Re: change in behaviour? Is this a regression in function?

Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Gavin Atkinson wrote:
>> Was this an intentional change or will it be fixed in 7.4?

> It was an accident and will be fixed in 7.4.

No, it was deliberate --- the difficulty of preserving the old behavior
seemed too high, and there were no complaints about it at the time.
http://archives.postgresql.org/pgsql-hackers/2002-09/msg00031.php

We've gotten a very small number of complaints subsequently, but AFAICT
most people either haven't noticed or like the space savings.

There has been some subsequent discussion of adding a WITH/WITHOUT OIDS
option to CREATE TABLE AS, eg
http://archives.postgresql.org/pgsql-hackers/2002-11/msg00336.php
http://archives.postgresql.org/pgsql-hackers/2002-11/msg00357.php
but AFAICT the amount of effort needed is well out of proportion to the
value of the feature.  You can always do a plain CREATE TABLE and then
fill the table with INSERT/SELECT, if you need to have OIDs.

            regards, tom lane