When cdnify("http://s3.amazonaws.com/bucket-with-cdn.com/path+with+spaces/cover.jpg") knows about a CDN for 'bucket-with-cdn.com', it returns an invalid URL: "http://cdn/path with spaces/cover.jpg")

This is because S3Uploader.bucket_and_filename assumes you are interested in "path+with+spaces/cover.jpg" as the name of the item on S3, not the URL to the item. It calls urllib.unquote_plus on the filename. Since cdnify is interested in the item's URL, not its name, unquote_plus should not be called here.

I have a fix for this but I need to write tests for it.