Hi,
On 2022-07-04 21:56:24 -0400, Tom Lane wrote:
> > For non-parallel tests I think we currently print the test name before running
> > the test, which obviously doesn't work well when needing to print the 'ok'
> > 'not ok' first.
>
> Is this still a consideration? We got rid of serial_schedule some
> time ago.
Not really for the main tests, there's a few serial steps, but not enough that
a bit additional output would be an issue. I think all tests in contrib are
serial though, and some have enough tests that it might be annoying?
> > I wonder if for parallel tests we should print the test number based on the
> > start of the test rather than the finish time?
>
> I think we need the test number to be stable, so it had better be the
> ordering appearing in the schedule file. But we already print the
> results in that order.
I remembered some asynchronizity, but apparently that's just the "parallel
group" line.
Greetings,
Andres Freund