Re: On login trigger: take three

On Tue, Mar 15, 2022 at 4:52 PM Daniel Gustafsson <daniel(at)yesql(dot)se> wrote:
> Yeah, that was the previously posted v25 from the author (who adopted it from
> the original author). I took the liberty to quickly poke at the review
> comments you had left as well as the ones that I had found to try and progress
> the patch. 0001 should really go in it's own thread though to not hide it from
> anyone interested who isn't looking at this thread.

Some comments on 0001:

- In general, I think we should prefer to phrase options in terms of
what is done, rather than what is not done. For instance, the
corresponding GUC for row-level security is row_security={on|off}, not
ignore_row_security.

- I think it's odd that the GUC in question doesn't accept true and
false and our usual synonyms for those values. I suggest that it
should, even if we want to add more possible values later.

- "ignoreing" is mispleled. So is gux-ignore-event-trigger. "Even
triggers" -> "Event triggers".

- Perhaps the documentation for the GUC should mention that the GUC is
not relevant in single-user mode because event triggers don't fire
then anyway.

- "Disable event triggers during the session." isn't a very good
description because there is in theory nothing to prevent this from
being set in postgresql.conf.

Basically, I think 0001 is a good idea -- I'm much more nervous about
0002. I think we should get 0001 polished up and committed.

--
Robert Haas
EDB: http://www.enterprisedb.com