Would it be considered ok to promote s/assert* to good for "public" use (atm the docstring says : "Do not call this directly, use 'assert'.")

in cases where you want the check to always happen no matter the value of check-assert/compile-asserts

or is there a better way? ping @alexmiller

I guess if it would be used that way it could warrant a rename maybe

I seem to remember that I could generate with a provided example. Something like (gen/fill-in ::spec {::id "all generated will have this id, other values are generated"})

Can't remember what it was called though... anyone?

@andreas862

(s/def ::id string?)
(s/def ::foo (s/keys :req [::id]))
(s/exercise ::foo 2 {::id #(s/gen #{"a" "b"})})
;=>
;([{::id "a"} {::id "a"}]
; [{::id "b"} {::id "b"}])

@misha yes, thank you!

(defn gen
  "... Optionally an overrides map can be provided which
  should map spec names or paths (vectors of keywords) to no-arg
  generator-creating fns. These will be used instead of the generators at those
  names/paths. Note that parent generator (in the spec or overrides
  map) will supersede those of any subtrees. 

but it forces me to keep generators and specs separate, which ends up to be cleaner

I have some function

(defn foo
  [a]
  {:pre [(s/valid? ::bar a)]}
...)

I just get an AssertionError

@pablore I use s/explain-data instead of s/valid. it returns nil if everything is ok, and a datastructure, if there are any errors.

So you can use it in if or when.

read through this too, may be it'll give you some ideas https://clojure.org/guides/spec#_using_spec_for_validation afaik, specs are replacement of :pre/`:post`-conditions

s/explain works better than I expected! I know of s/fdef but I’m writing a really complex reducing function with lot of if-cases.

(s/check-asserts true)
(s/assert string? :a)
;; clojure.lang.ExceptionInfo: Spec assertion failed
;;    val: :a fails predicate: :clojure.spec.alpha/unknown