[gaim-migrate @ 13840] This should be more robust when dealing with different versions of iconv - we parse out the BOM and explicitly tell iconv what to do. This was lifted from the gtkhtml source. Ethan was concerned that some iconv implementations might be confused with the naming of the explicit charsets (UCS-2LE and UCS-2BE), so we should keep that in mind if people are having problems. This fixes the problem I was having that was caused by the BOM being removed by iconv during the UCS-2 to UTF-8 conversion. There is also some distracting whitespace fixing here to obscure any mistakes that I might have made. committer: Tailor Script <tailor@pidgin.im>