Remote: Preserve BOM in UTF-8 when found

[bpasero]

The contract basically is, depending on what encoding is picked to save a file:

• UTF-8: we do not automatically add a UTF-8 BOM unless the file has a BOM on disk
• UTF-8 with BOM: we add the UTF-8 BOM because the user explicitly asked for it
• UTF-16 LE: we always add a UTF-16 BOM because this makes it easier for us to not falsely think the file is binary due to 0-bytes within the file
• UTF-16 BE: we always add a UTF-16 BOM because this makes it easier for us to not falsely think the file is binary due to 0-bytes within the file

This logic could go into the encoding helper we already have if possible.

[bpasero]

I pushed a change for the easy cases (f50c390). There is still the remaining (more tricky) one to actually preserve the BOM when it is already there for UTF-8 only.