Suggested change for TreeOp::Put

[Dave Bryson]

Not a big deal, but one suggestion I have is to changeTreeOp::Put(&'a[u8]) to TreeOp::Put(Vec<u8>). Why?

1. The consumer Node.value is already a Vec<u8>
2. It simplifies TreeOp a bit (lifetime is not needed)
3. It's easier to use Put(Vec<u8>) from external calls - (again simplifies lifetime stuff)

[Matt Bell]

That was chosen for performance - I hate unneccessary clones. Would switching to AsRef([u8]) solve this?

[Matt Bell]

Ah but you're right that it's getting cloned into the node - maybe you're right and we can just change this 👌

[Dave Bryson]

Made the change to try it out and see the perf difference. It's in my treeop branch

Bench results (just showing Put related ones):
test bench_put_insert_random: 186,201,818 (&[u8]) --- 168,084,786 (Vec<u8>) ns/iter
test bench_put_insert_sequential: 23,737,681 (&[u8]) --- 24,472,033 (Vec<u8>) ns/iter
test bench_put_update_random: 206,077,496 (&[u8]) --- 194,155,280 (Vec<u8>) ns/iter
test bench_put_update_sequential 25,539,183 (&[u8]) --- 24,806,745 (Vec<u8>) ns/iter

Performance wise they're pretty close. I'll settle for 23,809 random inserts per/sec 😄

Minimal touchpoints to make the change - 👍 to the designer of this thing!

[Matt Bell]

Closing this because #15 now uses Vec<u8>'s where we would have used &[u8]'s. In the future there might be a few places where we can optimize by using slices, but with a AsRef<[u8]> bound so that we can still accept Vec<u8>.