UNIQUE INDEX difference between 7.2 and 7.3

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Your name : Kevin Houle
Your email address : kjh(at)cert(dot)org

System Configuration
- ----------------------
Architecture : i686

Operating System : RH9, 2.4.20-19

PostgreSQL version : PostgreSQL-7.3.4 (RPMS from PGDG)

Please enter a FULL description of your problem:
- -------------------------------------------------

There is an email attachment (md5: 5cc780da645df9516235d43d1cf1e8b5)
which contains a file with two SQL INSERT commands to insert two rows
into a test table. The table should look like this:

CREATE TABLE tbl_test (
testcol text,
unique (testcol)
);

The databases in my testing are using SQL_ASCII encoding.

Executing the two attached INSERT statements produces a duplicate
key error on PostgreSQL 7.3.2 (as distributed by Red Hat) and 7.3.4
(as distributed by PGDG). Running on PostgreSQL 7.2.4 (PGDG) under
RH7.3, these two INSERTs work flawlessly. In all tests, I've executed
the INSERTs using

psql -d dbname -f filename.sql

The values in the two INSERTs are unicode filenames and they are
different. Something seems to have changed between 7.2 and 7.3 with
regard to how the values are handled, but I cannot find anything
in the docs to suggest what that change might be. So, I'm reporting
it as a bug in case it shouldn't have changed.

Regards,
Kevin

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE/OQM4u/NTC+XTbEkRAjtuAJ9DIFwrmRegc8jFkY/XKNxAjqywzACg3LnV
cosGfdzXiqcAhKJ1144Zhq0=
=9gDf
-----END PGP SIGNATURE-----