Rasmus Porsager
Rasmus Porsager
Would any of you mind resolving the current conflicts ? :)
@uasan oh shoot, that's actually not the only issue. Same goes for the other inferred types (Date, Buffer, Bigint) :-( I should have caught this for the v3 update and...
It was really an assumption on my part wrt. handling of values after supporting the PostgreSQL feedback, but I guess you can say I might have caught it if I...
@AkeleyUA the issue is that we can't really change it without potentially breaking someone who relies on the implicit inference. It would require a new major, and I haven't thought...
I think It's useful when calling functions that don't do the same implicit type casts as regular queries. I'm pretty sure someones setup will break if changing it, and the...
Yeah I'm very curious too ☺️ I think it's not possible because the types specified from our end are supplied in the `Parse` message which comes before the `ParameterDescription` message....
I mean at least not in an efficient way, that would require us to call `Parse` first with unknown types, and then create a new extended query calling `Parse` again...
Right, but we can't supply types to `Bind`?
It's not different from what Postgres.js does and you're still not able to tell the db about alternative types after the fact?
No you can only decide in bind if a parameter is text or binary encoded. That's not related to postgres types. The types are static for the statement.