Re: Meson far from ready on Windows

From: walther(at)technowledgy(dot)de
To: Andres Freund <andres(at)anarazel(dot)de>, Dave Page <dpage(at)pgadmin(dot)org>
Cc: Nazir Bilal Yavuz <byavuz81(at)gmail(dot)com>, PostgreSQL Developers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Meson far from ready on Windows
Date: 2024-08-18 14:30:11
Message-ID: c20aed6f-ec30-4a1f-83a7-ddf322059a8b@technowledgy.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Andres Freund:
>> That's not necessarily true. The nix package manager and thus NixOS track all dependencies for a piece of software. If any of the dependencies are updated, all dependents are rebuilt, too. So the security concern doesn't apply here. There is a "static overlay", which builds everything linked fully statically.
>
> Right. There's definitely some scenario where it's ok, I was simplifying a bit.
>
>> Unfortunately, PostgreSQL doesn't build in that, so far.
>
> I've built mostly statically linked pg without much of a problem, what trouble did you encounter? Think there were some issues with linking Kerberos and openldap statically, but not on postgres' side.

Mostly the "can't disable shared libraries / backend builds" part
mentioned below.

> Building the postgres backend without support for dynamic linking doesn't make sense though. Extensions are just stop ingrained part of pg.

I think there might be some limited use-cases for a fully-static
postgres backend without the ability to load extensions: Even if we get
libpq to build fine in the fully-static overlay mentioned above, a lot
of reverse dependencies have to disable tests, because they need a
running postgres server to run their tests against.

Providing a really simple postgres backend, with only minimal
functionality, would allow some basic sanity tests, even in this purely
static environment.

>> Lately, I have been looking into building at least libpq in that static overlay, via Meson. There are two related config options:
>> -Ddefault_library=shared|static|both
>> -Dprefer_static
>>
>> The first controls which libraries (libpq, ...) to build ourselves. The second controls linking, IIUC also against external dependencies.
>
> Pg by default builds a static libpq on nearly all platforms (not aix I think and maybe not Windows when building with autoconf, not sure about the old msvc system) today?

Yes, PG builds a static libpq today. But it's hard-to-impossible to
*disable building the shared library*. In the fully static overlay, this
causes the build to fail, because shared libraries can't be build.

>> Maybe it would be a first step to support -Dprefer_static?
>
> That should work for nearly all dependencies today. Except for libintl, I think. I found that there are a lot of buglets in static link dependencies of various libraries though.

To support prefer_static, we'd also have to look at our internal
linking, i.e. whether for example psql is linked against libpq
statically or dynamically. Once prefer_static controls that, that's
already a step forward to be able to build more of the code-base without
shared libraries available.

Best,

Wolfgang

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Alexander Lakhin 2024-08-18 16:00:00 Re: The xversion-upgrade test fails to stop server
Previous Message Peter Eisentraut 2024-08-18 14:07:48 Cirrus CI for macOS branches 16 and 15 broken