> On 28 Jun 2025, at 05:38, Michael Paquier <mich...@paquier.xyz> wrote:
>
> On Fri, Jun 27, 2025 at 02:45:58PM +0500, Andrey Borodin wrote:
>> I'm +1 on having full-fledged injection points in back branches
>> where possible. Right now I'm debugging a PG-17(probably) problem
>> where injection preloading would be handy (though functionality is
>> available via hacks, just a little more work).
>>
>> But are you going to backpatch all new features of injection points
>> in future? It's potentially x6 more work.
>
> That may be nice, but I'd be interested in seeing how things evolve
> across v17 first before taking a decision for older branches.
FWIW both multixact problem [0] and my recent corruption finding [1] would
benefit a lot from having ability to do injection points down to PG 12.
And while [0] is a bug that is detectable with several pgbenches, I have a good
sounding proof that [1] can't happen at all and no way to detect it without
waiting injection point (or similar hand-hacked functionality).
Best regards, Andrey Borodin.
[0]
https://www.postgresql.org/message-id/172e5723-d65f-4eec-b512-14beacb32...@yandex.ru
[1]
https://www.postgresql.org/message-id/B3C69B86-7F82-4111-B97F-0005497BB745%40yandex-team.ru
