Re: Question for Postgres 8.3

Clemens Schwaighofer <cs(at)tequila(dot)co(dot)jp> writes:
> * Disallow database encodings that are inconsistent with the server's
> locale setting (Tom)

> does this mean, if my server LOCALE is for example UTF-8.en_US, and I want
> to create a EUC_JP database it gets rejected? do I missunderstand that?

Nope, you have it correctly.

> Normaly my servers have default locale set to UTF-8.en_US but also have the
> locales for UTF-8.ja_JP and EUC_JP there, 99.9% of my databases are utf-8,
> but I have some clients that created EUC_JP databases, will the upgrade
> affect this?

I'm surprised your clients haven't been screaming about bogus sorting
and upper/lowercasing behavior.

If you want to support multiple encodings, the only safe locale choice
is (and always has been) C. If you doubt this, troll the archives for
awhile --- for example, searching for locale+encoding in pgsql-bugs
should provide plenty of amusing reading matter. 8.3 is just refusing
to do things that are known to be unsafe in previous releases.

regards, tom lane