My client has a conversion program that takes ISO8859 text from round the world and converts it to UTF-8. Another one does the opposite. It is possible to get an event something like
01/02/12 01:23:45 converted "ISO-8859-8 characters" to "UTF-8 equivalent"
Where "ISO-8859-8 characters" could be Chinese or any foreign characters. How would splunk handle this mixed character set data?
And before you say try it, I don't have any example data to test yet.
I don't know, but probably you're going to get one or both strings mis-decoded. Splunk can read different character sets, but I suspect it is going to try to determine one character set for each source file.