Getting the right floating point precision is tricky in some computer programming languages, namely PHP the one I used for this, depending on which ones you're using indeed, there are also some specific modules dedicated for that making it easier like in Python I guess.

Make sure to cast results to a float

if is not digits in the string - u will have a warning

You can't
1-$s/100
Must
(100-$s)/100

didn't mark this as spoiler, because by math it is the same, computers may have little problem in far decimal places.

In sample test
,missing "{" on first line

Array
(
[0] => Array
(
[0] => 1
[1] => 6356
)

[1] => Array
(
[0] => 2
[1] => 3178
)

[2] => Array
(
[0] => 4
[1] => 1589
)

)

That's not what is expected for 40398736! An array of the sort [[1,6356],[2,3178],[4,1589]] was expected.
Failed asserting that false is true.

Print the input inside function.

Don't worry, translations were really inconsistent, so they'll have to be standarized some day. Done for all, except for Haskell and NASM.

Didn't want to start a brawl, just pointed that out, because beginners can be confused because of that. Regards, Luke.

Ok, I'll remove the function declaration of the ones that already have it then.

Sorry, if the point of this kata is to let users declare the function, maybe just a comment mentioning that is better. This kata is really inconsistent :(

Sure.