when exporting files, all nation-specific characters
are exported incorrectly. to me it seems that the
exporter is aware only of ASCII characters and
assumes that any characters above first 127 are in
english code page (which is of course not true for
any non-english Windows).
interestingly however, the exporter also cannot
handle unicode encoded files with nation-specific
characters - ASCII characters are exported correctly,
non-ascii (nation-specific) are screwed.
it does not matter which output format you choose, it
does not work for all of them.