Fork me on GitHub
#datomic
<
2019-08-15
>
tyler00:08:15

Should a fresh connection be retrieved for every request with datomic cloud or should you cache the connection?

tyler00:08:38

Hm that’s what I thought. Seeing something that looks like a memory leak though when retrieving a db connection and db value every request. Will dig into it more.

kenny01:08:51

The docs for as-of (https://docs.datomic.com/client-api/datomic.client.api.html#var-as-of) say: > Returns the value of the database as of some time-point. What is "time-point"? A Date? Epoch millis? Datomic's t?

kenny01:08:23

I'm assuming it's similar to Datomic peer: https://docs.datomic.com/on-prem/clojure/index.html#datomic.api/as-of > t can be a transaction number, transaction ID, or Date.

Brian Abbott06:08:24

Is anyone from Datomic Support currently available? We are having a critical outtage at the moment.

Brian Abbott07:08:42

Please see Cognitect Support Ticket 2327 🙂 Thank you!

jaret14:08:52

Hi Brian, I responded on the ticket a few hours ago.

Brian Abbott23:08:02

Thank you again so much Jaret!

mkvlr08:08:30

@jaret my colleague @kommen submitted the issue to <mailto:[email protected]|[email protected]>. Got a request to create our zendesk accounts but got running into a XSS error when trying to set my password on

jaret14:08:44

It looks like there were eventually able to create the ticket. Did they resolve the error? Was it transient or are they still having issues registering?

mkvlr14:08:27

just tried again, still doesn’t let me assign a password

mkvlr14:08:41

Blocked a frame with origin "" from accessing a frame with origin "". Protocols, domains, and ports must match.

jaret14:08:27

What browser are you using and do you use adblock (can you try without if so to confirm)?

grzm14:08:16

I have a transaction function that may return empty tx-data if certain criteria aren't met. I'd like to take a particular action only if there were datoms transacted. My understanding is that every transaction that does not fail will have at least a single datom in the tx-data of the result for the txInstant of the transaction. From what I've observed, I believe I can test whether the count of the datoms in the tx-data of the result is greater than one to determine if any additional datoms were asserted or retracted. Is this something I can rely on? Anyone have a better approach?

mgrbyte15:08:31

Assuming 1 datom transacted, you could also check if :db/txInstant is the only attr asserted in the "transaction entity". If you're not using reified transactions then this additional check is probably moot.

grzm16:08:39

Yeah, I'd like to save doing an additional database lookup: I guess I could assume that the :db/id of the :db/txInstant attribute is unchanging and not do a lookup, but inspecting the datom attribute db/id in a tx-data that includes only a single datom seems redundant if I know that each non-anomalous transaction result will have at minimum a single :db/txInstant attribute. Thanks for thinking through this with me.

tyler17:08:46

Is there a recommended way to hook up a java agent to a query group? Running into a strange memory issue with ions and we are having a hard time debugging with the datomic memory monitor on the provided dashboard.

Laverne Schrock20:08:12

We have one Datomic Cloud deployment running version 477-8741, and another deployment running 480-8772. On the older version we are able to make assertions about the entity with id 0, but in the newer version we cannot, due to the following error :

{:cognitect.anomalies/category :cognitect.anomalies/incorrect,
 :cognitect.anomalies/message "Boot datoms cannot be altered: <actual datom omitted, let me know if you need it>",
 :db/error :db.error/datom-cannot-be-altered}
This seems like reasonable behavior, but it doesn't seem to be documented in the changelog. Does anyone know if this is an expected change?