Trivial optimisations

[PaintNinja]

Upstream PR: https://github.com/MinecraftForge/EventBus/pull/52

[alcatrazEscapee]

So, the benchmarks on the original PR are:

Name	Base (ns/op)	New (ns/op)	Change (%)
EventBusBenchmark.testClassLoaderCombined	38.549459 ±1.472707	37.653003 ± 0.907105	-2.32%
EventBusBenchmark.testModLauncherCombined	38.508817 ±1.478818	37.399871 ± 0.603555	-2.87%
EventBusBenchmarkNoLoader.testNoLoaderCombined	145.976518 ± 0.882545	149.555166 ± 2.168568	+2.45%

Notably, the first two are both within the margin of error stated and the final one has gotten slightly worse by more than the margin of error? This doesn't suggest that they are trivial optimizations and rather are trivial refactors with no concrete optimizing potential. Looking at the code, it appears to be:

• Making some more fields final
• Making an inner class static
• Typing ListenerLists.allLists as ArrayList instead of List
• Replacing a .forEach with a for(int i = 0; ...) style loop
• Replacing a for(T t : ) loop with a for(int i = 0; ...) style loop
• Replacing an ArrayList<T> with a T[]

Is there any theoretical basis for any of these to be actual optimizations? Because from what I remember from that time cpw looked at different loop structures there is no performance improvement to be expected from the idea of "let's replace all for( : ) with for(int i..)!", and so this would just be a refactor into a less clean + readable version. The practical benefit of replacing ArrayList<T> with T[] did not seem to materialize in the results. And the other changes seem like trivial code cleanup tasks, and not optimizations of any sort.

[Technici4n]

These are code style changes if anything.