It turns out that the position of '[Content_Types].xml' is not fixed in
this format, and whereas MS Office adds it in the beginning in general,
when files are generated using Google Docs, or RubyXL, this is not
necessarily true. They might add it at the end, so all these would be
detected as application/zip.

This fixes #39

The original images were licensed under CC-BY-SA-2.5, considered
non-free by the Debian project
https://commons.wikimedia.org/wiki/File:Phalaenopsis_%28aka%29.jpg
The new files
http://www.public-domain-image.com/free-images/flora-plants/flowers/orchid-flower-pictures/white-and-red-orchid
are licensed under the CC public domain

The freedesktop magic doesn't work for the Office 2007+ files, it
detects them all as application/zip. I have magic which detects them,
but it requires looking at 2,000 byte ranges, too much for their
project (which has to support things like file browsers). Since you
already warn that `by_magic` is slow, it'd be nice to support it here
at least.

This would allow the paperclip project to use mimemagic and stop
relying on the `file` binary:

  - https://github.com/thoughtbot/paperclip/issues/1530#issuecomment-43375497

Support is enabled through an overlay, required with

    require 'mimemagic/overlay'

Documentation is in the README