file upload improvements

[sbwalker]

• uses Fetch API rather than XHR
• Passes file part identifiers to server using headers rather than as part of the file name (to better hide implementation details from client)
• uses asynchronous approach rather than synchronous
• uses Promises to manage the entire process and return whether or not the upload process succeeded or failed
• since the status of the entire upload process is available on the client, there is no need for polling anymore (eliminating one of the most problematic aspects of the upload process)
• Promises also prevent the server from being flooded with requests
• chunk size is configurable (however increasing its value from the default of 1 MB can result in timeouts in low latency environments)
• saves memory by not creating an array containing all of the file chunks