i added a skeleton for the custom error type in the initial code and an explicit import of it in the tests. also added sample tests for exceptions. adding the same edge twice is not an edge case in this kata, as the description says:

Parallel edges and self-loops are permitted

This date strigified format is valid, but not the Date itself. And issue about description ambiguity has been raised above and below, closing...

The point of this cata is regexp practice and understanding Gregorian calendar.

Fortran Translation Kumited - please accept :D

Agreed.

Added test to all languages and updated description.

That's correct! See my fork that fixes the issue.

Also, this should be specified in the description and reenforced in the tests.

Cheers

nice one

simplicity can cause trouble IMO

besides what's the harm of just making it clear

i dont agree with the top solution.
If you got an r = 0 as your param in the random test like I do, that solution fail.

I suggest adding that into the sample test.
like arithmetic_sequence_elements(1, 0, 5)
I am pretty sure the top one failed on this

i dont agree with this solution.
If you got an r = 0 in your param like I do, this solution fail

Yes I agree on this.
Better bearing performance in mind

you can do this in one line by datetime.datetime.strptime("","%d-%m-%Y")