BUG #7805: Inconsistency, multiple versions

The following bug has been logged on the website:

Bug reference: 7805
Logged by: Gery Deft
Email address: gerydeft(at)gmail(dot)com
PostgreSQL version: 9.2.2
Operating system: Windows Server, Windows 7,Ubuntu linux
Description:

Hi,

Versions: 9.2.*,8.4.*
Our project uses PostgreSQL as main database system, but recently we found
that PG have serious inconsistency problems.
We have about 60 connections to the database, reading and writing it most of
the time.
It does not matter if we use inline SQL, stored procedures or
pg_advisory_lock, after a few hundreds of thousand records the database will
contain duplication in a key field.
Example:
CREATE TABLE <table>
(
id serial NOT NULL,
name character varying(100),
...
CONSTRAINT domain_pkey PRIMARY KEY (id),
CONSTRAINT domain_name_key UNIQUE (name)
---------------------------------
SELECT id, name,length(name) FROM <table> where name like 'samedata%';
3161132;"samedata";8
7821530;"samedata";8
1962653;"samedata";8

I made ASCII comparison and they are identical.

This bug corrupts the workflow of our project, please fix it as soon as
possible.

Thanks,
Gery