Re: CREATE TABLE/ProcessUtility hook behavior change

On 4/30/24 21:15, jian he wrote:
> On Tue, Apr 30, 2024 at 4:35 PM David Steele <david@pgmasters.net> wrote:
>>
>> On 4/30/24 12:57, jian he wrote:
>>> On Tue, Apr 30, 2024 at 10:26 AM David Steele <david@pgmasters.net> wrote:
>>>>
>>>> Since bb766cde cannot be readily applied to older commits in master I'm
>>>> unable to continue bisecting to find the ALTER TABLE behavioral change.
>>>>
>>>> This seems to leave me with manual code inspection and there have been a
>>>> lot of changes in this area, so I'm hoping somebody will know why this
>>>> ALTER TABLE change happened before I start digging into it.
>>>
> 
> I just tested these two commits.
> https://git.postgresql.org/cgit/postgresql.git/commit/?id=3da13a6257bc08b1d402c83feb2a35450f988365
> https://git.postgresql.org/cgit/postgresql.git/commit/?id=b0e96f311985bceba79825214f8e43f65afa653a
> 
> i think it's related to the catalog not null commit.
> it will alter table and add not null constraint internally (equivalent
> to `alter table test alter id set not null`).

Yeah, looks like b0e96f31 was the culprit. Now that it has been reverted 
in 6f8bb7c1 the pgAudit output is back to expected.

To be clear, I'm not saying this (now reverted) behavior was a bug, but 
it was certainly a change in behavior so I thought it was worth pointing 
out.

Anyway, I guess we'll see how this goes in the next dev cycle.

Thanks for helping me track that down!

Regards,
-David