> On 16 Mar 2026, at 17:11, Alena Rybakina <lena.ribackina@yandex.ru> wrote:
>
> I moved the tests there. Regression tests are unfortunately not an option because the statistics are not stable.
I think there's no need to test for correct numbers. It would be totally enough
to just for sane numbers or even any numbers at all.
If the test just invokes the function without segfaulting - that's already by far
better than no test at all.
Of course, test that verifies expected behavior is better, if it's possible.
Best regards, Andrey Borodin.
