> Sometimes Mojibake can be exploited for awesome. Oh lord please no! Having recently abused iconv to correct some mojibake in audio file metadata tags only t

> Sometimes Mojibake can be exploited for awesome. Oh lord please no! Having recently abused iconv to correct some mojibake in audio file metadata tags only to discover that there was no perfect solution for one set and another had lost data making them incomplete I now hate people and software that does this. You will lose the metadata telling you the encoding at some point. Then you have gibberish. What will a text system do with an invalid char? Drop it? Replace it? Error out? Copy it verbatim? Assume it's some other encoding? Keep it all as utf8 except when interacting with the Windows API.