Re: type cache cleanup improvements

On Fri, Oct 25, 2024 at 11:35 AM Andres Freund <andres(at)anarazel(dot)de> wrote:
> On 2024-10-22 20:33:24 +0300, Alexander Korotkov wrote:
> > Thank you, Pavel! 0001 revised according to your suggestion.
>
> Starting with this commit CI fails.
>
> https://cirrus-ci.com/task/6668851469877248
> https://api.cirrus-ci.com/v1/artifact/task/6668851469877248/testrun/build/testrun/regress-running/regress/regression.diffs
>
> diff -U3 /tmp/cirrus-ci-build/src/test/regress/expected/inherit.out /tmp/cirrus-ci-build/build/testrun/regress-running/regress/results/inherit.out
> --- /tmp/cirrus-ci-build/src/test/regress/expected/inherit.out 2024-10-24 11:38:43.829712000 +0000
> +++ /tmp/cirrus-ci-build/build/testrun/regress-running/regress/results/inherit.out 2024-10-24 11:44:57.154238000 +0000
> @@ -1338,14 +1338,9 @@
> ERROR: cannot drop inherited constraint "f1_pos" of relation "p1_c1"
> alter table p1 drop constraint f1_pos;
> \d p1_c1
> - Table "public.p1_c1"
> - Column | Type | Collation | Nullable | Default
> ---------+---------+-----------+----------+---------
> - f1 | integer | | |
> -Check constraints:
> - "f1_pos" CHECK (f1 > 0)
> -Inherits: p1
> -
> +ERROR: error triggered for injection point typecache-before-rel-type-cache-insert
> +LINE 4: ORDER BY 1;
> + ^
> drop table p1 cascade;
> NOTICE: drop cascades to table p1_c1
> create table p1(f1 int constraint f1_pos CHECK (f1 > 0));

Thank you for reporting this.
Looks weird that injection point, which isn't used in these tests, got
triggered here.
I'm looking into this.

------
Regards,
Alexander Korotkov
Supabase