Re: PG20 Minimum Dependency Thread

From: Jacob Champion <jacob(dot)champion(at)enterprisedb(dot)com>
To: Devrim Gündüz <devrim(at)gunduz(dot)org>
Cc: Jelte Fennema-Nio <postgres(at)jeltef(dot)nl>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: PG20 Minimum Dependency Thread
Date: 2026-07-16 15:21:43
Message-ID: CAOYmi+=VJJ=1a-wHwWd2mTgQ3bkHzAUHCFssosF7h1y8TEN+iw@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Tue, Jul 14, 2026 at 11:52 PM Devrim Gündüz <devrim(at)gunduz(dot)org> wrote:
> So I think from RPM packaging point of view Python 3.11 is the base on
> SLES 15, and the rest can do even more.

Well, that's cool. In that case 3.9 seems fairly easy, and going
higher than 3.9 is mostly an argument about how much temporary churn
we want for buildfarm owners rather than a question of how much pain
it causes for you. Thanks!

Running tally of my personal votes for PG20:
- Exclude Solaris and Oracle Linux from the standard LTS guarantees
- Declare OpenSSL 1.1.1 out of support, regardless of LTS guarantees
- Adopt N=2 so we can drop SLES 12 and Ubuntu 22
- Declare Python 3.9 the floor for build scripts (SLES 15.6 operators
will have to upgrade, and we'll need to double-check testing a base
Python version that's different from the build version)

Python 3.9 won't quite get us to Meson 1.2.0, but I'm not the right
person to make cost/benefit arguments for that area.

--Jacob

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Paul A Jungwirth 2026-07-16 15:35:39 Re: Fix RLS checks for UPDATE/DELETE FOR PORTION OF leftover rows
Previous Message Andrey Borodin 2026-07-16 15:13:03 Re: PGLZ Compression Optimization