In version 1.3.1 for Win32: The output in -s mode looks very strange, it consists of pairs <word in input encoding> <stem in dictionary encoding>. Shouldn't it be the same encoding like system default ANSI one or UTF-8?