I made a mistake. I didn't realize that Android's FileTransfer does not do any UTF-8 encoding of the URL and relies on the user encoding it themselves (default behaviour of java.net.URL http://docs.oracle.com/javase/6/docs/api/java/net/URL.html).
Thus, I'm reverting this change: https://issues.apache.org/jira/browse/CB-907 (the test case with the Chinese characters there should fail on Android also, I think) I tried to see what Android UTF-8 encodes Chinese characters to, but: new java.net.URL("http://dummy.org/中文名.txt";).toString() ... just returns the same string. I was thinking that maybe we could do encodeURI (in JavaScript) on the URL passed to FileTransfer, but then that will just add the restriction that the URL _cannot_ be encoded already (since we don't want to double encode). I think the best thing to do is just simply add in the docs that the dev *must* encode the URL first using encodeURI. issues: https://issues.apache.org/jira/browse/CB-907 https://issues.apache.org/jira/browse/CB-1083
