@U05254DQM Is this a deliberate bug to fall out of testing later? https://github.com/practicalli/leveraging-spec/blob/master/src/practicalli/bank_account_design_journal.clj#L56

:args should be a sequence spec, using s/cat: https://github.com/practicalli/leveraging-spec/blob/master/src/practicalli/bank_account_design_journal.clj#L422

Do you understand why this doesn't work? https://github.com/practicalli/leveraging-spec/blob/master/src/practicalli/bank_account_design_journal.clj#L337-L344

s/or requires labels for the alternatives: (s/or :qualified (s/keys :req [...]) :unqualified (s/keys :req-un [...]))

Oh yes, labels. At 2.30am I tend to forget things :white_frowning_face: I find fdef a bit opaque, some more reading and experimenting to do. Thanks for spotting the bug in last-name, not intentional. Thanks for the review

I got the spec/or expression to wrap the :req and :req-un versions working. I have my fdef :args validating after instrumenting. Still working on testing the :ret in the fdef...

Right, :args is for instrument -- checking that your code is calling the function correctly -- and :ret/`:fn` are for generative testing, to make sure your function behaves correctly.

With instrument, :ret and :fn are ignored. With check, :args is used to generate conforming arguments and then the function is called and the :ret and :fn specs are checked against the return value and against the arguments/return value respectively.

A lot of people don't feel the docs are clear enough about :args/`instrument` and :ret/`:fn`/`check`

I’m comfortable with instrument for now. I don’t seem to have something right with check yet, but will look again tomorrow. Thanks.