Vladimir Sitnikov

@rlbdv , the analysis sounds quite right, however, there's an issue with the current way of handling literals. The important point is that `org.postgresql.core.v3.SimpleParameterList#toString(int, org.postgresql.core.v3.SqlSerializationContext)` should not fall into SQL-injection...

Here's my attempt to optimize it via ChatGPT: https://chatgpt.com/share/689a3548-60f4-800f-bab8-6550415687c4 That was "5 thinking" model, and I still had to ask it multiple times to remove allocations. At the same time,...

>but didn't see how we might handle a prepared statement parameter that's an array of bytea values without "bytea" PGobjects, offhand. @rlbdv , have you tried `conn.createArrayOf("bytea", byte[][])`? See *...

The update is blocked by https://github.com/lburgazzoli/gradle-karaf-plugin/issues/88

I am afraid it might result in logging lots of extra log messages (the code would generate a message for every execution), and it will be useless if the query...

Can we somehow test this?

What do you mean by "properly"? Frankly, I would suggest to use `java.time` types instead of `java.sql` types. Is there a reason you have to use `java.sql` types without casts?

`Oid.DATE`, `Oid.TIMESTAMP`, and `Oid.TIMESTAMPTZ` differ. At the same time, it is hard to tell if `.setTimestamp` means `TIMESTAMP` or `TIMESTAMPTZ`. I'm not sure if it is the only reason for...

> So basically, this is a known behaviour of the driver and there is no way to improve it? I'm afraid it would be extremely hard to improve the behavior...

Have you considered using a tested dependency rather than reinventing the cleaner? (e.g. see https://github.com/juanplopes/lazycleaner ). --- I would suggest updating the API so the user could call something like...