Try again now. I lower limit even more. There's probably a pic 9(3) that needs to be pic 9(4) somewhere in the test setup.

This translation is unmanageable. I've fixed the ref sol, but had to lower random test cases. If you want more random test cases, try to get in touch with original author of the translation on Discord.

I'll try to fix it, but I haven't touched COBOL in ages.

I can't see what is the language you have problems with? Seeing your code would also help.

I fixed random tests: Now it is enforced that X >= Y. Your solution should be able to pass all tests now.

This looks to be an issue in the kata.

Should be fixed. Please resubmit your solution.

Should be fixed in the current version, please check it.

Please give clear examples of failed tests to help fixing the issue.

Thanks for reporting the issue. It should be fixed now. When you are sure to have found a bug somewhere, don't hesitate to raise it like an issue, and please give an example of bad tests to facilitate the fix (almost all COBOL katas give explicit logs providing outputs and inputs).

You were right indeed. Thanks for reporting the bug. It should be fixed now.

I am the author of the translation. Can you give me an example of bad test logs?