This is my favourirte Kata Series so far. Many thanks to the Sensei and all contributors. Cheers :)

Let me know if you find any issues with this kata. Feed backs are welcome :)

Error messages should show inputs. (Expected result is already shown by assert_equals, messages like "It should be X: Y should equal X" are redundant.)

Examples:

Sample tests and small fixed/random tests:

fixed_partitions(4, 3): None should equal {(1, 1, 2)}

Large tests:

fixed_partitions(20, 55): wrong result

this also seems fixed in current draft

timing seems to be fixed consistently

Time: 617ms Passed: 48Failed: 0

Also, you would like all small tests to run independently, regardless on errors in some of the tests, while you want to exit early on first error for large tests. You can decide how to handle medium tests, one way or the other.

I see, this is true. I have unpublished it. Thanks for your feedback.

No, tests still expect a mix of (set,length), 0.

There is stll to much variability.

Times of an empty solution (return 0):

time out
time out
time out
 4135 ms
 1567 ms
time out
 3792 ms
 5278 ms
 8654 ms
 7208 ms
10730 ms
12021 ms
11668 ms
 1612 ms
 5453 ms
time out

• Reference solution is too slow

• Random tests shold bo grouped in batches of small (debugging, with visible solution), medium (correctness), large (performance). Performance tests should have low time variability.

I have implemented it now. Thanks, once again, I have learnt something new.

Thank you for pointing this out. I will try to effect these changes now.

Returning different types (tuple/int) is bad practice.

Returning a set and its length is redundant.

The function should return a set (an empty set if there aren't tuples). If you want a tuple with length (not recomended), the function should always return a tuple ((set(), 0) if no tuples).

I have worked on the test cases to prevent this. Thanks for pointing this out. Let me know if the problem persists or if there are other issues.

Hello! Thanks for pointing this out, I have edited the test code to not print the solutions to the terminal to prevent this. Let me know if there is any other issue.

Memory overload in error feedback

Max Buffer Size Reached (1.5 MiB)