The C++ randomized tests are ridiculously more difficult than the Javascript ones. Here are a few randomized tests from the Javascript cases:

money=2962, coins=[342, 381, 358, 400]
money=0, coins=[14, 482, 58]
money=1504, coins=[376]
money=0, coins=[108, 358, 319, 439]

and the madness that are the randomized C++ cases:

money=45290, coins=[1, 2, 2052, 2414, 3758, 3768, 4685, 6607, 6852, 9886, 11599, 12309, 12537, 14983, 19645, 21142, 21952, 22199, 25886, 26096, 26608, 28288, 29523, 30689, 36574, 37546, 39867, 40640, 41033, 43250]
money=71905, coins=[1, 2, 1091, 3875, 8351, 10746, 11519, 14011, 22480, 22954, 23537, 26308, 30694, 35073, 40862, 41930, 42306, 47965, 50558, 54438, 57510, 57864, 61605, 63530, 67173, 67536]
money=71119, coins=[1, 2, 2473, 24198, 24931, 31298, 39847, 42517, 50857, 54231, 59625, 66036]

I ended up implementing an optimization just for these 1, 2, big_number, ..., big_number cases, which funnily enough is enough to pass them (assuming that the rest of your code is somewhat efficient). Otherwise a recursive solution simply isn't performant enough.

Fixed.

That was a typo, fixed

That won't work with inputs starting or containing zeroes ([0, 1, 2] will simply return 0)

Yes, there was a merge conflict because of that. Forked and fixed the conflict here:

https://www.codewars.com/kumite/5cd3387ffb6f25456f8806b4?sel=5cd887ca5bf451002485da8e

Have you modified the description between now and then?

https://www.codewars.com/kumite/5cd86e4ecedf620011e59dfa?sel=5cd86e4ecedf620011e59dfa

^ Javascript translation :)

JS has Math.PI, but it only goes up to the 64bit float precision, so no guarantee past the 12th-13th digit I'd say

Javascript here, you should provide some sample test cases. I'm flagging this as an issue, as it is essential for a Kata to provide them

Mathematically, the sum of nothing is zero (it's called an empty sum). The product of nothing is 1 (this one is called an empty product, sometimes marked with _)

Wait a minute...

CodeWars should mention that somewhere (it doesn't yet). All my apologies, it is indeed about the C implementation

There is an indexOf method on arrays too :)

Nevermind, read it wrong