The vector content should stay the same

refers to the fact that the vector must still contain the same elements after being sorted, i.e. you cannot add, delete, or replace elements. such a vector can still be wave-sorted, there is no contradiction

fixed there by displaying doubles to full precision

is your cube round?

Approved, thanks!

Approved.

Approved, thanks!

Rust translation

After some more fiddling, I've got that exactly floating point formula

price += ceil(rental / people_left * days_span)

produces the expected result. Changing the order, like first multiplying and then dividing, gives a totally different outcome.

And no, this is not a spoiler, this is an absolutely needed information to complete this kata - you should not just round up, but do calculations in double-precision floating point in this specific order. With the current tests, it should be in the description (or tests should be changed).

Can anyone explain HOW THE HELL should numbers be rounded? I got 5 medium tests failing, everything else (125 tests) fine. I printed data and got this:

{770: 91453, 1155: 139009, 2387: 297527, 3003: 380232, 6237: 834168, 8470: 1162526, 9086: 1257637, 9779: 1370268, 10010: 1409898, 12320: 1829503, 12859: 1933530, 14938: 2361527, 15862: 2565335, 16016: 2601916, 18172: 3156727, 18557: 3264807, 21714: 4239689, 21945: 4318948, 22715: 4616168, 23177: 4819976, 24255: 5374787, 24871: 5755229, 25795: 6468557, 26026: 6825221, 27412: 11105189} should equal {770: 91453, 1155: 139009, 2387: 297527, 3003: 380232, 6237: 834169, 8470: 1162527, 9086: 1257638, 9779: 1370269, 10010: 1409899, 12320: 1829504, 12859: 1933531, 14938: 2361528, 15862: 2565336, 16016: 2601917, 18172: 3156728, 18557: 3264809, 21714: 4239691, 21945: 4318950, 22715: 4616170, 23177: 4819978, 24255: 5374789, 24871: 5755231, 25795: 6468559, 26026: 6825223, 27412: 11105191}

The difference is in day 6237. The corresponding log:

On day = 770, 1 leaves paying 91453 (rounded from 770*3088/26=1188880/13)
On day = 1155, 1 leaves paying 139009 (rounded from 385*3088/25=237776/5)
On day = 2387, 1 leaves paying 297527 (rounded from 1232*3088/24=475552/3)
On day = 3003, 1 leaves paying 380232 (rounded from 616*3088/23=1902208/23)
On day = 6237, 1 leaves paying 834168 (rounded from 3234*3088/22=453936)

In expected result, day 6237 should be 834169.
Can someone open my eyes and show how 380232 + 3234 * 3088 / 22 (which is 834168) is 834169 after rounding up? What kind of black magic is expected to be used?

Rust translation

approved

C++ test fails:

Expected: equal to 0.026314
Actual: 0.026314

Really?

Fixed EE

Changed the random speed to positive, optimized solution to nonsense, removed all unneeded high-level abstractions, and run rustfmt over all the code.

Good enough now?

fixed