it's the second one that's expected
that wording is terrible though because it refers to the assertion code which the solver doesn't get to see and it's all too easy to see "expected x" and then not read the rest

I failed about 10 tests.
For example
Testing for: "hilpbwsfcpkfygvdjzmbb"
expected [ 4, 'f' ] to deeply equal [ 1, 'b' ]
Why is [ 4, 'f' ] expected if the minimum in the string is clearly [ 1, 'b' ]?

Fixed.

Fixed. My apologies for the mistake.

See the issue above.

Haskell: the expected value for "bbaa" should be Just (1, 'b') but the reference solution returns Just (1, 'a').

The output should have been an optional tuple of an int and a char, not a nullable, necessarily heterogeneous, array. This is overly monolinguistic.

I'm not sure what can be done now. Python ( and, proposed, Haskell ) correctly use a tuple, but the description suffers.

Haskell translation

this translation modifies the description

intersting , thanks anthor

Nice kata to warm up

Yeah, I completely agree with this.

Really cool kata :D

Yes thanks, I fixed it.

Java: Since the second item of the array is just a character, why not expect a char instead of a string?