Per the description:

Test constraints:

30 sets of random source paints will be tested.
Each source paint will have an R value and a G value ranging from 0 to 1.
Each set of paints will contain up to 10 unique paints. (I create 10, then dedupe.)
Each set of paints will be tested by sampling a 40x40 grid. i.e. the point (0,0) will be tested, then (0,0.25), then (0,0.5) etc, up to (0,0.975), at which point the test will roll over to (0.25,0). Your solution should be able to handle these 48000 validate queries rather quickly - my solution gets them in around 1-2 seconds.

This isn't a bug. You're not passing because you're breaking the tests, you're passing because the operation you're doing in your solution is a valid way to recolor all the nodes. It's not one that I like, but it's valid, and trivializes the issue.

Perhaps the problem should force certain nodes to start as certain colors, in order to disable this trick? If so, that should be a separate kata. I'm certainly not coding it up here.

Regeardless, nice solution. Resolving.

Now deepcopying inputs prior to administrating tests, so mutation doesn't affect validation. Ty for callout.

Thank you for your kind compliment. I try.

thanks <3

also nice job my guy :-D

• Fixed up description.
• Correct re: overlap. If I get a string ATCGAAAT, that means that the repeated string would be ATCGAAATATCGAAAT, not ATCGAAATCGAAAT. If there is an AT at the front and end, that is because it is repeated.
• Re: Test cases: Fine, I added 3 more. I'm not going to get torn up over missing test cases, the code in the actual test cases is pretty comprehensive, and has useful debug messages. Also, some of the code for checking error_prone tests is revealing of how to deal with error_prone cases IMHO, so it shouldn't be shown.

ty

Bump.

• Could you please link to docs on the new testing framework? I'm unfamiliar with new CodeWars updates.
• Strings only get output if tests fail. If your solution actually works, no output is generated, so no issue. If you solution doesn't work, better to have slow tests that describe your failure than fast tests that tell you nothing.

It isnt as complicatedWhen you get it you will feel dumb, I assure you. I sure did.

Always return the minimum number of steps possible to converge all 3 coins.

Thanks for compliment :)

.

Also, Bellman Ford is for weighted graphs. Which this is not. Optimization is possible.

Fixed.