Unicode support in java.io.File.listFiles()

Question

I am trying to list all the files in a given directory using the File.listFiles() method, yet for some reason the File objects returned have invalid paths when a file has unicode characters in its filename.
example .. a test directory has these files
tiga-dj_kicks.mp3
tr\374by_trio-elevator_music.mp3
if i call dir.listFiles() i will get ...
tiga-dj_kicks.mp3
tr?by_trio-elevator_music.mp3
why is java converting my unicode into "?"s?
i have seen other java apps that don't have this problem ... does anyone know why this is happening?
in terms of my system ... i am on linux 2.4.20 in the US and i have tried may jdk versions including 1.3.1 and the latest 1.4.2 as well as blackdown 1.4.1

Answer 1

It's true that System.out may not like unicode characters, so I instead I have been using a JTextArea and I know that supports my unicode.
**eek ... after a preview it seems that my unicode isn't going to show up here either :( i'll wrap brackets around what was actually displayed as unicode in my environment.
here is another test I have run ... println(String) just appends to the text area
String original = new String("A" + "\u00ea" + "\u00f1" +
               "\u00fc" + "C");
println(original);
File f = new File(original);
println(f.getPath());
f.createNewFile();
String[] files = new File(System.getProperty("user.dir")).list();
for(int i=0; i < files.length; i++)
println(files);
Output ...
[[A???C]]
[[home/ag92114/workspace/test/unicode-test/A???C]]
Test.java
Test.java~
Test.class
A???C
As you can see ... the text area correctly displayed the unicode for the String and File objects that I constructed myself, but when I list back the file I just created then my unicode is lost.
listing the directory on the shell yields ...
[ag@home:~/workspace/test/unicode-test ] ls
./ ../ A???C Test.class Test.java Test.java~
So possibly the file is not even written to the native filesystem with the unicode??
Finally I tried opening xemacs and touching a file with a unicode filename and it works fine and displays in the shell just fine, yet when I list the dir contents in java then the unicode is lost. I touched the file [[A?]] using xemacs.
shell view ...
[ag@home:~/workspace/test/unicode-test ] ls
./ A???C Test.class Test.java~
../ [[A?]] Test.java
java lists files as ...
Test.java
Test.java~
Test.class
A???C
A?
normally this is where I would concede that java just isn't capable of handling unicode in the filesystem, but I know that isn't true because I have tried other applications (jEdit, limewire) that both seem capable of listing and displaying directories and files that contain unicode. i just wish i knew how they were doing it.