On Wed, Jun 5, 2024 at 8:01 PM Heikki Linnakangas <hlinnaka@iki.fi> wrote:
> I'm very much in favor of a runtime toggle. To be precise, a
> PGC_POSTMASTER setting. We'll get a lot more testing if you can easily
> turn it on/off, and so far I haven't seen anything that would require it
> to be a compile time option.
I was thinking about global variable annotations. If someone wants to
build without multithreading, I think that they won't want to still
end up with a ton of variables being changed to thread-local. So I
think there has to be a build-time option controlling whether this
build supports threading. I suspect there will be other people who
want to just shut all of this experimental code off, which is probably
going to be a second driver for a build-time toggle. But even with
that, we can still have a GUC controlling whether threading is
actually used. Does that make sense to you?
Supposing it does, then how does the extension-marking system need to
work? I suppose in this world we don't want any build failures: you're
allowed to build a non-thread-aware extension against a
threading-capable PostgreSQL; you're just not allowed to load the
resulting extension when the server is in threading mode.
--
Robert Haas
EDB: http://www.enterprisedb.com