Use TextDecoder for Buffer.toString?

[cmdcolin]

I have some pretty large buffers (~45MB) I'm decoding, and they can create a large "GC pressure"

With TextDecoder start: 42ms

With buffer.toString() start: 838.85302734375ms

I know there have been discussions about TextEncoder/TextDecoder in the past, but thought I'd mention the performance boost here

[gsvarovsky]

Linking to #60 for reference

[feross]

I'm open to merging a PR that switches to TextDecoder. If you would like to see it implemented, please send a pull request!