A valid time with a trailing newline (e.g. '12:30:00\n') should be added as a fixed test. Many solutions use regex patterns like /^([0-1]\d|2[0-3])(:[0-5]\d){2}$/, but the '$' token matches on different things depending on language, notably Python and Java, which will match the position before a terminal newline as well as the end of the string. Also, annoyingly, just replacing the '$' with '\Z' isn't always effective, as in Python, '\Z' matches only the end of the string, while in Java it's just a synonym for '$'.

Java translation!

JavaScript and Python use rounding + strict equality instead of approximate equality. (Ruby is OK)

Tests round answers to 6 decimal places.

should also be removed from the description

In Crystal the method name should be distance_between_points (snake case) not distanceBetweenPoints (camel case). Changing it will make it impossible to pass the final tests.

Haskell translation

Tests should use approximate equality instead of rounding + strict equality. Affected languages:

• JavaScript
• Ruby
• Python
• Crystal

In JavaScript, I intentionally used a truncated form of Pi in order to point out that an incorrect solution could pass the kata.

(not sure if this amounts to a legit issue but thought I would post it anyway, see what the community thinks)

Adding fixed tests / altering the RTG would solve this.

Groovy Translation

python: assert_approx_equals should be used.

On java full suit of tests asserts this input "0X1e5" but on kata instructions is only defined for prefix "0x" and not "0X". I hope it helps. Maybe that information should be included on specifications.

Where i can leave my code to ask about his workable ?

Error messages in C++ present values with insufficient precision: see the question below.

Hi Taw,

Thanks for the Kata. By reading the challange too complex it tought me a lot.

Just as an idea:
Build in a check that is the trainee creates a "x class" in the solution; to point them in the right way.

Enjoy