Internationalizing Basic HTTP authentication browser dialog for UserID

Question

Is it possible to have multibyte user ID for Basic HTTP authentication? Based on RFC2617 user ID has to be *Text, which basically is ASCII. But I thought maybe someone has a workaround for this limitation. Our entire web app is internationalized, we use UTF-8 as encoding for JPS pages and request processing, and that all works fine, but there is one area where we use Basic HTTP authentication, and so far I was not able to find a way to internatianalize that. Once the resource is reqested, we process request in the servlet and if the user is not authenticated we send authentication challenge response to the browser. Response encoding is set to UTF-8. After user enters the credentials, I process those in the same servlet , again using UTF-8. Of course when I tried to input the japanese ( multibyte)userID, the authentication is failing. I think the browser is corrupting DBCS data once it Base64 encodes it... Does anyone have ideas whether it is possible to internationalize this at all?

Answer 1

You'll probably need your own ServletFilter to process the authentication header, since servers will mostly decode headers in the locale encoding, regardless of any charset in the Content-type header of the request. Getting browsers to use UTF-8 encoding before base64 might be a bit tricky though.
It is probably better to use form based login. The procedure for getting UTF-8 encoded form parameters is a well understood FAQ for this forum.