Re: pg_upgrade fails with non-standard ACL

27.09.2019 15:51, Bruce Momjian wrote:
> On Fri, Sep 27, 2019 at 04:22:15PM +0900, Michael Paquier wrote:
>> On Thu, Sep 26, 2019 at 04:19:38PM -0400, Bruce Momjian wrote:
>>> On Thu, Sep 26, 2019 at 05:16:19PM -0300, Alvaro Herrera wrote:
>>>> On 2019-Sep-26, Bruce Momjian wrote:
>>>>> Well, right now, pg_upgrade --check succeeds, but the upgrade fails. I
>>>>> am proposing, at a minimum, that pg_upgrade --check fails in such cases,
>>>> Agreed, that should be a minimum fix.
>>> Yes.
>> Agreed as well here. At least the latest patch proposed has the merit
>> to track automatically functions not existing anymore from the
>> source's version to the target's version, so patching --check offers a
>> good compromise. Bruce, are you planning to look more at the patch
>> posted at [1]?
> I did look at it. It has some TODO items listed in it still, and some
> C++ comments, but if everyone likes it I can apply it.

Cool. It seems that everyone agrees on this patch.

I attached the updated version. Now it prints a better error message
and generates an SQL script instead of multiple warnings. The attached
test script shows that.

Patches for 10, 11, and 12 slightly differ due to merge conflicts, so I
attached multiple versions.

--
Anastasia Lubennikova
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company