Well, the input is defined strictly as 1000 ≤ year ≤ 9000.

See https://www.postgresql.org/docs/current/queries-values.html

The query is roughly equivalent to select * from (values (''good''), (''ok''), (''bad'')) as t.

There is a memory leak, the buffer is not freed. It is missing the free(buffer); before the return.

I have just made a more concise solution ;) https://www.codewars.com/kata/reviews/57cb045535ef62f92d0000f1/groups/610d7e5b96e25100013cae0d

This solution will fail when given "0" as $str.

Does not work for the argument $str = "0". Returns empty array, instead of ["0_"];

Indeed, thank you for pointing this out.

Was the last line of your example meant to be remove_char(str2, str1);? Your solution would overflow in that case.

For example, assuming the str1 consists of 6 characters (including null terminator) The str2 is correctly allocated 4 characters (to fit two less characters). When passed to your implementation of the remove_char the call strcpy(dst, src + 1); writes 5 characters to dst/str2 thus overflowing.

The problem is the function overflows when the user calls it with reasonable arguments, ie a valid src and a dst with the exact size.

The kata description suggests that input will consist of 2 characters or more, so calling strlen(src) in this case makes sense.

It will overflow, if the dst size is set to exactly fit the trimmed string.

It is not a double array. It is array[key] construct, where the array content is 'silver' => 10, 'gold' => 5, null => 20 and the key that is accessed is the result of @$bands[3].

There is no heads-up about not using the built in function for array reversing, at least in the PHP version.

"Output should have 2 digits precision."
Currently, at least in PHP, that is a lie. Do not round the output, or you will fail the test.

Funny how I came up with almost the same strings.