feat: updated version that uses ES6 crypto API

[PiiXiieeS]

Here you are a pull request that fixes the concerts of Math.random() not guaranteed to be a high-quality RNG.
Anyway, I understand your remark and the fact that this may not be a serious issue for the use cases managed by this library, like framebus that we have started to use in our projects.

Credits for https://stackoverflow.com/questions/105034/how-to-create-a-guid-uuid

[PiiXiieeS]

fix(uuid): remove arrow function
move from arrow function to standard function(c) to avoid issues with minifying with gulp task

[crookedneighbor]

This is looking good. My one concern is for browser compatibility.

Uint8Array is not supported in IE9.

Crypto is not supported in IE9 or 10.

There are parts of the Braintree web SDK that still technically support IE9, so we'll need to wait until the next major version bump of the Braintree SDK to incorporate these changes. We're working on that now, so it shouldn't be too long. Until then, I'll leave this PR open.

Thanks so much for working on this! Hoping to get it merged in within the next couple of months.

[crookedneighbor]

Also, using https://jsbench.me/, it shows it as being 97% slower to run it this way, though for Braintree's use case, I don't think that's a big deal. Running it a few times is still fast enough to not be noticeable by the end user.

@braintree/team-sdk would love to hear y'all's thoughts on this when you get a second.

[hacklschorsch]

@braintree/team-sdk would love to hear y'all's thoughts on this when you get a second.

I find it amusing that the SDK of a payment company uses Math.random() for randomness :) Hope nobody of those 340 k downloads per week on NPMJS associates the company name with serious business and uses your library to shoot a hole in their foot 🙏