add list of major features to the v19 release notes
Hackorum builds and tests every patch posted to the lists, not only commitfest submissions. This is Hackorum's own CI rather than the PostgreSQL project's, and it is still under testing - please report anything that looks wrong.
Here's a first try at $subject. I used the 19beta1 announcement as a
guide.
--
nathan
Attachments:
v1-0001-Add-list-of-major-features-to-the-v19-release-not.patchtext/plain; charset=us-asciiDownload+78-2
Hi,
yay for having this in place earlier this time.
Some comments:
On Wed, Jul 01, 2026 at 04:38:14PM -0500, Nathan Bossart wrote:
+ <listitem> + <para> + Logical replication now + <link linkend="logical-replication-sequences">replicates sequence values</link>, + and it can be enabled without a server restart when + <xref linkend="guc-wal-level"/> is set to <literal>replica</literal>. + </para> + </listitem>
Maybe it's because I'm not a native speaker, but I tripped over the
second part and parsed it as a function of the replicate sequence
values, not logical replication. Would it help readers to drop the "it",
i.e. write "and can be enabled [...]"?
+ <listitem> + <para> + Faster performance in many areas, including automatic scaling of the + number of <link linkend="guc-io-max-workers">I/O worker processes</link>, + quicker foreign-key checks, and further planning and execution + optimizations. + </para> + </listitem>
(This is kinda a catch-all statement and might go well as last item,
dunno).
+ <listitem> + <para> + Data checksums can now be + <link linkend="checksums-online-enable-disable">enabled or disabled while the database is running</link>. + </para> + </listitem>
I think we still don't say "the database" when we talk about a
server/instance/cluster. It should either be "the cluster" or "the
database server".
+ <listitem> + <para> + <xref linkend="sql-update"/> and <xref linkend="sql-delete"/> can now + change or remove data for just part of a time range via the new + <literal>FOR PORTION OF</literal> clause. + </para> + </listitem>
I think this sentence should include something about temporal
ranges/keys, or maybe just s/time range/temporal range/?
Michael
On Thu, Jul 02, 2026 at 01:15:39PM +0200, Michael Banck wrote:
On Wed, Jul 01, 2026 at 04:38:14PM -0500, Nathan Bossart wrote:
+ <listitem> + <para> + Logical replication now + <link linkend="logical-replication-sequences">replicates sequence values</link>, + and it can be enabled without a server restart when + <xref linkend="guc-wal-level"/> is set to <literal>replica</literal>. + </para> + </listitem>Maybe it's because I'm not a native speaker, but I tripped over the
second part and parsed it as a function of the replicate sequence
values, not logical replication. Would it help readers to drop the "it",
i.e. write "and can be enabled [...]"?
Fixed.
+ <listitem> + <para> + Faster performance in many areas, including automatic scaling of the + number of <link linkend="guc-io-max-workers">I/O worker processes</link>, + quicker foreign-key checks, and further planning and execution + optimizations. + </para> + </listitem>(This is kinda a catch-all statement and might go well as last item,
dunno).
Moved.
+ <listitem> + <para> + Data checksums can now be + <link linkend="checksums-online-enable-disable">enabled or disabled while the database is running</link>. + </para> + </listitem>I think we still don't say "the database" when we talk about a
server/instance/cluster. It should either be "the cluster" or "the
database server".
Fixed.
+ <listitem> + <para> + <xref linkend="sql-update"/> and <xref linkend="sql-delete"/> can now + change or remove data for just part of a time range via the new + <literal>FOR PORTION OF</literal> clause. + </para> + </listitem>I think this sentence should include something about temporal
ranges/keys, or maybe just s/time range/temporal range/?
How about something like this?
Support for temporal updates and deletes via the new FOR PORTION OF
clause.
--
nathan
Attachments:
v2-0001-Add-list-of-major-features-to-the-v19-release-not.patchtext/plain; charset=us-asciiDownload+78-2
Import Notes
Reply to msg id not found: 6a46485d.2a10144b.ad66e.cd2bSMTPIN_ADDED_BROKEN@mx.google.com
On Thu, Jul 02, 2026 at 09:36:46AM -0500, Nathan Bossart wrote:
Subject: [PATCH v2 1/1] Add list of major features to the v19 release notes.
Barring more feedback, I'd like to proceed with this patch sooner than
later.
--
nathan
On Tue, Jul 28, 2026 at 10:39:19AM -0500, Nathan Bossart wrote:
On Thu, Jul 02, 2026 at 09:36:46AM -0500, Nathan Bossart wrote:
Subject: [PATCH v2 1/1] Add list of major features to the v19 release notes.
Barring more feedback, I'd like to proceed with this patch sooner than
later.
Sure, I am usually not involved with the major feature list.
--
Bruce Momjian <bruce@momjian.us> https://momjian.us
EDB https://enterprisedb.com
Do not let urgent matters crowd out time for investment in the future.
On 7/31/26 3:01 PM, Nathan Bossart wrote:
Committed.
Thanks for drafting this!
I think there's a notable omission, specifically around the `ALTER TABLE
... MERGE/SPLIT PARTITIONS` as this has historically been a challenging
operation for folks. I added proposed copy here (which is mostly a
rewording of the release note).
This has been up for awhile, so I don't want to suggest removing items
for this list. That said, I'd like to suggest a few
listing/wording/ordering tweaks, primarily to tighten up a few of the
explanations and align with the style from previous release notes.
Thanks,
Jonathan
Attachments:
v2-doc-release-19-major.difftext/plain; charset=UTF-8; name=v2-doc-release-19-major.diffDownload+34-30
On 2026-08-22 Sa 3:56 PM, Jonathan S. Katz wrote:
On 7/31/26 3:01 PM, Nathan Bossart wrote:
Committed.
Thanks for drafting this!
I think there's a notable omission, specifically around the `ALTER
TABLE ... MERGE/SPLIT PARTITIONS` as this has historically been a
challenging operation for folks. I added proposed copy here (which is
mostly a rewording of the release note).This has been up for awhile, so I don't want to suggest removing items
for this list. That said, I'd like to suggest a few
listing/wording/ordering tweaks, primarily to tighten up a few of the
explanations and align with the style from previous release notes.
I don't know how the ordering was arrived at. I think we should lead
with blockbuster items. Everyone will have their own opinion, but in my
mind the blockbuster items are REPACK and pg_plan_advice.
cheers
andrew
--
Andrew Dunstan
EDB: https://www.enterprisedb.com
On Sat, Aug 22, 2026 at 05:50:56PM -0400, Andrew Dunstan wrote:
On 2026-08-22 Sa 3:56 PM, Jonathan S. Katz wrote:
This has been up for awhile, so I don't want to suggest removing items
for this list. That said, I'd like to suggest a few
listing/wording/ordering tweaks, primarily to tighten up a few of the
explanations and align with the style from previous release notes.I don't know how the ordering was arrived at. I think we should lead with
blockbuster items. Everyone will have their own opinion, but in my mind the
blockbuster items are REPACK and pg_plan_advice.
There wasn't a whole lot of science around the ordering. We did
intentionally put the catch-all performance item at the end, but that was
about the extent of conscious thought about it.
Jonathan's proposed changes seem fine to me at a glance. I'll let that
marinate on the lists a bit more before applying.
--
nathan
On 8/25/26 10:30 AM, Nathan Bossart wrote:
On Sat, Aug 22, 2026 at 05:50:56PM -0400, Andrew Dunstan wrote:
On 2026-08-22 Sa 3:56 PM, Jonathan S. Katz wrote:
This has been up for awhile, so I don't want to suggest removing items
for this list. That said, I'd like to suggest a few
listing/wording/ordering tweaks, primarily to tighten up a few of the
explanations and align with the style from previous release notes.I don't know how the ordering was arrived at. I think we should lead with
blockbuster items. Everyone will have their own opinion, but in my mind the
blockbuster items are REPACK and pg_plan_advice.
"Blockbuster" means different things to different people - when I
propose ordering, I'm typically looking for an optimization around how
many users does the change impact and how resounding that impact is.
That said, in the proposed patch I did shuffle the order around that
algorithm - if needed I'm happy to detail the thinking more, but I like
to be careful around this given people have put a lot of work and effort
into all the patches and did work specifically because they want to make
things better/easier for users.
There wasn't a whole lot of science around the ordering. We did
intentionally put the catch-all performance item at the end, but that was
about the extent of conscious thought about it.Jonathan's proposed changes seem fine to me at a glance. I'll let that
marinate on the lists a bit more before applying.
Thanks; happy to make additional tweaks if need be.
Thanks,
Jonathan
On Tue, Aug 25, 2026 at 09:52:32AM -0600, Jonathan S. Katz wrote:
On 8/25/26 10:30 AM, Nathan Bossart wrote:
Jonathan's proposed changes seem fine to me at a glance. I'll let that
marinate on the lists a bit more before applying.Thanks; happy to make additional tweaks if need be.
A number of things have been (or are being) reverted, so this will need to
be reworked. We might want to bump one or two features into the list while
we're at it.
--
nathan
On Tue, Sep 15, 2026 at 07:05:12AM -0500, Nathan Bossart wrote:
On Tue, Aug 25, 2026 at 09:52:32AM -0600, Jonathan S. Katz wrote:
On 8/25/26 10:30 AM, Nathan Bossart wrote:
Jonathan's proposed changes seem fine to me at a glance. I'll let that
marinate on the lists a bit more before applying.Thanks; happy to make additional tweaks if need be.
A number of things have been (or are being) reverted, so this will need to
be reworked. We might want to bump one or two features into the list while
we're at it.
FYI, I have not modified that list since I am not the owner of that part
of the doc. However, I didn't see anything that needs removal so I
assume others have updated it.
--
Bruce Momjian <bruce@momjian.us> https://momjian.us
EDB https://enterprisedb.com
Do not let urgent matters crowd out time for investment in the future.
On Tue, Sep 15, 2026 at 09:24:08AM -0400, Bruce Momjian wrote:
On Tue, Sep 15, 2026 at 07:05:12AM -0500, Nathan Bossart wrote:
On Tue, Aug 25, 2026 at 09:52:32AM -0600, Jonathan S. Katz wrote:
On 8/25/26 10:30 AM, Nathan Bossart wrote:
Jonathan's proposed changes seem fine to me at a glance. I'll let that
marinate on the lists a bit more before applying.Thanks; happy to make additional tweaks if need be.
A number of things have been (or are being) reverted, so this will need to
be reworked. We might want to bump one or two features into the list while
we're at it.FYI, I have not modified that list since I am not the owner of that part
of the doc. However, I didn't see anything that needs removal so I
assume others have updated it.
I think the only remaining item that is pending removal is the online
checksums one, and I'm willing to bet that Daniel will take care of that.
What needs rework is Jonathan's patch.
--
nathan
On 15 Sep 2026, at 16:38, Nathan Bossart <nathandbossart@gmail.com> wrote:
I think the only remaining item that is pending removal is the online
checksums one, and I'm willing to bet that Daniel will take care of that.
I did.
--
Daniel Gustafsson
On 9/16/26 5:31 PM, Daniel Gustafsson wrote:
On 15 Sep 2026, at 16:38, Nathan Bossart <nathandbossart@gmail.com> wrote:
I think the only remaining item that is pending removal is the online
checksums one, and I'm willing to bet that Daniel will take care of that.I did.
Here's v3 of the major features list. This merges in the changes from
the v2 proposal upthread with the other changes to the list over the
past several weeks.
Thanks,
Jonathan
Attachments:
v3-doc-release-19-major.difftext/plain; charset=UTF-8; name=v3-doc-release-19-major.diffDownload+26-28
On Fri, Sep 18, 2026 at 07:23:11PM +0800, Jonathan S. Katz wrote:
Here's v3 of the major features list. This merges in the changes from the v2
proposal upthread with the other changes to the list over the past several
weeks.
Thanks for the patch.
+ New <link linkend="sql-repack"><command>REPACK</command></link> command
nit: extra space before "command"
+ that reclaims disk space and reorganizes table contents, which can help + increase query performance. <literal>REPACK CONCURRENTLY</literal> + allows repacking without blocking reads and writes to the table.
s/which can help increase/which can improve
+ Logical replication and can now be enabled without a server restart when + <xref linkend="guc-wal-level"/> is set to <literal>replica</literal>.
nit: extra "and"
+ Support <quote>read-your-writes</quote> on standbys with the new
Hm. Can we be more descriptive here? Maybe something like "Support for
read-your-writes query patterns on standbys via the new..."
+ <link linkend="sql-wait"><command>WAIT FOR</command></link> command.
Per commits 1a562a2661b and e527bf17d04, I think we're just calling it
"WAIT" and not "WAIT FOR".
+ <link linkend="autovacuum-priority">user-defined prioritization of autovacuuming</link>.
IMHO this makes it sound like users will be able to choose the priority of
each individual table, when we actually are just giving them a handful of
coarse controls. I think we ought to emphasize the scoring system instead.
In theory, most users shouldn't need to adjust anything and the
prioritization will "Just Work" for them.
--
nathan