Conversation
Notices
-
@kaniini did we email the postgres list yet to lodge a complaint and request they patch/backport a change that doesn't cause data loss? I'll jump on that thread and voice support.
-
@rin @kaniini it will be a bug if we're loud enough.
-
@kaniini @rin please write to the list, i can get people to join in on it. they need to see that many people find this terrible and it's the opposite of what Postgres is supposed to be known for: protecting your data
-
@kaniini @rin also Kevin Grittner lives in town. If I have to I will find him and speak to him in person.
-
@mewmew @kaniini @rin eli5: there's a JSON function in PostgreSQL that instead of returning an error it wipes out all the data in that field
-
@mewmew @kaniini @rin there is a very hand-wavy "because you should know these specific semantics" explanation that is a poor excuse for the existing behavior
it's unbefitting of the postgres reputation. they need to fix it.
-
@mewmew @kaniini @rin you really shouldn't have to do this data validation / error checking in your application. that's what sql is for. it's a huge waste for us to have to query the existing data first, then decide if we should execute the originally planned query.
-
@kaniini @mewmew @rin i know, but every application using jsonb_set() on the planet will have to do this too after learning the hard way. it's pretty insane.
-
@kaniini @mewmew @rin yes, this is the right thing to do. provide stern constructive criticism and advice to everyone so they can save their asses when they google this
-
@kaniini @mewmew @rin i understand your reservations about getting involved in this discussion, but it helps if it's written by someone who understands the full breadth of technical details involved.
sure, I get what it's doing but if they want to challenge me on some specific technical minutiae I will fail to produce a coherent response :)
Ruby Rhod
All Bobinas P4G content and data are available under the