Deducing the content type for files

[lnicola]

If we're so inclined, we could read the first bytes of the file and use the mime_sniffer crate to and pick a content type automatically.

Alternatively, we could have an extension / MIME type mapping (mime_guess crate).

[shepmaster]

The playground guesses via filename, so I'm in favor. I had trouble trying to implement this in a reusable manner, however (although I don't remember why).

[carllerche]

I am also in favor of supporting this 👍