looping over instances of data types.  For example, `numericTypes` holds `DataType` instances for all supported numeric types.  These constants will help us to raise the level of abstraction in our tests.  For example, it's now very easy to write a test which is parameterized by all common data types.

Author: Josh Rosen <joshrosen@databricks.com>

Closes #7176 from JoshRosen/sql-random-data-generators and squashes the following commits:

f71634d [Josh Rosen] Roll back ScalaCheck usage
e0d7d49 [Josh Rosen] Bump ScalaCheck version in LICENSE
89d86b1 [Josh Rosen] Bump ScalaCheck version.
0c20905 [Josh Rosen] Initial attempt at using ScalaCheck.
b55875a [Josh Rosen] Generate doubles and floats over entire possible range.
5acdd5c [Josh Rosen] Infinity and NaN are interesting.
ab76cbd [Josh Rosen] Move code to Catalyst package.
d2b4a4a [Josh Rosen] Add random data generator test utilities to Spark SQL.