__unicode__(): When converting to a unicode string, we need to
preserve spaces in the encoded/unencoded word boundaries.  RFC 2047 is
ambiguous here, but most people expect the space to be preserved.
Really closes SF bug # 640110.
1 file changed