WIP: Regexp Unicode support

[andygrove]

Part of #5549

Changes in this PR:

• New regexp fuzz tests that generate unicode input data
• RegExp will fall back to CPU if LANG is not en_US.UTF-8
• An exception will occur if any executors has a different LANG setting to the driver

Driver error when LANG is not UTF-8

      !Expression <RegExpReplace> regexp_replace(name#4, [ae], _, 1) cannot run on GPU because regular expression support is disabled because environment variable 'LANG' is not set to 'en_US.UTF-8' and this would result in incorrect handling of Unicode data
        @Expression <AttributeReference> name#4 could run on GPU

Driver & Executor LANG mismatch

java.lang.RuntimeException: Driver and executor LANG mismatch. Driver LANG is 'en_US.UTF-8' and executor LANG is 'cockney'.
	at com.nvidia.spark.rapids.RapidsExecutorPlugin.init(Plugin.scala:218)

[andygrove]

build

[andygrove]

build

[andygrove]

build

[andygrove]

build

[andygrove]

build

[andygrove]

build

[andygrove]

Closed in favor of #5776