tried without any predefined looping functions

Be aware that the stack space here can become huge for bigints. It's a better practice to use tail recursion for constant stack space.

Please mark your post as having spoiler content next time,

Read the description again:

MUTATE the input array/list, don't return anything.

There can be valid reasons to mutate inputs ( eg. performance. did you ever stop and wonder why eg. default JS sorting behaviour is modification in place? ). So if a kata specifically specifies ( and tests! ) this behaviour, you can have your doubts, but it makes a valid kata.

I fully agree side-effect-free programming has its advantages ( as well as its disadvantages, but you're not even mentioning those ). But single-mindedly forbidding valid language features in kata, or even kata about valid language features, is just closed-minded. Everything's a trade-off, including side-effect-free programming.

Updated JS version for this issue ( both, actually ).

Related, but not exactly the same thing: should probably give a useful failure message when solver returns a new value instead of mutating the input.

This is a 7kyu kata; solvers may actually not know the difference.

Try reading up on "side effect".

MUTATE the input array/list

It might be a good idea to research how arguments are passed by reference in python. I'd suggest googling for it.

The solution here is simple, but these test cases test the side effects of this function and not an actual return value, how you'd expect in a proper test scenario.

I'm not sure how to manipulate the global values from the test case in python.

Thanks for the clarification. They are very useful and broaden my knowledge on this subject.

Imagine someone is having two last names like it's normal in hispanic cultures. The code only works within the limitations of the exercise requirements.