That is a clever use of replace and split! Well done.

ocelot269 would benefit more if you would explain why it is inefficient...

If I knew that split() does a better job in removing " " * n than split(" ").
Well I guess that's why I here. To learn, I guess.

... check your damm '()'... I overlocked that there was a ')' missing in my print statements like so:
print(iter_pi(0.01 == [100, 3.1315929036]))
(I have tried to avoid the syntax error by adding a ')' at the end -fatal.
I assumed that my code was just too slow and surrendered.

Then I saw that there were proper solutions similar to mine.
After a short check I knew that my code was okay but my print not.

What a bad experience. Though not your fault ofcourse.

I think for a level 6 kata, performance should not play that a big role...

Wow! I'm blinded by science!

I find 5kyu overrated as well. It was fun anyway.

removed

For C the extra test case

tester(858993458, 18446744043644780550U);

far exceeds the range of input used for the Java version of this Kata.

That has turned this Kata into a "performance" Kata, so now it limits what kinds of solutions can work. I don't think that was intent of the original Kata.

Polynomial regression is kinda overkill but works just fine

approved!

You could use a font-independent illustration in the description, e.g.

X X X X X
. . . . X
X X X . X
X . . . X
X X X X X

Ruby translation kumited -- please check and approve!

This one made me bite my tongue more than once....