That's ezxactly my point: it mentions characters, which for Go in general means just about anything in the Unicode tables, of which Latin1 is a tiny part, and ASCII, which is what is actually being tested, and even tinier one. Hence the suggestion to either modify the wording to make it clear the kata is only about single-byte characters sets, or include tests with multiple-byte per character, which will make all SBCS-only solutions break.

I passed it too, but remember because something works does not mean it is correct. The exact same code failed expecting that [""] result (and getting [] instead) then succeeded on a second run, while nothing in the kata suggests [""] can be a valid output in any case.

There appears a bug in the random tests: on one submission, it expected [""] instead of [], which does not match the kata specification. Resubmitting the same code did not trigger the error, probably because the same original source was not generated.

I didn't print the input and it didn't happen the next time I submitted, so I can't see what the default value causing this is.

Ah, I think I understand: added those.

Thanks for your message. I'm afraid I do not understand what I've missed here: the test cases (both full and basic) are the same, as far as I can see. Could you give an example of what I should change ?

The conversion is not needed: under Go language spec, that division is an integer (floor of result).

The tests should include multi-byte characters to break solutions relying on plain byte handling.

The tests should include at least one non-ASCII initial, to break solutions assuming runes are bytes.

Go: description requests an array but expects a slice. It should be reworded to specify a slice return.

The Go translation tests for hello world instead of hello world! with an exclamation point, contrary to the description.

There is something unclear in the problem description. It says the eggs are almost identical (emphasis added), meaning they are not identical, so tries with one egg would not apply to another one: any series of try with one egg is independent from the series with another egg. Is this actually what is meant here ? In the issues thread, someone claims to assume they are actually identically behaving.