Hi,
On 2022-04-06 10:37:29 +0200, Peter Eisentraut wrote:
> On 06.04.22 04:06, Andres Freund wrote:
> > On 2022-04-04 23:32:50 +0200, Peter Eisentraut wrote:
> > > This has been committed.
> >
> > It's somewhat annoying that made pg_regress even more verbose than before:
> >
> > ============== removing existing temp instance ==============
> > ============== creating temporary instance ==============
> > ============== initializing database system ==============
> > ============== starting postmaster ==============
> > running on port 51696 with PID 2203449
> > ============== creating database "regression" ==============
> > CREATE DATABASE
> > ALTER DATABASE
> > ALTER DATABASE
> > ALTER DATABASE
> > ALTER DATABASE
> > ALTER DATABASE
> > ALTER DATABASE
> > ============== running regression test queries ==============
> >
> > Unfortunately it appears that neither can CREATE DATABASE set GUCs, nor can
> > ALTER DATABASE set multiple GUCs in one statement.
> >
> > Perhaps we can just set SHOW_ALL_RESULTS off for that psql command?
>
> Do you mean the extra "ALTER DATABASE" lines? Couldn't we just turn all of
> those off? AFAICT, no one likes them.
Yea. Previously there was just CREATE DATABASE. And yes, it seems like we
should use -q in psql_start_command().
Daniel has a patch to shrink pg_regress output overall, but it came too late
for 15. It'd still be good to avoid further increasing the size till then IMO.
Greetings,
Andres Freund