LSP: could we use utf-8 without BOM when `-lsp --encoding utf-8` is specified?

[ntdiary]

1. when I launch lsp server with --encoding utf-8 and then send a request, server writes BOM to stdout first:
   
   because Encoding.GetEncoding("utf-8") gets utf-8 encoding with BOM
   

   omnisharp-roslyn/src/OmniSharp.Stdio.Driver/Program.cs

   Lines 27 to 29 in 258e5d0

   	var encoding = Encoding.GetEncoding(application.Encoding);
   	Console.InputEncoding = encoding;
   	Console.OutputEncoding = encoding;

2. If --encoding utf-8 is not specified and my system's default encoding is also utf-8, the server does not write a BOM to stdout.

Considering that LSP currently only supports utf-8, I'm curious if it's possible to ensure the server consistently use utf-8 without BOM across all platforms (just for lsp case). This would make client-side parsing slightly simpler. BTW, neovim has removed the logic for filtering out BOM in lsp client, so it will throw an error when receiving the message with BOM.

💖