reg-query-sub registered subscription is not returning latest value

[jleonard-r7]

from REPL (after a change in the UI from 24h to 1w):

state.core=> @(rf/subscribe [::timerange])
{:timerange "24h"}
state.core=> (d/q '[:find ?t . :where [:timerange :timerange/timerange ?t]] @connection)
{:timerange "1w"}

sub is defined as:

(rp/reg-query-sub
   ::timerange
   '[:find ?t . :where [:timerange :timerange/timerange ?t]])

and the event which sets the new value is defined as:

(rp/reg-event-ds
   ::set-timerange-from-query-param
   [rf/trim-v]
   (fn-traced [_ [timerange]]
              [[:db/add :timerange :timerange/timerange (timerange/query-string-param->timerange timerange)]]))

[jleonard-r7]

note also that i am using a listener to write the data to local storage:

(d/listen! connection :persistence
             (fn [tx-report]
               (when-let [db  (:db-after tx-report)]
                 (rf/dispatch [:state.local-storage/set-item storage-key (db->string db)]))))

and local storage itself contains the correct value.

[denistakeda]

Hello @jleonard-r7 . Thank you for reporting this. I'll have a look

[jleonard-r7]

Here is a minimal repro:
jleonard-r7@ec9351b

Just cd into that example and run:
clj -A:repl

and then in the browser that appears, click the "Click Me" button.

[denistakeda]

Thank you @jleonard-r7 . Will check this out

[denistakeda]

Hello @jleonard-r7 . I've checked the example and indeed, it works not as expected. So-far posh library is unable to watch the entities based on keyword and :db/ident rule. I'm going to support that but so far you have two workarounds:

1. Use predefined integer ids instead
2. Use the data storage for singleton values

Thank you for the reporting and provided repo!

[jleonard-r7]

Would you mind briefly mentioning the planned fix for this and whether it would also address #47 ?

I'm beginning to hit these caching issues all across my codebase to the point that re-posh is not really usable in it current state (and that's after introducing a top-level :root object as a workaround to this, #41, the first caching issue I discovered).

If you could briefly tell me the planned fix, I could carry it out myself. Otherwise, I will unfortunately probably have to refrain from using re-posh for the time being.

Many thanks in advance for any guidance you can offer.