generate random v4 UUIDs (107 bytes)

README.md

UUID

Returns a random v4 UUID of the form xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx, where each x is replaced with a random hexadecimal digit from 0 to f, and y is replaced with a random hexadecimal digit from 8 to b.

Instance, based on recursion, can be found here.

annotated.js

function(
  a,b                // placeholders
){
  for(               // loop :)
      b=a='';        // b - result , a - numeric variable
      a++<36;        // 
      b+=a*51&52  // if "a" is not 9 or 14 or 19 or 24
                  ?  //  return a random number or 4
         (
           a^15      // if "a" is not 15
              ?      // genetate a random number from 0 to 15
           8^Math.random()*
           (a^20?16:4)  // unless "a" is 20, in which case a random number from 8 to 11
              :
           4            //  otherwise 4
           ).toString(16)
                  :
         '-'            //  in other cases (if "a" is 9,14,19,24) insert "-"
      );
  return b
 }

index.js

function(a,b){for(b=a='';a++<36;b+=a*51&52?(a^15?8^Math.random()*(a^20?16:4):4).toString(16):'-');return b}

LICENSE.txt

            DO WTF YOU WANT TO PUBLIC LICENSE
                    Version 2, December 2004

 Copyright (C) 2011 Alexey Silin <[email protected]>

 Everyone is permitted to copy and distribute verbatim or modified
 copies of this license document, and changing it is allowed as long
 as the name is changed.

            DO WTF YOU WANT TO PUBLIC LICENSE
   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION

  0. You just DO WTF YOU WANT TO.

package.json

{
  "name": "UUID",

  "description": "Generates random UUIDs",
  "contributors": [
    "jed",
    "subzey",
    "tsaniel"
  ],
  "keywords": [
    "UUID",
    "ID",
    "random" 
  ]
}

@LeverOne, - operator should be faster than ^ (and it is, for about 10% in node.js). ^ uses Double ↔ Int32 conversion twice thus making significant overhead.

UPD: I was wrong, ^ is faster indeed in IE9. There's no visible difference between ^ and - cases in other browsers.

Cut one byte:

function(a,b){for(b=a='';a++<36;b+=4<<~a*6.5?(a^15?8^Math.random()*(a^20?16:4):4).toString(16):'-');return b}

@subzey, fork this gist and put your code, please!
I don't want to speak the words that you know yourself. :))

I suppose, this change is minor enough.

Unfortunately, gists can be forked but cannot be merged back. In order to avoid huge amount of hardly-ever traversable branches, let's imagine I've made a fork and then send you a pull request.

// 107
function(a,b){for(b=a='';a++< 36;b+=a * 51&52?(a^15?8^Math.random()*(a^20?16:4):4).toString(16):'-');return b}

Everything is just beginning! // <-- No, this is the end.

Great!

just incredible. @LeverOne, did you try using the negative number coercion trick instead of using hyphens?

@jed: It won't work when the number is 0.

good point, @tsaniel.

negative integers failed also.

Collisions are high in this sector.

Included into Ophal, congrats for the great work!
ophal/core@d7a4c66

Wow.
@LeverOne how you found out this a*51&52? I mean you was staring at bit patterns of 9, 14, 19 and 24 and suddenly came out with this 52? Or how?

USER TICKET #1
feature request

Even Out Character Distribution

As a user, not a developer, of this fine node module, I have to say after reading the installation instructions I have but one gripe, and it is a very specific (if minor) one.

For installing a new module (in this case typing it into my editor by reading it from my napkin that I took to my friend's or something) I do not at all mind using all of the keys on my 105 key keyboard, but I'd love to avoid using one key many times and another not at all. You know, in order to have the color on the caps wear off evenly. If I have to install this module a bunch of times I mean.

It's fine for numbers and symbols to be a bit more frequent than letters, which I use a lot for other languages, mostly natural ones. But can this be evened out a bit in general? The characters seem repetitive at first glance.

Thank you for the module though. Oh, and a mnemonic would be nice so I don't always need the napkin for installation. Might make that another ticket.

Okay, I THINK I'm on track to understanding the trick behind the a*51&52 magic now (I suspect it functions similarly to this technique to determine if a number is a power of 2 - note that, due to the bodiless design of this for, where all evaluation is done in the post-loop "increment" statement, this procedure uses one-based indices and never has to deal with the zero case); I'm going to make a tool to see if I can refine the method for discovery.

Shout out to arrow functions shaving off 5 characters.

(a,b)=>{for(b=a='';a++<36;b+=a*51&52?(a^15?8^Math.random()*(a^20?16:4):4).toString(16):'-');return b})

Awesome tricks btw.

@LiamKarlMitchell You have an extra ) at the end of your function. You can save another byte 😄