Clojurians Log v2

re-implemented the schema visitor using walker: m/accept is now a postwalk and and there is new m/find-first to do lazy scanning for schemas. Though of doing a prewalk too, but coudn’t find any valid use case for such and would be more complex for schema extenders to add support for it. Not sure how useful the m/find-first is and should merge this: https://github.com/metosin/malli/pull/216

should the m/accept be just m/walk in the future? or m/post-walk?

(deftest find-first-test
  (let [schema [:map
                [:x int?]
                [:y [:vector [:tuple
                              [:maybe int?]
                              [:or [:and {:salaisuus "turvassa"} boolean?] int?]
                              [:schema {:salaisuus "vaarassa"} false?]]]]
                [:z [:string {:salaisuus "piilossa"}]]]]

    (let [walked-properties (atom [])]
      (is (= "turvassa" (m/find-first
                          schema
                          (fn [s _in _options]
                            (some->> s m/properties (swap! walked-properties conj))
                            (some-> s m/properties :salaisuus)))))
      (is (= [{:salaisuus "turvassa"}] @walked-properties)))

    (let [walked-properties (atom [])]
      (is (= "vaarassa" (m/find-first
                          schema
                          (fn [s _in _options]
                            (some->> s m/properties (swap! walked-properties conj))
                            (some-> s m/properties :salaisuus #{"vaarassa"})))))
      (is (= [{:salaisuus "turvassa"}
              {:salaisuus "vaarassa"}] @walked-properties)))))

Hi all! Is there a possibility to merge definitions declaratively - for example in an edn definition given that the definitions to merge are defined in a registry?

I see I can do

[:and
 [:ref :test/object]
 [:ref :test/annotated]
 [:map ....]]

@alpox nothing just now, but sounds like a good idea. Could you write an issue of that?

@ikitommi Done: https://github.com/metosin/malli/issues/217 thanks for the response!

what's the idiomatic way to specify a nonempty string in a schema? should I actually use a custom registry as in the README?

currently I'm doing this

(def nonempty-string
  [:and
   string?
   [:fn {:error/message "should be nonempty string"}
    (fn [x] (pos? (count x)))]])

one minor problem with that is that malli/explain gives me two error messages: "should be string" and "should be nonempty string". I tried to set the :error/message in the :and instead to get only one error, but that didn't work

(I'm using the malli version that reitit 0.5.2 requires, i.e. 0.0.1-20200525.162645-15, if that matters)

I'm happy to upgrade to any version that is compatible with reitit 0.5.2

@euccastro [:string {:min 1}], will need to release new version to work with reitit

thank you!

@ikitommi is there any chance you can release a new version of reitit with the latest malli?

another option for early adopters would be to release cutting-edge versions of malli under a separate namespace and artifact id (e.g., malli.pre-alpha) so we can use that for our own code without fear of breaking reitit. that would also reduce the pressure to update reitit unless new malli functionality is actually needed there

please ignore this if it sounds like too much of a hassle 🙂