Re: New Object Access Type hooks

Andrew Dunstan <andrew@dunslane.net>

From: Andrew Dunstan <andrew@dunslane.net>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Mark Dilger <mark.dilger@enterprisedb.com>, Julien Rouhaud <rjuju123@gmail.com>, Joshua Brindle <joshua.brindle@crunchydata.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>, Robert Haas <robertmhaas@gmail.com>, Jeff Davis <pgsql@j-davis.com>, Joe Conway <joe@crunchydata.com>, Stephen Frost <sfrost@snowman.net>
Date: 2022-03-22T16:09:20Z
Lists: pgsql-hackers
On 3/22/22 12:02, Tom Lane wrote:
> Andrew Dunstan <andrew@dunslane.net> writes:
>> That seems quite weird. I'm not sure how it's getting loaded at all if
>> not via shared_preload_libraries
> Some other animals are showing this:
>
> diff -U3 /home/postgres/pgsql/src/test/modules/test_oat_hooks/expected/test_oat_hooks.out /home/postgres/pgsql/src/test/modules/test_oat_hooks/results/test_oat_hooks.out
> --- /home/postgres/pgsql/src/test/modules/test_oat_hooks/expected/test_oat_hooks.out	2022-03-22 11:57:40.224991011 -0400
> +++ /home/postgres/pgsql/src/test/modules/test_oat_hooks/results/test_oat_hooks.out	2022-03-22 11:59:59.998983366 -0400
> @@ -48,6 +48,8 @@
>  SELECT * FROM regress_test_table;
>  NOTICE:  in executor check perms: superuser attempting execute
>  NOTICE:  in executor check perms: superuser finished execute
> +NOTICE:  in executor check perms: superuser attempting execute
> +NOTICE:  in executor check perms: superuser finished execute
>   t 
>  ---
>  (0 rows)
> @@ -95,6 +97,8 @@
>                        ^
>  NOTICE:  in executor check perms: non-superuser attempting execute
>  NOTICE:  in executor check perms: non-superuser finished execute
> +NOTICE:  in executor check perms: non-superuser attempting execute
> +NOTICE:  in executor check perms: non-superuser finished execute
>   t 
>  ---
>  (0 rows)
> @@ -168,6 +172,8 @@
>                        ^
>  NOTICE:  in executor check perms: superuser attempting execute
>  NOTICE:  in executor check perms: superuser finished execute
> +NOTICE:  in executor check perms: superuser attempting execute
> +NOTICE:  in executor check perms: superuser finished execute
>   t 
>  ---
>  (0 rows)
>
>
> I can duplicate that by adding "force_parallel_mode = regress"
> to test_oat_hooks.conf, so a fair bet is that the duplication
> comes from executing the same hook in both leader and worker.
>
> 			



OK, thanks. My test didn't include that one setting :-(


If I can't com up with a very quick fix I'll revert it.


cheers


andrew


--
Andrew Dunstan
EDB: https://www.enterprisedb.com




Commits

  1. Update some tests in 013_crash_restart.pl.

  2. Avoid freeing objects during json aggregate finalization

  3. SQL/JSON constructors

  4. Force NO_LOCALE / UTF8 for test_oat_hooks tests

  5. Temporarily disable installcheck for test_oat_hooks module

  6. Fix new Object Access hooks test

  7. Add String object access hooks

  8. Add a test module for Object Access hooks