Made script AMD/CommonJS compatible

[Rendez]

In our case, we had cases where we wanted to run the script more than once. This is why I've also taken the change to wrap the module into a UMD factory-style closure. It's the cleanest way to allow exposure of the method, in order to be called at convenience.

[jonathantneal]

@Rendez, excellent work. How might this handle the xhr.onload issue that @rjmunro was hoping to resolve? Also, could this wrapper be added as part of a build process?

I’m looking into https://github.com/bebraw/grunt-umd

[Rendez]

Yeah, it's a great idea. I am very happy about your feature/2.0.0 branch.

ps: What do you mean about the xhr.onload issue? why does it need to pass this context with bind()?

[jonathantneal]

The issue seemed to do with IE11 throwing an error Unable to get property 'splice' of undefined or null reference.

See: whiteoctober@bf3184a#diff-857c32776517cbdd3f9f587bd3561c73L64

I don’t think there was a isolated example provided though.

[Rendez]

If that's true, the XHR object returned to the onload callback is probably a copy of the original XHR object. Without debugging I would assume this is a bug in IE11, and probably IE10 too. Is there a reason to not try onreadystatechange? Perhaps that works out better.

[jonathantneal]

The XHR issue has been resolved in 2.0.0 in #62. Please review.