В письме от понедельник, 7 марта 2022 г. 20:04:49 MSK пользователь Greg Stark
написал:
> I don't think this is worth spending time adding tests for. I get what
> you're saying that this is at least semi-intentional behaviour and
> desirable behaviour so it should have tests ensuring that it continues
> to work. But it's not documented behaviour and the test is basically
> testing that the implementation is this specific implementation.
>
> I don't think the test is really a bad idea but neither is it really
> very useful and I think it's not worth having people spend time
> reviewing and discussing. I'm leaning to saying this patch is
> rejected.
This is a regression test. To test if behaviour brocken or not.
Actually I came to the idea of the test when I wrote my big reloption patch.
I've broken this behaviour by mistake and did not notice it as it have not
been properly tested. So I decided it would be goo to add test to it before
adding and big patches.
--
Nikolay Shaplov aka Nataraj
Fuzzing Engineer at Postgres Professional
Matrix IM: @dhyan:nataraj.su