On Sun, Feb 08, 2009 at 10:38:16AM -0500, Tom Lane wrote:
> I believe the only real "fix" is to guarantee that messages are sent
> as untranslated ASCII until we have sent an encoding indicator at
> the end of the startup sequence. Which has its own pretty clear
> downside: no more translation of authorization failures.
I, for one, would be perfectly fine with this. It has
several advantages:
- it does allow to "know" the encoding right from the beginning
- it does not require client library changes (IOW the change does
not need to trickle down: server -> libpq -> psycopg2 -> GNUmed)
- the type of (auth) failure can be deduced from the (now always
English) message string which, again, allows for client-side
translation, if so desired
Karsten
--
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346