either way, the suggestion won't be implemented, as it would invalidate too many results on an approved kata

for me, it was not clear at all from the description, which sequences to include and which to discard

.

Approved

perhaps the restrictions aren't tight enough

resolved

Well, there's no happiness without suffering; congrats on solving!

Though you're making me question whether I should add a barrier against hardcoding...

biggest number in sample test is if i am not wrong something like ~400th Fibonacci number. in random tests they starting from ~500 000th Fibonacci number, so there is no way to calculate square root of that in reasonable amount of time

i will translate to python probably tomorrow. by the way how do you test squares ? because if you test it searching for root there is no chance to handle those numbers

I have now added some random tests! Thanks for the reminder.

Added 100 random tests.

thank you!
I will read through this and implement random tests.

It looks like generating random arrays takes most of the time. The ref solution is very fast.

Maybe since now input arrays are deeply copied, we need to reduce the number of huge random tests.

Is it still relevant? I made a correction a few minutes ago.