On subtitle upload, Unicode detection fails if all non basic Latin characters are the same

[franontanaya]

Steps.

1. Upload a UTF-8 SRT with any number of the same non basic Latin character, e.g. á

Result: The file is treated as ISO-8859-1x and the character is corrupted.

1. Add a different non basic Latin character, e.g. é

Result: The file is treated as UTF-8 and loaded correctly.

I would look for an off-by-one error either on our code or in Django/Python functions.