Vladimir Sitnikov

Frankly speaking, I can't easily test with PG 8.4, however, I guess the key is `simple query mode` which you seem to miss. Try adding `preferQueryMode=simple`: https://github.com/pgjdbc/pgjdbc/blob/43e449ec4eb24132c0d6bc7b92fc68443b52c708/.github/workflows/main.yml#L177-L185 However, I do...

An alternative option would be excluding "8.4+simple query" combo from the test suite. E.g. `matrix.exclude({query_mode: {value: 'simple'}, pg_version: '8.4'});` near https://github.com/pgjdbc/pgjdbc/blob/43e449ec4eb24132c0d6bc7b92fc68443b52c708/.github/workflows/matrix.js#L203-L209

We can try inventing a better "prepared statement batching". 1. JDBC spec requires that batch is executed only with `executeBatch` call. In other words, the spec does not allow sending...

@panchenko , technically speaking, I would agree that current memory consumption is not predictable. I do agree it might be painful to recreate 8k arrays again and again. So should...

@davecramer , when using a connection pool you never know which connection becomes a victim. I guess with bigger values caching the buffer would not help much, however large cached...

> it would need to use the privileged user Use of privileged users is fine for unit tests, see `TestUtil.openPrivilegedDB`

The tests were not executed as a part of test suite. We tend to register all test classes into test suite: https://github.com/pgjdbc/pgjdbc/blob/89fe182463b60511c94766115f6cdbdae36e9539/pgjdbc/src/test/java/org/postgresql/test/jdbc2/Jdbc2TestSuite.java This is the way your test class will...

> since it only gives correct results when the Offset is UTC. @alexanderkjall , can you please clarify what you mean by "only"? Could you provide a case when it...

This is a "feature added", thus it can't fit into 42.2.2. I suggest closing 42.2.2 (review the remaining PR, release it) before we continue discussion on 42.3.0. Does it make...

There's one PR left https://github.com/pgjdbc/pgjdbc/milestone/16 , plus release notes prepare&review.