PostgreSQL 13 Beta 1 Release Announcement Draft

On Mon, May 18, 2020 at 7:29 PM Jonathan S. Katz <jkatz@postgresql.org>
wrote:

Attached is a draft of the release announcement for the PostgreSQL 13
Beta 1 release this week.

We could call out the additional commits that Tom has done for wait event
renaming, re: compatibility - next to "Rename some recovery-related wait
events".

These are the relevant commits, I think:

Rename SLRU structures and associated LWLocks.
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=5da14938f7bfb96b648ee3c47e7ea2afca5bcc4a

Rename assorted LWLock tranches.
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=36ac359d3621578cefc2156a3917024cdd3b1829

Drop the redundant "Lock" suffix from LWLock wait event names.
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=14a91010912632cae322b06fce0425faedcf7353

Mop-up for wait event naming issues.
https://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=3048898e73c75f54bb259323382e0e7f6368cb6f

Thanks,
Lukas

--
Lukas Fittl

Hi,

On 5/18/20 10:29 PM, Jonathan S. Katz wrote:

Hi,

Attached is a draft of the release announcement for the PostgreSQL 13
Beta 1 release this week.

The goal of this release announcement is to make people aware of the new
features that are introduced in PostgreSQL 13 and, importantly, get them
to start testing. I have tried to include a broad array of features that
can noticeably impact people's usage of PostgreSQL. Note that the order
of the features in the announcement are not in any particular ranking
(though I do call out VACUUM as being one of the "most anticipated
features"), but are my efforts to try and tell a story about the release.

Please let me know your thoughts, comments, corrections, etc. and also
if there are any glaring omissions. I know this is a bit longer than a
typical release announcement, but please let me know your feedback
before the end of Wed. May 20 AOE (i.e. before the release ships).

Thanks everyone for your responses (a note on that in a sec). I have
attached an update to the release, which will be close to and/or the
final copy that goes out tomorrow.

Based on the feedback received, I accepted & declined changes based on
a) how it impacts our users at large b) if an expression was actually
wrong vs. personal preference of explanation. For (a), there is an
existing note in the announcement to read the release notes as there may
be some features that are more interesting to the reader than what may
be described. As mentioned, the goal is to try to have the release
announcement as a springboard into what people can use/test.

I also received an interesting amount of off-list feedback this
iteration. While I generally do not mind chatting (this week an
unfortunate exception and I apologize if I did not reply back to you), I
do ask as part of the collaborative process to try to keep feedback on
list so as to limit duplication of comments.

Anyway, I am happy to receive and incorporate additional feedback up
until we launch tomorrow ~12 UTC.

Thanks! Happy Beta eve :)

Jonathan

On Wed, May 20, 2020 at 06:11:08PM -0400, Jonathan S. Katz wrote:

This release includes more ways to monitor actibity within a PostgreSQL

activity

partition its "accounts" table, making it easier to benchmark workloads that
contains partitions.

contain

No need to respond :)

Thanks,
Justin

On 5/20/20 6:42 PM, Justin Pryzby wrote:

On Wed, May 20, 2020 at 06:11:08PM -0400, Jonathan S. Katz wrote:

This release includes more ways to monitor actibity within a PostgreSQL

activity

...that one is embarrassing. Thanks.

partition its "accounts" table, making it easier to benchmark workloads that
contains partitions.

contain

Adjusted.

No need to respond :)

Happy to :) A lot of the early craziness of the week has now subsided.

Thanks!

Jonathan

Hi Jon,

I noticed a couple minor inconsistencies:

".datetime" -> elsewhere functions are formatted as `.datetime()`

libpq -> `libpq`

The link to the release notes on its own line is the same as the
inline link, if that makes sense. In other places with links on their
own line, the full URL is in the link text.

Also, for "indexes that contain many repeat values", "repeated" might
sound better here. It's one of those things that jumped out at me at
first reading, but when trying both in my head, it seems ok.

Regarding "streaming `pg_basebackup`s", I'm used to the general term
"base backups" in this usage, which seems a distinct concept from the
name of the invoked command.

Thanks!

--
John Naylor https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Hi John,

On 5/21/20 12:12 AM, John Naylor wrote:

Hi Jon,

I noticed a couple minor inconsistencies:

".datetime" -> elsewhere functions are formatted as `.datetime()`

libpq -> `libpq`

The link to the release notes on its own line is the same as the
inline link, if that makes sense. In other places with links on their
own line, the full URL is in the link text.

Also, for "indexes that contain many repeat values", "repeated" might
sound better here. It's one of those things that jumped out at me at
first reading, but when trying both in my head, it seems ok.

Regarding "streaming `pg_basebackup`s", I'm used to the general term
"base backups" in this usage, which seems a distinct concept from the
name of the invoked command.

Thanks for the suggestions. I ended up incorporating all of them.

Stay tuned for the release...

Jonathan

Congrats to all for the release of a new major version!

Two questions:
- Why is VACUUM together with FETCH FIRST WITH TIES, CREATE TABLE LIKE,
ALTER VIEW, ALTER TABLE, etc in Utility Commands section?
Shouldn't there be a separate section for SQL changes? (or keep one
section but rename the Utility to include all?)

Add FOREIGN to ALTER statements, if appropriate (Luis Carril)

WHAT IS THIS ABOUT?

- The "WHAT IS THIS ABOUT?" should be removed, in my opinion.

Again, congrats for another release of the best database in the world.

Pantelis Theodosiou

On Thu, May 21, 2020 at 12:44 PM Jonathan S. Katz <jkatz@postgresql.org>
wrote:

On Thu, May 21, 2020 at 3:20 PM Pantelis Theodosiou <ypercube@gmail.com>
wrote:

Congrats to all for the release of a new major version!

Two questions:
- Why is VACUUM together with FETCH FIRST WITH TIES, CREATE TABLE LIKE,
ALTER VIEW, ALTER TABLE, etc in Utility Commands section?
Shouldn't there be a separate section for SQL changes? (or keep one
section but rename the Utility to include all?)

Add FOREIGN to ALTER statements, if appropriate (Luis Carril)

WHAT IS THIS ABOUT?

- The "WHAT IS THIS ABOUT?" should be removed, in my opinion.

Again, congrats for another release of the best database in the world.

Pantelis Theodosiou

On Thu, May 21, 2020 at 12:44 PM Jonathan S. Katz <jkatz@postgresql.org>
wrote:

Thanks for the suggestions. I ended up incorporating all of them.

Stay tuned for the release...

Jonathan

Apologies, I realized a minute too late that my comments are about the
Release Notes and not the Announcement.
However, since the link to Notes makes them no visible to more eyes, they
could be checked again.

Pantelis Theodosiou

On 19/05/2020 04:29, Jonathan S. Katz wrote:

Hi,

Attached is a draft of the release announcement for the PostgreSQL 13
Beta 1 release this week.

Hi,

Maybe I'm too late, but in this paragraph :

`psql` now includes the `\warn` command that is similar to the `\echo`

command

in terms of outputting data, except `\warn` sends it to stderr. And in

case you

need additional guidance on any of the PostgreSQL commands, the

`--help` flag

now includes a link to

[https://www.postgresql.org](https://www.postgresql.org).

is it --help shouldn't be /help ?

Same thing in the release note
(https://www.postgresql.org/docs/13/release-13.html) :

Add the PostgreSQL home page to command-line |--help| output (Peter

Eisentraut)

as it probalbly refer to 27f3dea64833d68c1fa08c1e5d26176a579f69c8, isn't
it ?

regards,

--
Sébastien