More accurate office file detection.

[liudonghua123]

I saw the implementation of pdf viewer chrome extension, it does not only rely on file extension in url but also consider the content-type of the response.

If you are creating a chrome extension. maybe you can use a similar way. 😄

https://github.com/mozilla/pdf.js/blob/b194c820bfbf30734589bb37f7f632bcc7790352/extensions/chromium/pdfHandler.js#L73-L94