GordonBGood

@rahulmutt, yes on letting the benchmarks decide... I had found `System.arraycopy` but it can only copy between arrays of the same type, so a cast copy would require manual conversion....

@rahulmutt, If we were to use the `sun.misc.Unsafe` class for fast byte copies, isn't there some risk that it might not be available on some platforms such as OpenJDK or...

@rahulmutt, as to primitives, the very best way would to make the "standard" Haskell/Eta primitives just do what we are doing built on top of them or in by-passing them;...

@rahulmutt, I am wondering if you can help me: I've emulated primitives and implemented arrays using standard Java Object arrays and have used your `jobjectArrayNew#`, `jobjectArraySet#`, `jobjectArrayAt#`, and `alength#` primitives...

@rahulmutt, I've done some more testing and believe I see that your "primitives" for boxed arrays just uses your `Java.Array` module so should be about the same speed as Java...

@rahulmutt, Re: the occurring SOE calling Arrays using your primitives that hook into the `Java.Array` module, I'm starting to think that it may be a manifestation of the same problem...

@rahulmutt, as to the strictness problem I had before, I can't repeat it in a small few line example and it only occurs when I use it as a benchmark...

@rahulmutt, on the subject of your implementation of pinned `addr#` using a wrapped version of `ByteBuffer`, why do you need to emulate a pinned `ByteArray` or memory outside of the...

@rahulmutt, I know how projects like this go, and how you BDFL(s) are so buried into it that sometimes things get implemented just because they happen; it's what I bring...

@rahulmutt, I had never used cabal much, just the ghc compiler, so didn't recognize it in Etlas. I see in the Eta repo that you have been able to reuse...