Message-ID: From: "vlsi (@vlsi)" To: "pgjdbc/pgjdbc" Date: Thu, 14 May 2026 06:58:09 +0000 Subject: Re: [pgjdbc/pgjdbc] PR #3062: feat: type cache rework, codec API, and composite-type round-trip In-Reply-To: References: List-Id: X-GitHub-Author-Login: vlsi X-GitHub-Comment-Id: 4448418808 X-GitHub-Comment-Type: issue_comment X-GitHub-Issue: 3062 X-GitHub-Repo: pgjdbc/pgjdbc X-GitHub-Type: comment X-GitHub-Url: https://github.com/pgjdbc/pgjdbc/pull/3062#issuecomment-4448418808 Content-Type: text/plain; charset=utf-8 @davecramer , @sehrope , @bokken , The PR seems to add full Struct / Array support, proper type resolution (name -> oid), and SQLData support. It looks like this goes in a pretty good shape modulo remaining cleanups 🎉 🎉 🎉 I still plan to review the changes after `docs: PGTYPE_SPEC overview` (those changes are "test failure fixes", "`@Nullable` fixes", etc). The commits up to "docs: PGTYPE_SPEC overview" are coherently structured (I asked claude to restructure WIP commits), and the rest are fixups to be reviewed (and later squashed into the proper commits). I plan to drop something like "hey, codex, the PR claims to support Struct/ Array / SQLData, so go ahead and generate user-level tests for it using JDBC / Hibernate / jOOQ / Spring Data ...". I'm not sure if it would identify meaningful issues, however, it could surface some. I need to double-check `CallableStatements` as well (I haven't paid much attention to them). I haven't paid attention to "structs/composites in updateable resultsets". It is yet another todo. /cc @lukaseder. It would be nice to hear from you as well. Hopefully this should ease struct/array/binary handling with pgjdbc. PS. To everybody: this is almost fully machine-generated code. I did review it, and I steered it. Sorry for issuing such a big change, however, I think it is ready for the general review. If you feel "oh, slop code detected", please ping me. `PGTYPE_SPEC.md` is a part of the initial plan (a lot of iterations between claude and codex), and I intend to remove it.