Kata

The link to the 3th chapter of this series points to the 1st one.

missing fixed test: something like the following:

Start states [ 8 ]
Transitions [
  [ 0, 'b', 6 ], [ 0, 'b', 8 ],
  [ 1, 'a', 5 ], [ 1, 'b', 9 ],
  [ 2, 'b', 4 ], [ 2, 'b', 2 ],
  [ 3, 'b', 9 ], [ 3, 'b', 0 ],
  [ 4, 'a', 5 ], [ 4, 'a', 8 ],
  [ 5, 'b', 6 ], [ 5, 'a', 8 ],
  [ 6, 'a', 0 ], [ 6, 'b', 0 ],
  [ 7, 'b', 1 ], [ 7, 'a', 4 ],
  [ 8, 'a', 7 ], [ 8, 'b', 7 ],
  [ 9, 'b', 2 ], [ 9, 'a', 3 ],
  [ 1, '', 3 ],  [ 9, '', 3 ],
  [ 4, '', 9 ],  [ 3, '', 8 ],
  [ 3, '', 1 ],  [ 1, '', 3 ],
  [ 2, '', 8 ],  [ 3, '', 3 ]
]
Aceppt states [ 0 ]
input:     aa

my solution fails often, but not on every run of the random tests (note: I don't know the expected result there, because my code ends up in a recursion error)

Hi,

• In the test suite, you should avoid to print actual info to the console. For instance, the logs you do in the random tests could actually be in the message of the it block. Well, I guess there is too much data to write in there... Maybe just remove the logs. If the user needs the infos, he can print them on his side. On the other hand, you could be a bit more precise about what's going on in the random tests, changing the message of the it block to something like X tests on a new NFA or something like that.
• It would be better to separate the concerns in the random tests: so for you redefine the whole logic of the solution inside the loop defining the it blocks of the random tests. 'Would be better to have a separate function (defined in the describe block, but not right inside the for loop) that you call with the appropriate arguments. This way, the tests will be easier to maintain by others.

Hi again,

• at least in the sample tests: actual and expected are swapped.
• description isn't better than the first one (except for the fact that now, I already know what it is about, sort of...)

cheers

See question below. Why this many tests? Is this an extreme performance kata?

How many tests do you have? I'm timing out after 123,000 tests. Perhaps you should add a "performance" tag.