Horton

Re: Allow database owners to CREATE EVENT TRIGGER

David G. Johnston <david.g.johnston@gmail.com>

View thread
From: "David G. Johnston" <david.g.johnston@gmail.com>
To: Steve Chavez <steve@supabase.io>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Aleksander Alekseev <aleksander@timescale.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2025-04-21T04:13:47Z
Lists: pgsql-hackers 
On Sunday, April 20, 2025, Steve Chavez <steve@supabase.io> wrote:

> > Also, this looks unconventional…
> > EventTriggerCacheItem *item = (EventTriggerCacheItem*) lfirst_oid(lc);
>
> Just noticed the mistake there, I would have expected a compilation error.
> New patch attached with the following change:
>
>   EventTriggerCacheItem *item = lfirst(lc);
>
> On Sun, 20 Apr 2025 at 22:55, Steve Chavez <steve@supabase.io> wrote:
>
>> Sorry, attached the output file.
>>
>>
You can remove role member_1 and trigger..1 and “create table foo” from the
nosuper script without any loss of test coverage.  Or member2 trigger2
table_bar along with the alter event trigger command which doesn’t need to
be exercised here.  Ownership is all that matters.  Whether come to
directly or via alter.

Actually, leave the other member around, but not granted ownership, and
both create tables, to demonstrate that a non-superuser and non-owner is
unaffected by the trigger.

David J.