Way too easy for 5kyu !

I'm not really familiar with the review process. Are the suggested changes alreadylive on the website or is the old code still running there? In any case on the website I'm currently still getting the same type of error:

expected:<466805.28571428574> but was:<466805.28571428556>

expected:<780934.9302325582> but was:<780934.930232558>

Sorry for closing that issue prematurely, I should've noticed that the tolerance is too small... 🤦‍♂️

Please review & scrutinize Java fork

epsilon of 1e-15 is used but is too small because that would be +-1 21st digit where a double only offers 15 or so

java 17

I believe I fixed the issue that a solution can generate multiple solutions with the same value. Please try again.

In what language?

The issue with the precision at random tests got once resolved due to user simply using the wrong return type.However even using the right return type (as far as I can see) a similar precision issue still occurs (even with much lower deviation as in the older case).
Some examples:

expected:<853157.2307692308> but was:<853157.2307692306>

expected:<672661.1111111111> but was:<672661.1111111109>

expected:<579131.1153846154> but was:<579131.1153846155>

Did you add all necessary imports, specifically for java.util.stream.IntStream?

error appeare when I tried solve task using IntStream:
src/main/java/Solution.java:6: error: cannot find symbol
return IntStream.range(0, arr1.length).map(i -> arr1[i] - arr2[i]).....
^
symbol: variable IntStream
location: class Solution
1 error

Approved

JS translation 🤖🤖

• used mocha/chai + lodash for randomness
• used fuzzy assertion intstead of strict assertion
• added random tests resolving this issue
• provided inputs on failure

Sorry, I got C# and C++ mixed up. Now the C++ translation is ready.

C++ translation

Fake News