Re: pg_recvlogical: Prevent flushed data from being re-sent after restarting replication

On Sat, Jan 10, 2026 at 12:53 AM Mircea Cadariu
<cadariu(dot)mircea(at)gmail(dot)com> wrote:
> Thanks for the updated patch set!
>
> I applied them and it builds correctly and the test passes on my laptop.

Thanks for the test! Barring any objections, I will commit the patches.

> About termination on Windows, I've noticed the following in file 006_logical_decoding.pl:
>
> # some Windows Perls at least don't like IPC::Run's start/kill_kill regime.
> skip "Test fails on Windows perl", 2 if $Config{osname} eq 'MSWin32';
>
> It seems that previously, in a similar situation the decision was to skip it on Windows.
>
> If we adopt the same approach for this test it should be fine, the production code in the patch is not Windows-specific.

Yes, I also noticed that Windows-specific handling while investigating
the test failure. Since IMO it's worth verifying that pg_recvlogical behaves
as expected on Windows, I'd prefer the approach taken in the v4 patch.

> Alternatively if we keep the approach in v4, we can consider updating 006_logical_decoding.pl as well so they're consistent.

That's possible. But TBH I'm not sure how much effort is justified here.
The test uses pg_recvlogical to activate the slot and doesn't really test
pg_recvlogical itself. It's unclear how valuable it is to additionally run
this test on Windows...

Regards,

--
Fujii Masao