Every test reuses the same input object, so modifying the input will break the fixed tests, which would be very confusing.

I've deep-frozen the configuration objects in a new version of this kata. Not sure if it's smart to mutate function arguments in the first place though :P

Also, transitions in the random tests can be modified.

They too are now deep-frozen.

Btw, you're replying to an issue I made earlier and then deleted, so the issue being raised here is completely different (don't ask me why CW behaves like this, it's weird like that). This is actually about a separate issue on the input.

Is there something in the phrasing of the description that could be improved?

Yes, the actual problem here is that we never actually see the input corresponding to the FSM for the sample tests, not even in the test code (since it's defined in Preloaded), and it was never mentioned what FSM is being tested there (which is the example in the description). So the expected behaviour is pretty much undecipherable before one runs the test once and logs the input.

Every test reuses the same input object, so modifying the input will break the fixed tests, which would be very confusing.

Also, transitions in the random tests can be modified.

It's not specified whether multiple source states could have a transition by the same name, possibly to different target states.

@schsam woah, that's pretty cool! I guess I should I learn it.

Agreed

The test cases should tell what values they expect instead of just simply stating "[format] to [format] failed"

The description should say what entries we may receive (in other words: define what 'convert to integer makes no sense' means; how am I supposed to guess in advance how to handle a single array with a single number? and what if two numbers or more ?, etc.). One needs to guess uncertainly from test cases.

Needs random tests.

assertEquals should be used.

Sorry for not actually taking a look at the kata in detail, however browser compatibility problems are very unlikely here. All tests are run on the server in a pretty well-defined environment.

For details the kata author would need to chime in...