Re: Schema variables - new implementation for Postgres 15 - Mailing list pgsql-hackers

From Greg Stark
Subject Re: Schema variables - new implementation for Postgres 15
Date
Msg-id CAM-w4HM3xm-pv1SFrgVc+VShAvZ8yM4c8F4Bj0dBcfOJztDEEQ@mail.gmail.com
Whole thread Raw
In response to Re: Schema variables - new implementation for Postgres 15  (Julien Rouhaud <rjuju123@gmail.com>)
Responses Re: Schema variables - new implementation for Postgres 15
List pgsql-hackers
On Sun, 26 Mar 2023 at 07:34, Julien Rouhaud <rjuju123@gmail.com> wrote:
>
> This feature can significantly increase log size, so it's disabled by default.
> For testing or development environments it's recommended to enable it if you
> use session variables.

I think it's generally not practical to have warnings for valid DML.
Effectively warnings in DML are errors since they make the syntax just
unusable. I suppose it's feasible to have it as a debugging option
that defaults to off but I'm not sure it's really useful.

I suppose it raises the question of whether session variables should
be in pg_class and be in the same namespace as tables so that
collisions are impossible. I haven't looked at the code to see if
that's feasible or reasonable. But this feels a bit like what happened
with sequences where they used to be a wholly special thing and later
we realized everything was simpler if they were just a kind of
relation.

-- 
greg



pgsql-hackers by date:

Previous
From: Andres Freund
Date:
Subject: Re: Option to not use ringbuffer in VACUUM, using it in failsafe mode
Next
From: Greg Stark
Date:
Subject: Re: Temporary tables versus wraparound... again