Re: Adding facility for injection points (or probe points?) for more advanced tests

On Fri, Jan 05, 2024 at 08:38:22AM +0900, Michael Paquier wrote:
> On Thu, Jan 04, 2024 at 04:31:02PM -0600, Nathan Bossart wrote:
>> Rather than defining a module somewhere that tests would need to load,
>> should we just put the common callbacks in the core server? Unless there's
>> a strong reason to define them elsewhere, that could be a nice way to save
>> a step in the tests.
>
> Nah, having some pre-existing callbacks existing in the backend is
> against the original minimalistic design spirit. These would also
> require an SQL interface, and the interface design also depends on the
> functions registering them when pushing down custom conditions.
> Pushing that down to extensions to do what they want will lead to less
> noise, particularly if you consider that we will most likely want to
> tweak the callback interfaces for backpatched bugs. That's also why I
> think contrib/ is not a good idea, src/test/modules/ serving the
> actual testing purpose here.

Ah, so IIUC we'd have to put some functions in pg_proc.dat even though they
would only be used for a handful of tests in special builds. I'd agree
that's not desirable.

--
Nathan Bossart
Amazon Web Services: https://aws.amazon.com