I've rearranged it a couple of times, unfortunately it makes sense to me (obviously), are there any obvious reworks that would make it cleaner.

I've removed the 31st of February unnecessary statement.

I'll be cleaning up some of the assertions in the simpler tests to make them easier to understand (which should cover the code side of the suggestion)

Edit: I've done some rearranging and adding of headers. Hopefully should be a little clearer now?

In regards to the test assertions, do you have any suggestions to make it more readable?

The issue with changing the assertions is, I could update it to compare each line in the output, but in the case of issues in multiple fields it would not be clear (only the first failure would be rendered). It would also muddy the water as to the output format as it would look like a list more than a string output.

Edit: Have added specific formatting tests, and updated the tests that are expecting a single field to be outputted correctly to only assert that line. The random tests will still use the old compare everything assertion.

Could you be more specific in regards to which of Blind4Basics issues are coming up for you as well? Each of the issues has been marked as resolved so the same issues should not be possible.

both random choices should now differ.

I very cleverly left justified 15 and had a space after it...

I knew that one would come back to bite me... Will fix the range generator

ooops, fixed

Random tests added.

Working on the random crontab generator now.

Awesome-sauce.

When I run that test i get:

should handle stepped ranges

minute          1 3 5 7 9
hour            0
day of month    1
month           1
day of week     1

Edit: Potentially it hasn't noticed the republish for you yet?

Added a note for that.

Fixed the issue with stepped ranges, can you try it again.

Bug confirmed offline, tracking down the root cause now.

Edit: found the offending line, and can see what I did wrong, just gotta work out how to undo it.