Re: AIX support

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Peter Eisentraut <peter(at)eisentraut(dot)org>
Cc: Srirama Kucherlapati <sriram(dot)rk(at)in(dot)ibm(dot)com>, Aditya Kamath <Aditya(dot)Kamath1(at)ibm(dot)com>, Michael Paquier <michael(at)paquier(dot)xyz>, Noah Misch <noah(at)leadboat(dot)com>, Andres Freund <andres(at)anarazel(dot)de>, Robert Haas <robertmhaas(at)gmail(dot)com>, "pgsql-hackers(at)lists(dot)postgresql(dot)org" <pgsql-hackers(at)lists(dot)postgresql(dot)org>, "hlinnaka(at)iki(dot)fi" <hlinnaka(at)iki(dot)fi>, "tristan(at)partin(dot)io" <tristan(at)partin(dot)io>, "postgres-ibm-aix(at)wwpdl(dot)vnet(dot)ibm(dot)com" <postgres-ibm-aix(at)wwpdl(dot)vnet(dot)ibm(dot)com>
Subject: Re: AIX support
Date: 2026-02-19 23:55:12
Message-ID: 2708568.1771545312@sss.pgh.pa.us
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I wrote:
> Peter Eisentraut <peter(at)eisentraut(dot)org> writes:
>> Regarding patch 0004, please see the thread
>> https://www.postgresql.org/message-id/flat/e8aa97db-872b-4087-b073-f296baae948d%40eisentraut.org
>> which is developing a more general and more robust version of that patch.

> Ah. I don't object to having a more general version, as long as
> the AIX patch can auto-disable static libraries rather than having
> a gotcha "it breaks with a mysterious error if you don't say this".

Attached is an updated AIX patchset based on the assumption that
you'll first commit the v3 patch in that thread (or some close
relative of it). v7-0001 below is exactly that v3 patch, and then
the rest are rebased on top of it.

With that patch, v6-0004 became vestigial, amounting to just the line
+ build_static_lib = false
plus some commentary. So I folded it into the new
v7-0004-Various-minor-adjustments-for-AIX.patch and also merged the
requested addition of /opt/freeware/lib into that patch. The other
patches are the same as before except for some very minor rebasing
adjustments.

Also, now that p9-aix1-postgres1 has a full Python installation
(thanks for that!), I tried --with-python, and sure enough it did
not work. But the cause turns out to be quite trivial: python.m4
has hard-wired knowledge about the set of plausible shlib file
extensions, and ".a" is not one of them. So v7-0008 attached
fixes that. The meson side seems fine already.

To be clear, I anticipate squashing all the AIX-specific bits
(0003-0008) into one commit in the end. But 0001 and 0002
seem worth committing separately in case they affect any other
platforms. In the meantime, keeping the bits separate seems
like it might ease review.

To my mind this is pretty nearly ready to go. It's waiting
on Peter to commit his patch (0001), and I'd really like to
retest on AIX 7.3. But cfarm119 is still offline and I won't
wait for it forever.

regards, tom lane

Attachment Content-Type Size
v7-0001-meson-allow-disabling-building-installation-of-st.patch text/x-diff 11.8 KB
v7-0002-Cope-with-AIX-s-alignment-woes-by-using-_Pragma-p.patch text/x-diff 61.5 KB
v7-0003-Revert-removal-of-AIX-support.patch text/x-diff 27.7 KB
v7-0004-Various-minor-adjustments-for-AIX.patch text/x-diff 6.8 KB
v7-0005-Force-64-bit-builds-on-AIX.patch text/x-diff 9.5 KB
v7-0006-Add-some-definitions-needed-for-a-clean-compile.patch text/x-diff 2.0 KB
v7-0007-Fix-inconsistencies-in-the-set-of-installed-files.patch text/x-diff 4.6 KB
v7-0008-Fix-libpython-configure-check-for-AIX.patch text/x-diff 1.8 KB

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Manni Wood 2026-02-20 00:09:27 Re: Speed up COPY FROM text/CSV parsing using SIMD
Previous Message Melanie Plageman 2026-02-19 23:41:43 Re: Checkpointer write combining