Greg Stark <greg.stark@enterprisedb.com> writes:
> So it would still be possible to byass this check by cloning a
> database into SQL_ASCII and then cloning it into the desired encoding?
> Doesn't sound like it really accomplishes much.
Well, it accomplishes preventing stupid encoding violations. The point
came to mind when I saw a post a little bit ago on -general in which
the poster seemed to imagine that CREATE DATABASE would convert
encodings for him. Since that is not happening in the foreseeable
future, I think we need to prevent the system from acting like it
would work.
If we wanted to be entirely anal about this, we could allow SQL_ASCII
destination with a different source encoding, but not the reverse.
However, we currently consider that you're on your own to ensure sanity
when using SQL_ASCII as far as locale goes, so I'm not sure why the
policy would be different for encoding.
regards, tom lane