fixed (and added edge cases to fixed part of full tests)

Already approved by someone.

About the test cases in the fork: you're calling my solution function instead of your own function, and then everything passes... which is to be expected. Your code is not even called. Only in the sample tests do you call your own code.

About the random thing being equal each run; that is of course by design. If you want to change that, srand() on time.

About your test failing; you can check for yourself what the inputs are and what the output should be. If your code cannot reproduce it, the issue is really with your code.

As for forking, I don't really get your intention. I will have a quick look at it, but don't have the time to fully analyze it. You indicated it shouldn't be accepted, hence I'll reject it.

Actually, I tested everything with + EPS, then I tried to remove it before the final submission and it passed, so I don't know if it would have passed every time or not. I've added it back now.

Um... depends on the implementation, there could be many 'edge' cases.
You could simply raise the times from 20 to 100 if you wish o_O

It's not slow, just an edge case like

[0.06063410508099362, 0.16237818048291924, 0.12982549834001908, -0.1589744979205776, -0.19717508632071187, -0.12542437944438573, -0.26636647957973736, 0.19592829895911112]

will cause it stuck into an infinite loop
because the initial step is too huge for cases like this.

Somehow it doesn't even pass the sample tests, yet it's not invalidated O_o

Thanks! I'm off at a conference right now, but will look into this this weekend after I ger back. I certainly used date objects myself.

Thanks, I'm glad you enjoyed it! At the time I was working on it I had ideas for both a part 2 and a part 3, and hope to get back to part 2 at some point (although the details of part 3 have completely slipped my mind). Though there's also the issue that the cat dynamics here have completely changed again since I wrote part 1, hmmm ...

Yeah, most users don't report issues at all.

I wrote something about it in the kata comment (to mark it as spoiler).

https://www.codewars.com/kata/base91-encoding-and-decoding/discuss/go#5aecb7bd783bb4329e0000f7

All of the solutions from the past seems to use ' for 90 so the issue was there from the beginning.

I don't think so. I only changed %s to %d.

Resolved

Cool. I have no idea why this happened, but I'm glad it worked :)