feat: Use Caffeine for caching compiled scripts in JSR223 samplers instead of commons-collections4 LRUMap

[vlsi]

Motivation and Context

Caffeine is a much more robust caching solution, so it makes sense to use it instead of the older LRUMap.

[owenpahl]

@vlsi, I know this change is old and is already in master for a while now.
But I was looking through it as I was curious about the change to Caffeine, and I can't see anywhere in the source where compiled scripts are added to the new cache. There only appears to be two references to the cache, one to see if a script is present

jmeter/src/core/src/main/java/org/apache/jmeter/util/JSR223TestElement.java

Line 249 in 5380e59

CompiledScript compiledScript = COMPILED_SCRIPT_CACHE.get(newCacheKey, compiler);

and emptying the cache at the end of a test

jmeter/src/core/src/main/java/org/apache/jmeter/util/JSR223TestElement.java

Line 357 in 5380e59

COMPILED_SCRIPT_CACHE.invalidateAll();

[vlsi]

get creates a value if needed: https://www.javadoc.io/doc/com.github.ben-manes.caffeine/caffeine/2.0.3/com/github/benmanes/caffeine/cache/Cache.html