here is an example https://www.codewars.com/kata/reviews/5286a298f8fc1b7667000c1f/groups/55633cb19b780ca9fb000031

Kata is good, it's great you've made it.
I agree that adding tests for memory limitations is not a trivial thing. So I'll try some ideas, maybe would found how to do that properly ...

Sure, feel free to submit a pull request to improve the tests.
This kata was made almost 10 years ago, out of generosity on my free time, and is free.
You can't expect it to be bulletproof or with a gold level support.
And, if you try to think as the test developper, you'll see it is challenging to simulate memory limitation with the provided tools.

Btw, Which solutions violate the rule ?

The whole point of this kata is that it is possible to respect the rule and code efficiently. If choose not you, it's your problem, not really mine.

of course this shouldn't work, but works with some probability

.

now it works, cool!

yes, "Max Buffer Size Reached (1.5 MiB)" is because of the large number of test cases. Maybe it is possible to output only some part of the diff, I'm not sure how it works

in SQL for n = 40207 expected is 39800
why not 40200 ?

also I found that I get buffer overflow error when attempt so I see no diff. Not sure whether that could be fixed

in SQL ref solution returns a=4 b=20 res=... for input a=20 b=4
which is somewhat unexpected
If modifying the input is ok - that should be stated somewhere I think

to convince postgresql run my query before others

Why so polite? :-)

Your solution mutates the input, that's why you get this weird behaviour

Fixed Java, should no longer accept O(n^2) solutions.

oh, that's right logic :-)