Re: More speedups for tuple deformation

Hi,

On 2026-02-26 17:01:57 +0100, Álvaro Herrera wrote:
> On 2026-Feb-26, Andres Freund wrote:
>
> > I don't get this. We'll approximately never have a full suite of benchmark
> > tools that make sense to integrate at once. The reason we're here is because
> > over and over we've not actually merged quite useful tools, making do with
> > benchmarking-by-proxy and one off tools. This seems like a recipe for never
> > getting anywhere better.
>
> Eh, my first suggestion was to start the benchmark suite as a top-level
> thing, so put this module under src/benchmark rather than hide it with
> the bunch of smelly dudes in src/test/modules.

I don't particularly care about whether we put it in src/benchmark or
src/test/modules. I don't love either, so whatever.

But I think it's reasonably important that we don't go for "one extension for
every single C benchmark function", because that'll just make the overhead for
adding another benchmark function too high.

And I think we shouldn't install anything in there by default (like with
src/test/modules), as otherwise we'll have to care about the benchmark module
being perfectly safe, which I don't think we want to.

Greetings,

Andres Freund