| From: | Ashutosh Bapat <ashutosh(dot)bapat(dot)oss(at)gmail(dot)com> |
|---|---|
| To: | Steve Chavez <steve(at)supabase(dot)io> |
| Cc: | PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: [PATCH] Report column-level error when lacking privilege |
| Date: | 2026-03-30 15:21:40 |
| Message-ID: | CAExHW5u7H4nfXsPaaigr1TtB9HRDywJb9Jpn8uLReD1c_F7sVw@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi Steve,
On Mon, Mar 30, 2026 at 6:37 AM Steve Chavez <steve(at)supabase(dot)io> wrote:
>
> Hello hackers,
>
> When a role `xx` has `grant select (name) on items to xx;`, a generic table-level error is given:
>
> select * from items;
> ERROR: permission denied for table items
>
> With this patch, we now give:
>
> select * from items;
> ERROR: permission denied for column "id" of relation "items"
>
> This only when the user has column-level privileges, if it doesn't have any the same regular table-level error is given. This makes the most sense and also keeps current tests mostly the same.
>
> * It also works for UPDATE and INSERT.
> * Clears the TODO mentioned on lines
> https://github.com/postgres/postgres/blob/45cdaf3665bedfbabb908bb84284f3db26781ad3/src/backend/executor/execMain.c#L691-L693
> * This patch is on top of the patch mentioned on https://www.postgresql.org/message-id/CAGRrpzYP%2B3zEk__KZu-a5uWySfwgRFk6eoPXKrA5AdtBTXR%3Dng%40mail.gmail.com, which refactors the code to make it simpler to review.
I had started on this some time ago [1], but couldn't finish it for
the want of time. There are some suggestions from Nathan and Tom which
may be applicable to you approach as well. I am fine if you would like
to use my patches if those help.
--
Best Wishes,
Ashutosh Bapat
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Matthias van de Meent | 2026-03-30 15:26:35 | Re: [BUG] Excessive memory usage with update on STORED generated columns. |
| Previous Message | Sami Imseih | 2026-03-30 15:16:50 | Re: Add pg_stat_autovacuum_priority |