The following bug has been logged online:
Bug reference: 1987
Logged by: Paul Lindner
Email address: lindner@inuus.com
PostgreSQL version: 8.1beta3
Operating system: Fedora Core 4 x86_64
Description: UTF8 encoding differences hamper upgrades
Details:
I've been doing some test imports of UNICODE databases into Postgres
8.1beta3. The only problem I've seen is that some data from 8.0
databases will not import.
I've generated dumps using pg_dump from 8.0 and 8.1. Attempting to
restore these results in
Invalid UNICODE byte sequence detected near byte ...
Question:
Does the 8.1 Unicode sanity code accept the full set of characters
accepted by the 8.0 Unicode sanity code?
If not we'll see a lot of problems like the one above.
I believe this patch is the one causing the problem I see:
http://www.mail-archive.com/pgsql-patches@postgresql.org/msg08198/unicode.di
ff
Is there any solution other than scrubbing the entire dataset to
conform to the new (8.1) encoding rules?
