| From: | Jacob Champion <jacob(dot)champion(at)enterprisedb(dot)com> |
|---|---|
| To: | Bruce Momjian <bruce(at)momjian(dot)us> |
| Cc: | Andreas Karlsson <andreas(at)proxel(dot)se>, PostgreSQL-development <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: First draft of PG 19 release notes |
| Date: | 2026-07-16 20:38:00 |
| Message-ID: | CAOYmi+k5h_w5p6HsVJF5k+bNGmChMAhpGU_1gjJMxHg080XibA@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, Jul 15, 2026 at 2:57 PM Bruce Momjian <bruce(at)momjian(dot)us> wrote:
> You can see the updated output here:
>
> https://momjian.us/pgsql_docs/release-19.html
Thanks! I think one of those patches partially undid the other one; see below.
> Let me know if there should be anymore changes?
I've definitely kept you waiting long enough. See attached suggestions
for my entries.
First patch adjusts credit, expands/rewords some items [1], and
rewrites the new authdata hook section to hopefully better illustrate
why it belongs to libpq. Let me know what you think.
Second patch should just be motion, I hope: the pg_hba changes go into
the server config section and the authdata hook goes into libpq. It
also deprioritizes the PQgetThreadLock() section, which I imagine very
few people will care about in practice. :D
I think this replays some of what Andreas was hoping for, maybe with
more context. I'm not sure why "Allow OAuth validators to supply
failure details" is still with server configuration at the moment,
based on the conversation above, so maybe that also needs to be moved
back to Source Code. I didn't do that here, in case it was on purpose.
(I don't know if it's been discussed somewhere, but I find that the
standard Git diff algorithm can't handle these sorts of doc-motion
patches in a readable way. The attached patches were generated with
`git format-patch --histogram` instead.)
Thanks,
--Jacob
[1] "OAuth" is the overall feature, `oauth` is the HBA method name,
and `OAUTHBEARER` is the developer-facing SASL mechanism name.
Unfortunately (or fortunately?) none of them are acronyms. This is
kind of analogous to the distinction we'd have to make between
"Kerberos", `gss`, and the `GSSAPI` SASL mech (if we used it).
| Attachment | Content-Type | Size |
|---|---|---|
| 0001-19-release-docs-reword-expand-credit.patch | application/octet-stream | 3.7 KB |
| 0002-19-release-docs-move-sections-around.patch | application/octet-stream | 5.2 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Sami Imseih | 2026-07-16 20:49:44 | Re: Report index currently being vacuumed in pg_stat_progress_vacuum |
| Previous Message | Melanie Plageman | 2026-07-16 19:59:12 | Re: pg19b1: TRAP: failed Assert("pgstat_bktype_io_stats_valid(bktype_shstats, MyBackendType)") |