Re: [HACKERS] Invalid unicode in COPY problem

"John Hansen" <john@geeknet.com.au> writes:
> Tatsuo Ishii wrote:
>> We have developed patches which relaxes the character
>> validation so that PostgreSQL accepts invalid characters.

> That is just plain 100% wrong!!

That was my first reaction too.  Why would this be a good idea?
If someone does want an encoding-agnostic database, they can
set it as SQL_ASCII.

            regards, tom lane