The way I'm using the base58 required me to do a lot conversions between
[]byte slices and strings. Not anymore.
Also switching to []byte slices allowed some nice performance
improvements.
Before, with strings:
BenchmarkEncode-4 3000000 511 ns/op
BenchmarkDecode-4 5000000 300 ns/op
After, with []byte slices:
BenchmarkEncode-4 5000000 256 ns/op
BenchmarkDecode-4 20000000 107 ns/op
This package serves a single purpose, convert between integers and
base58 strings the same as a Flickr does. It has no need for supporting
custom alphabets passed in a runtime.
