Tested on your test instance: accents from command line interface are uploaded correctly. Accents from an UTF-8 encoded file works great too.
I did not test other file encoding for the NFO. But even if I test it, it should not be Windows specific. And I did not find a way to detect the encoding of a file in Python. (Other than call a lib which will brute force the detection.)
If it's still works on linux I think you can merge it.