:<- macro does not work across namespaces

[joshkh]

In the following example, using the :<- macro in the myapp.components.events namespace causes re-frame to throw an exception stating that the :main/some-items subscription handler cannot be found

(ns myapp.core
  (:require [myapp.events]
            [myapp.component.events]))

(ns myapp.events
  (:require [re-frame.core :refer [reg-sub]]))

(reg-sub
  :main/some-items
  (fn [db]
    (:all-items db)))

(ns myapp.component.events
  (:require [re-frame.core :refer [reg-sub]]))

(reg-sub
  :components/filtered-items
  :<- [:main/some-items]
  (fn [items]
    (filter odd? items)))

re-frame: no subscription handler registered for: " cljs.core.Keyword_hash: 1803312489cljs$lang$protocol_mask$partition0$: 2153775105cljs$lang$protocol_mask$partition1$: 4096fqn: "main/some-items"name: "some-items"ns: "main"proto: Object ". Returning a nil subscription.cljs.core.apply.cljs$core$IFn$_invoke$arity$2 @ core.cljs:3679
subs.cljc?rel=1472037082624:64Uncaught TypeError: Cannot read property 'call' of null

[stumitchell]

I would make sure that myapp.events is loaded before myapp.component.events to do this change the require in myapp.component.events to the following.

(ns myapp.component.events
  (:require [re-frame.core :refer [reg-sub]]
            [myapp.events]))

Let us know if that fixes it.

[mike-thompson-day8]

@joshkh Given the error message you posted, you really, really need to be using clj-devtools https://github.com/binaryage/cljs-devtools.

[mike-thompson-day8]

I'm about 100% sure Stu has answered your question. So I'm going to close this. Will reopen as necessary, but please direct support questions to clojurians.

[jamesnvc]

Something I'd like to add to this issue is that reg-sub is inconsistent as to when this will and will not work: Specifically, as in the example above, with one pair it fails, but if you have multiple pairs of :<- [:sub] it does work (since the implementation uses map in the multiple-sub cases, so the value isn't realized until later). It is very confusing to have something work and then stop working when removing subs - I submit it would be better if it either worked (preferable) or didn't consistently.

[danielcompton]

@jamesnvc that behaviour is fixed in #259, the behaviour of the single pair now matches the multiple pairs.

[jamesnvc]

@danielcompton perfect, thank you so much for the prompt reply!