It's difficulty comes from how unclear the directions can be.

Haskell fork with random tests.

Haskell translation

Empty object => {}

fun :)) +rep

Lua Translation

Hi there! If you want to solve hard katas, you should try kyus of higher than the lowest level. Hope this helps!

so hard

this is a prime example of why floating-point numbers should never be used to represent monetary amounts. the problem is that the reference solution does 111.0 // 3.7 to calculate the total number of beers, which yields 29.0, not 30; floor(111.0 / 3.7) would yield 30.

Sometimes my submission (using Python 3.11) is accepted, and sometimes the same submission fails. Here is one that failed:

beeramid(111, 3.7)

It reports it should be 3, but 4 (in my answer) = 111.
It happens whether I round the cost at each step to 2 decimal places or not.

rownum: 1
cumulativecost: 3.7

rownum: 2
cumulativecost: 18.5

rownum: 3
cumulativecost: 51.8

rownum: 4
cumulativecost: 111.0

So how is the expected answer 3 not 4? That sounds wrong to me.

thanks a lot!
it would be great to have typed functions to make much easier to understand the conditions of the kata.

I would be happy to updated them if that's possible.

Although not explicitly stated, the sample tests indicate that the price isn't always an integer, so your calculation might be wrong, and also you type hint the input (see what happened when in not strict mode?).

Closing as this is not an issue with the tests.