I've written a Java program to generate an m3u file based on a CD ripped from k3b which pretty much preserves special character encodings in artist, album and track names. I then place these m3u files on a server and generate a GWT web application where the m3u file name is the target of an HTML anchor tag. For 99+% of cases, this all works perfectly. For a few cases, special characters cause the link to fail.
One failing example is the Movits! album Äppelknyckarjazz (note the first character which gets encoded by a URI constructor as %C3%84). Since the client is GWT, view source does not show the link, :-( But when hovering over the link Firefox shows the correctly decoded URL. When clicking on the link, Firefox fails with: "...Äppelknyckarjazz.m3u was not found on this server" It is as though different character encoding schemes are at play but frankly my brain is hurting in trying to unravel the puzzle at this level.
So there are really two questions:
1) Is my problem an encoding scheme issue?
2) Assuming it is, how can I maintain consistency given the various pieces of the application (Java m3u generater, GWT client, Firefox browser, Apache web server).