Unmark gen_random_uuid() function leakproof

Hi,

On 2024-12-09 14:10:30 -0800, Masahiko Sawada wrote:

While reviewing UUIDv7 patch[1], I realized gen_random_uuid() is
marked leakproof even though it doesn't take arguments. The functions
without arguments don't need to be marked leakproof in principle. This
is the sole function that has no arguments and is listed in the "List
of built-in leakproof functions" in opr_sanity.sql. I've attached the
patch for fixing it and for better consistency with new UUID
generation functions discussed on that thread.

Seems like it'd make sense to add a test to opr_sanity.sql so we don't
reintroduce such cases?

Greetings,

Andres Freund

On Mon, Dec 9, 2024 at 2:23 PM Andres Freund <andres@anarazel.de> wrote:

Hi,

On 2024-12-09 14:10:30 -0800, Masahiko Sawada wrote:

While reviewing UUIDv7 patch[1], I realized gen_random_uuid() is
marked leakproof even though it doesn't take arguments. The functions
without arguments don't need to be marked leakproof in principle. This
is the sole function that has no arguments and is listed in the "List
of built-in leakproof functions" in opr_sanity.sql. I've attached the
patch for fixing it and for better consistency with new UUID
generation functions discussed on that thread.

Seems like it'd make sense to add a test to opr_sanity.sql so we don't
reintroduce such cases?

Thank you for the comment. It's a good idea. I've updated the patch.

Regards,

--
Masahiko Sawada
Amazon Web Services: https://aws.amazon.com

On Mon, Dec 9, 2024 at 2:48 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:

On Mon, Dec 9, 2024 at 2:23 PM Andres Freund <andres@anarazel.de> wrote:

Hi,

On 2024-12-09 14:10:30 -0800, Masahiko Sawada wrote:

While reviewing UUIDv7 patch[1], I realized gen_random_uuid() is
marked leakproof even though it doesn't take arguments. The functions
without arguments don't need to be marked leakproof in principle. This
is the sole function that has no arguments and is listed in the "List
of built-in leakproof functions" in opr_sanity.sql. I've attached the
patch for fixing it and for better consistency with new UUID
generation functions discussed on that thread.

Seems like it'd make sense to add a test to opr_sanity.sql so we don't
reintroduce such cases?

Thank you for the comment. It's a good idea. I've updated the patch.

I'm going to push the updated patch tomorrow, barring objections and
further comments.

Regards,

--
Masahiko Sawada
Amazon Web Services: https://aws.amazon.com

On Tue, Dec 10, 2024 at 1:52 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:

On Mon, Dec 9, 2024 at 2:48 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:

On Mon, Dec 9, 2024 at 2:23 PM Andres Freund <andres@anarazel.de> wrote:

Hi,

On 2024-12-09 14:10:30 -0800, Masahiko Sawada wrote:

While reviewing UUIDv7 patch[1], I realized gen_random_uuid() is
marked leakproof even though it doesn't take arguments. The functions
without arguments don't need to be marked leakproof in principle. This
is the sole function that has no arguments and is listed in the "List
of built-in leakproof functions" in opr_sanity.sql. I've attached the
patch for fixing it and for better consistency with new UUID
generation functions discussed on that thread.

Seems like it'd make sense to add a test to opr_sanity.sql so we don't
reintroduce such cases?

Thank you for the comment. It's a good idea. I've updated the patch.

I'm going to push the updated patch tomorrow, barring objections and
further comments.

Pushed (398d3e3b5b).

Regards,

--
Masahiko Sawada
Amazon Web Services: https://aws.amazon.com