Character conversion problem

Question

I've written an app that's a wrapper for rsync and one of its functions is to get a listing from the server and see if the files exist on the local machine. My problem is with filenames that have accented characters. �, � etc
On linux it works fine and rsync -r lists the file as:
J�rgHardt.rcd
But on windows it lists like this:
J\303\266rgHardt.rcd
When it creates the files on the disk it uses the correct characters though, so if I do a File.exists() using the above string it doesn't match. So my question is, is it possible to convert the "\303\266" back into an � and if so, how?
I tried URLDecoder.decode, but that sees it as 2 characters and returns "��". I've tried the rsync.exe from cygwin and from cwRsync but they give the same output.
Thanks.

Answer 1

No, the issue is how to convert \303\266 into an �. That's all I need to know.
My point above was that since the files have the correct characters and that they are displayed properly, then obviously windows supports unicode perfectly fine as it is. The cause is that the text output from the rsync.exe isn't displaying the characters, so when I read the stream from it's output I get \303\266.
Modifying windows locale etc isn't an option and nether is modifying the rsync program. All I can do is deal with the text that I get from it.