Kris Jurka wrote:
> For the record, I'm in favor of changing our use of initial setup
> encoding from US-ASCII to UTF-8. While it doesn't solve the root of
> the problem, it does allow people to use non-ascii user and database
> names if they set them up appropriately and doesn't seem to harm
> anything.
That does seem like the simplest approach and should cover all possible
usernames (if you can't represent it in UTF-8 you can't get it into a
String, anyway), if not all possible encodings.
> The original patch's suggested use of the client's
> environment encoding seems random to me.
Agreed. I don't like the idea of "configure the server to use UTF-8 like
this.. except if you have one of these encodings as your JVM default in
which case you need to do something else like this.. and maybe how your
cluster needs to be configured will change if you change JVM versions or
environment or driver version".
-O