@trashy_incel claims fixed for all languages, if not then please reraise.

Fixed for Python

Fixed for JS

C, C++, C#, Java, JS, Python, Ruby, TS.

That is to say: all of them.

From a quick skim of the tests they all have insufficient coverage as well.

I have an open JS fork which fixes both, and there's an open Python fork which fixes the test coverage but doesn't omit the empty list test.

Can you specify the language that has the issue, please? People can't read your mind, y'know.

I had the same problem, turns out I missed one of the blocking paths, E blocking H -> B...

Here it says it was: https://moonsighting.com/gregorian-calendar.php?1084

Like in 1084, there were 52 weeks and every days were 52 times.

Where did you get this? 1084 was a leap year and had 366 days. As 366 % 7 == 2, it's not possible that all 7 days have equal frequencies.

Also look at this:

{THURSDAY=52, MONDAY=52, SUNDAY=52, TUESDAY=53, WEDNESDAY=52, SATURDAY=52, FRIDAY=52}
//                                  ^^^^^^^^^^

This kata is not about precise chronology. I'll change the random test range in every translation to prevent similar issues/questions in the future.

Regards,

suic

Can confirm what B4B said, just checked and my old solution still passes the test suite in 6.8-7.5s for both Java 8 and Java 11.

you forgot that part (at least in some specific cases):

passing over a point that has already been 'used' like (G and C passing over E).

Hi,

It may happen that j8 is faster than J11 for some kind of task, yes. But a correct/reasonable solution passes consistently the tests with around 8s in J11 (and roughly the same in j8, actually). Meaning you really need to improve your code. Would be useful to spot what kind of structures are slowing J11 compared to J8, tho.