Things already work just as you suggested, and on timeout test output panel presents what tests managed to print.

What is not shown, though, is what tests wanted to print, but did not manage to. That's how buffered output works. When a process is killed on timeout, contents of internal buffers are lost and there's nothing to show. That's why flushing the printed logs explicitly usually helps in such situations.

Hopefully you can find something usefull here: Troubleshooting-your-solution

I have a suggestion pertaining the testing process in kata - specifically, the lack of any feedback whenever the code execution on the server times out.
This specifically annoyed me because I'm struggling with a kata where my code is timing out, and I want to debug it, to test at least how far my code got before it timed out, but whenever the code times out, you only receive one message - that the code execution timed out. Not even the console log.
I don't think it would be particularly hard to simply return what the program wrote before it was terminated automatically by the server, even though I don't know how exactly it works.

My solution is "failing" a random test because the error of 3.725290298461914E-9 is less than the 1e-10 tolerance.

Ridiculous.

Pretty neat, but were 4000 tests really necessary? lol

thank you for asking and answering that. i had also that stupid question

Closing the issue (it was a question btw).

are the tasting cases thoroughly checked?
I'm pretty sure that a prime pair with a gap of 6 exists between 100 and 110, and it's [101, 107]. Both of these numbers are primes, and the gap between them is 6. Why is the expected output null?
Edit - nevermind, I've looked at the instructions - the kata defines a pair of primes as two consecutive primes. Between those two numbers is another prime, 103, and that's why the expected output is null. Sorry for the stupid question.