2024-11-25 re-frame | Clojure Slack Archive

re-frame

Vladimir Pouzanov 2024-11-25T16:53:19.198389Z

hi o/ I'm getting strange issues with using re-frame and reagent form-2 components and they seem to be falling apart in pretty random ways. Just checking if I'm doing something stupid. Tried this:

(defn login-name []
  (let [!input-name (atom nil)
        send-code #(rf/dispatch [:login/send-otp-code (-> @!input-name .-value)])
        loading? @(rf/subscribe [:login/loading?])
        error @(rf/subscribe [:login/error])]
    (fn []
      [...])))

and the state gets widely desynchronized. This however, seems to always work:

(defn login-name []
  (let [!input-name (atom nil)
        send-code #(rf/dispatch [:login/send-otp-code (-> @!input-name .-value)])]
    (fn []
      (let [loading? @(rf/subscribe [:login/loading?])
            error @(rf/subscribe [:login/error])]
        [...]))))

The first approach seems to work sometimes if I move the deref into the function. But again, is wobbly. Is that something I shouldn't do? All examples I've found use form-1 components, but I need the :ref binding in this component and it seems wrong to run that one through re-frame.

✅ 1

p-himik 2024-11-25T16:55:43.170339Z

Don't deref the subs in the outer fn, deref them in the inner fn. So definitely the second code block. So you can either go with the second code block (something I would do myself), or you can subscribe in that outer let and @ in the inner let.

Vladimir Pouzanov 2024-11-25T16:57:01.947199Z

thanks. I tried @ in the inner lets and got strange results and never figured where my state falls apart and the second version always works so I'm happy with a double-let. Was just having a sanity-check because I haven't seen people do that 🙂

p-himik 2024-11-25T16:57:59.356529Z

The only difference between subscribing in the outer fn and in the inner fn might come from passing arguments to subscriptions. But your code samples don't do that.

Clojurians Log v2

re-frame