added a retry for 500 internal service error

[lhr0909]

According to the Handling Error section of the MWS documentation:

The common response to a 500 or 503 service error is to try the request again. Such service errors are usually only temporary and will resolve themselves. If you want to retry an operation call after receiving a 500 or 503 error, you can immediately retry after the first error response.

We recently encountered a lot of 500 errors using the module, so I thought after I added this to my fork, I should contribute back.

Also fixed the Travis CI so it stops yelling at build errors. Also made sure the documentation in README conforms to the correct Date signature.

[lhr0909]

Also if you guys need a hand to maintain this, happy to do some of that.