basejump

v0.0.5
Built to convert 128 bit buffers to base 62 strings. Supports base 36 and buffers of arbitrary size.
base 62 base 36 128 bit flake

baseJump

Build Status Coverage Status Version npm npm Downloads Dependencies

Disclaimer

This project combines the 3 things I'm worst at:

  1. Math
  2. C++
  3. Computers

That said, I have verified the output against third party tools that will freely convert between different numeric bases. I did not include these in the tests because reasons.

Huge Thanks To Matt McCutchen

I have one dependency that is included in this code, namely the venerable and blessed Matt Mccutchen's Big Integer Library. Matt graciously hosts this source at: http://mattmccutchen.net/bigint/. If you need Maths in C++, I recommend this over using OpenSSL's BigInteger lib.

API

It's so easy even I can use it:

var jump = require( 'baseJump' );

// Node buffers can be treated as / thought of as an array of 8 bit numbers.
// Don't believe me? Call buffer.toJSON somtime. Anyway, passing one of those
// arrays to baseJump is how we pass along data.
var one = jump.toBase62( [ 100, 100, 200 ] );

// one == '000000000000000000t6Sy'
// Because of my own selfish selfishness, it auto-pads to 22 characters since that's
// the max number of characters required to represent a 128 bit unsigned integer. Maybe. Who knows.
// Certainly not me. Math is hard.
// The good news is you can control this:
var two = jump.toBase62( [ 100, 100, 200 ], 10 );

// two == '000000t6Sy' Easy.

// And base 36 calls work the same, btw, toBase36 defaults to 26 spaces.
var three = jump.toBase36( [ 100, 100, 200 ], 10 );

// three == '000007THES' ta-da

Frequently Imagined Questions (FIQ)

How Do I Get This Crap To Install on Windows

You'll need a pre-3.0 Python version installed an in your PATH. You'll also need a C++ compiler. This will be annoying to set up if you don't have one. But you're on Windows and that's just how they do.

  • VS 2010 C++ Express
  • VS 2010 SP1 - why isn't this just in the download? So that everything they do will remain a consistent hot bucket of non-sense :)

Why?

Flake ids should be 128 bits as the good Boundary intended. Or whoever came up with it. Anyway, I want flake ids and I want them today and in 128 bits and rendered as base 62 encoded, lexicograhpically sortable strings. And I want it to be as fast as possible. That is the only reason anyone should ever hurt themselves with C++.

BaseJump?

Yeah. So. Take the code and call it something else. If other people ask me, I might be able to add different base conversions. I tried to make it so it'd be easy for me. But I see a lot of folks asking for 62 and 36 and almost nothing supports either (in Node anyway) or does it sloooooowwwwwllllyyyyy.

Why Didn't You Just Use [some other thing]

I tried. Those other things were slow and/or irritating and/or required a ton of dependencies.

Metadata

  • MIT
  • Whatever
  • Alex Robson
  • released 4/17/2014

Downloads