Re: [multithreading] extension compatibility

On 05/06/2024 23:55, Robert Haas wrote:
> On Wed, Jun 5, 2024 at 4:32 PM Tristan Partin <tristan(at)partin(dot)io> wrote:
>> Not entirely sure how I feel about the approach you've taken, but here
>> is a patch that Heikki and I put together for extension compatibility.
>> It's not a build time solution, but a runtime solution. Instead of
>> PG_MODULE_MAGIC, extensions would use PG_MAGIC_MODULE_REENTRANT. There
>> is a GUC called `multithreaded` which controls the variable
>> IsMultithreaded. We operated under the assumption that being able to
>> toggle multithreading and multi-processing without recompiling has
>> value.
>
> That's interesting, because I thought Heikki was against having a
> runtime toggle.

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 don't think PG_MODULE_MAGIC_REENTRANT is a good syntax. It all looks
> great as long as we only ever need the PG_MODULE_MAGIC line to signal
> one bit of information, but as soon as you get to two bits it's pretty
> questionable, and anything more than two bits is insane. If we want to
> do something with the PG_MODULE_MAGIC line, I think it should involve
> options-passing of some form rather than just having an alternate
> macro name.

+1, that would be nicer.

--
Heikki Linnakangas
Neon (https://neon.tech)