It's bad practice, but I love how optimized it is.

excelente me sorprende a veces la capcidad en la que los codigos de php se pueden comprimir a tal punto en que solo ocupan una sola linea de codigo.

Currently my code will sometimes pass, sometimes failed.
And if it failed, the wrong one would be like this:

38655
expected '38.65s' to equal '38.66s'

Or like this:

325297327
expected '3.76d' to equal '3.77d'

Then how about 23135?
56.385 round down to 56.38.
23.135 round up to 23.14.
By the way, these two message is from different code.

Resolved.

56385ms is 56.385s, as you can see in example 1.555 should not be rounded to 1.56. So the answer for this will be 56.38s.

(eg. 1.555 => 1.55, but 1.556 => 1.56) What? Round up 6!? OK...
56385: expected '56.39s' to equal '56.38s'
23135: expected '23.13s' to equal '23.14s'
WTF?

It may seems like these katas are similar, but each of it contains unique logic of formating time. Also, the purpose of my kata is to simulate node's console.time. For that reason it could be interesting :)

Thanks a lot. I will add some tests later.

I think it needs more boundary tests in the fixed submisson tests.

At the moment some are covered for seconds, but not for the other units.

Also could do with boundary cases for removing of excess 0's, like 5100 returns "5.1s" and not "5.10s". These might be caught in Random tests, but not guaranteed to be.

Other than that, thought it was a nice little kata with a clear description.

Ok, I have updated the description. I hope it's clear now :)

The answer also needs to be rounded.

How can users know it? It's not explained how to round the answer.

It's 70 seconds, so the expected output is 1.17m (1.17 * 1000 * 60 ~ 70000). The answer also needs to be rounded.