Tumblr ripper did not account for potential none image urls in the
original_size field. This is caused by larger images being sent to AWS
If there is no extension i now get the response’s url and download the
file from AWS
You have to enter the URL in a weird format, though.
For example, to rip `fiddle.se`, you need to try to rip:
`http://fiddle.se.tumblr.com`
Tags and /post/ rips are still supported
