just brilliant

sort is bad, but reverse is not that bad.

Nice one

Hi,

return the time to the second during which...

Reading this, I except to have to use rounding. Tho, it's flooring that is expected. => missing in the description. That, or you clearly explain that you're supposing that the move of the minute hand is not continuous, and it moves a tiny bit at the end of each second (which might cause more troubles than talking about flooring directly...)

Yeah we discussed elswehere and recognized I was being particularly daft.

There is at most one alignment per hour, no ambiguity.

For example, on a 12-hour clock:
hour=1 -> "1:05:27", previous: 12:00:00 (hour=12), next: 2:XX:XX (hour=2)

At the top, but this is not stated in the description. Could the description be edited?

Your fix in the random tests solved this issue, thanks.

Thanks, I've adjusted the random tests to account for the discrepancy

Where is the minute hand initially?

I get the same results as you with my initial solution. See my topic below. I believe it's a discrepancy in solutions that calculate in minutes vs those that use seconds as a base unit.

The random tests expect incorrect results sometimes:

given 13 on a 51 hour clock
expected '13:15:36' to deeply equal '13:15:35'

Maybe you could phrase it better. But I'll take it.

Some extra line breaks, commas, and a mention of the second parameter, might make it more readable ( and not actually incomplete ).

Also, try to make "h:mm:ss" look like a string? "h:mm:ss"

Probably could phrase it better, rounding is specified by returning the 'second during which the hands cross', if you round up that'd be the second after? And if rounding up was allowed there be no situation where you'd return an empty string