SharedMimeInfo::guess_mime_type already combines guesses from the filename, metadata, and content, so there is no need to do a filename based guess separately. Also, the result of guessing from the filename is always preferred, defeating the purpose of doing a content-based guess in the first place.
Fix #32.
SharedMimeInfo::guess_mime_typealready combines guesses from the filename, metadata, and content, so there is no need to do a filename based guess separately. Also, the result of guessing from the filename is always preferred, defeating the purpose of doing a content-based guess in the first place.