For the future please mention the language you are solving in. There was a bug in the TypeScript random test case generation. Now only positive numbers will be generated.

Description states that $-1600 \le r \le 1600$, however the test cases in Python (and possibly other language) only have positive radii. Possible typo?

Approved.

Also forgot that. ✅fixed

Kata is unbalanced between languages. Brute-force works easily in JS while it times out in Python (yeah because Python slow, a big test case runs in under 50 ms in JS while the same takes more than 2 seconds in Python).

For some reason the solution was present in the sample tests. I have removed it now.

A last thing: Please add assertion messages also to Sample Tests, then I will approve.

I have no idea how this works

Totally my bad. Should have looked into that a little deeper. TOP = 15 should be fine now since:

Long.MAX_VALUE -> 9223372036854775807
        10**15 -> 1000000000000000

20 is still too big but I tested and the refsol actually threats the number as unsigned so wrap-around would not be a problem.

Fixed ✅
TOP is 20 now

Random tests only generate tests where recipe and available ingredients match.

• 10 ** 32 overflows long. Groovy seems to cut it off at Long.MAX_VALUE though. Two randoms drawn from this range could overflow long easily and your refsolution will let them wrap around.
• sample test method is called returnsProduct.

Approved.

Fixed. Had an error in the random test generation, that's why there were no tests for the last column but the others could overflow.