(rf/reg-event-db
  ::toggle-starred
  (fn-traced [db [_ person-id]]
    (let [val (get-in db [:tracker :person-list :people person-id :starred])
          x (util/console-log val)]
      (assoc-in db [:tracker :person-list :people person-id :starred] (not val))
      )))

I run it first time, it's null

second time: true

third time: false

every other time: still false

The subscription and app-db in 10x never updates

if I remove the last key just to see what the get-in is returning, its just the starred property even if I never specify it -_-

I have code ghosts

To weed out the obvious thing - are you 100% sure person-id is always the same one and is always correct? Do you have any global interceptors registered?

Also, just in case - no need for get-in + assoc-in when you have update-in.

(update-in db [...] not)

Question: any particular reason it’s not

(rf/reg-event-db ::toggle-starred
  (fn-traced [db [_ person-id]]
    (update-in db [:tracker :person-list :people person-id :starred] not)))

Beat me to it

No reason in particular! 🙂 I started with that and broke it out to try and see what was failing but let me try it again that way

It won't change anything, the problem is elsewhere.

If you're sure about person-id and you're sure about there not being any global interceptors that might influence that value, then I'd try to come up with a minimal reproducible example. Otherwise, it's just crystal gazing.

Understood, there are no global incerceptors

Right, so now the event itself is not throwing any console errors, but the app-db inspector does not show the value updating

I can use the path inspector to get to :tracker :person-list :people 32 :starred]

32 is the person-id

(hand coded sample data)

fn-traced is something that doesn't change the result, right?

I was under the impression it was just to support re-frame 10x event tracing

I try not to assume too much when the namespace declaration is not provided. :)

I mean, wise 🙂

(ns pnx-web.events
  (:require
    [re-frame.core :as rf]
    [re-pressed.core :as rp]
    [pnx-web.db :as db]
    [day8.re-frame.tracing :refer-macros [fn-traced]]
    [pnx-web.util.core :as util]))

If it's a small public project available somewhere along with instructions or an obvious way to run it, I can take a look.

It is none of those things, unfortunately. 😕 It appears the db is not getting updated and nothing is obviously erroring, Am getting some warnings about seqs not having keys, but I don't think that would manifest this way

The part that gets me about this is you saying the first time you run it it’s still null

Oh right, that’s expected, since it’s logging before you’ve changed the value

ok yeah that makes sense

How are you reading the value? Through simple sub or through tooling?

(Or just through what the bit of debug outputs?)

10x, but Im messing with it in repl now and making some headway

I think one of the items in the db. is in an extra vector I didn't mean for it to be in and that's confounding some things

my people list is a vector of maps that are keyed by the id.. trying to use the indexed entity pattern

Oh maybe its the event is firing with strings, and the keys are numbers

That seems like its a problem

Which was the very first question I asked. :D

32 and "32", looks the same to me 😳 😞

Happens.

My hubris leads me to believe the problem is always more complicated than it ends up being! Learning clj is a humbling experience for sure, haha!

Thanks as always!

annnd of course it works now! 🙂

Clojure has different print statements. print and println produce output for human consumption. I learned early on not to use that ^^^, for this very reason. Better:

By default, pr and prn print in a way that objects
can be read by the reader

And FWIW, when in the browser, I always use js/console.log. "1" and 1 will be printed slightly differently, enough for me to notice, but there will be a benefit of being able to interactively study the printed objects, both CLJS and JS ones.

Thats great advice, I need to use console.log and println 🙂

I think kennytilton meant the opposite - not using println and instead using prn because it clearly shows the actual type.