[Daydream]

My problem is on Windows but I'm curious about it across all OSes, as I may expand my clients: for Kodi to properly read an NFO with diacritics (of any kind, from French characters to temperature degree sign), it MUST be encoded as UTF-8-BOM? Is that it or are there other (OS language, skin, etc) dependencies...?

This is in the context of Kodi 19, MariaDB and semi-automatically created NFOs (we're taking music videos here...) in a combo of Notepad++ and Autohotkey.

With a Unicode character but not saved as UTF-8-BOM Kodi refuses to identify the NFO file (pressing info on the video leads to "No information found for..." with the NFO file right next to it). Fixing the encoding with UTF-8-BOM makes things work, but since I'm working to automate all these things, I'd like to be sure.

[scott967]

My experience is UTF-8 (no BOM) works well.  From looking at the code, it looks like it will attempt to find an encoding declaration in the file, and also try the GUI charset encoding among other things.  But UTF-8 is preferred.  Having BOM is not required (for UTF-8) but won't stop proper reading of the file.

The only problem I've experienced is with CUE sheets (the CUE parser didn't like lines starting with the BOM)

scott s.
.