Valentin Waeselynck

Was just bitten by this as well... Here's another repro: ```clojure (def db (dts/empty-db {:my_id {:db/unique :db.unique/identity} :my_ref {:db/valueType :db.type/ref :db/cardinality :db.cardinality/many}})) (dts/with db [{:my_other_attr "hello" :my_ref [{:my_id 2 :my_ref...

I don't know enough about the implementation details of DataScript, but to me it seems more a matter of making sure the entities unify correctly based on unique attributes. (That...

> Another strategy is to have DB change its id after each transaction, effectively make all entity equal checks to return false (pessimistic case: if DB changed in any way,...

The main downside I see is that it will become harder to consume events for users. It might be worth it (I'm not debating that here), but if we do...

> @vvvvalvalval > > > ... harder to consume events for users. > > By this, do you mean harder to read for programmers? Harder to understand? Or some other...

@mike-thompson-day8 I've got several remarks, I'll make one comment for each. This first remark only expresses agreement, and provides supporting arguments. ## 1. Effects are better conj'd than assoc'd I...

## 2. Please do go with tuples, not maps @mike-thompson-day8 I concur that effects under `:fx` should be represented as tuples, not maps. I hope to get you to reconsider...

> sorry for offtopic, feels like the next step is to use datascript as app-db Already doing that by embedding DataScript in `app-db`. It's great. But it's good to complement...

## 3. About the `:db` effect: I recommend disallowing `:db` under `:fx` (at least initially) @mike-thompson-day8 Here I make a case for not allowing `:db` effects under the new `:fx`...

> I don't find the (future) need for metadata argument that compelling - too speculative. Perhaps that's a lack of imagination on my part. @mike-thompson-day8 Here's a random idea, still...