On ons, 2009-10-21 at 13:11 +0900, Itagaki Takahiro wrote:
> So, we should always ignore BOM sequence
> at the file head no matter what client encoding is used.
I think we can't do that. That byte sequence might be valid user data
in other encodings.