All of the JIRA shows national (Slovenian) characters (čšž) correctly, except for the "Show saved filter" portlet - screenshot follows.

The encoding is set to UTF-8.

The following conversion happens:
č -> ￄﾍ
š -> ￅﾡ
ž -> ￅﾾ

Looks like something like this happens:

String s = the_corrrect_string_in;
byte b[] = s.getBytes("UTF-8"); // so far so good

// this uses the platform default encoding, which is not
// UTF-8, but rather something like ISO-8859-2
String t = new String(b); // now, t is wrong