Clojurians Log v2

The reagent docs have this code example:

(defn lister [items]
  [:ul
   (for [item items]
     ^{:key item} [:li "Item " item])])

It seems like this is Reagent metadata?

It’s syntax for adding metadata to the vector being constructed

Reagent checks a few places for keys when converting a vector into a React element

One is props , so you could just pass in the key like any other prop

Another is metadata on the vector

to be clear, ^{:foo bar} is vanilla clojure, and puts a metadata map onto a collection or symbol, but reagent has a special interpretation of specific metadata keys

^:foo is often seen, it translates to ^{:foo true}

you can combine them

ins)cljs.user=> (def m ^:foo ^{:bar false} {})
#'cljs.user/m
(ins)cljs.user=> m
{}
(ins)cljs.user=> (meta m)
{:bar false, :foo true}

yeah I feel some of the docs are just outdated, thing is, when learning, it is hard to determine which thing is wrong

I want to use functional components by default because I use hooks a lot

The problem is that I use react-navigation and the screens need to be react components, so my root component needs to be reactified, it's a reagent component, then it uses navigators and screens from react-navigation that take react components as params

(defmacro defscreen
  [name component]
  `(def ~name
     (reagent.core/reactify-component
      (fn []
        @app.navigation/cnt
        [(fn []
           ~component)])
      app.navigation/functional-compiler)))

(defmacro defroutes
  [name navigator-creation-fn navigator]
  (let [screen-class (gensym "screen-class")
        navigator-class (gensym "navigator-class")]
    `(def ~name
       (let [[~navigator-class ~screen-class] (~navigator-creation-fn)]
         (reagent.core/reactify-component
          (fn []
            (~navigator ~navigator-class ~screen-class))
          app.navigation/functional-compiler)))))

Well, my guess would be that that specific one at least is not just outdated, it seems just plain wrong (as in: I have doubts Reagent ever behaved that way).

Ah that makes sense. Thanks @lilactown and @noisesmith

right, I meant, maybe it was correct at some point, otherwise it would not had been written (I want to think) haha

would be awesome to have some sort of build for the docs: if you use a piece of code as an example, should be extracted from a test suite or something

maybe once I learn the thing I should try doing something like that 🙂

cf. https://github.com/seancorfield/readme