DOCS - Server Applications [option] should be [option...]

Started by Peter Smith7 months ago7 messageshackers
Beta feature

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.

won't retrysuccessCI history

You can run a PostgreSQL built from this patch straight from Docker, with no checkout and no build:

docker run --rm -p 5432:5432 ghcr.io/hackorum-dev/postgres-patch:t53327
psql -h localhost -U postgres

Built from patchset v1 (message #1), July 29, 2026 at 05:21 PM.

Every patchset is also pushed to a branch of our PostgreSQL fork, so you can check out the same tree CI built. Without a PostgreSQL checkout:

git clone --branch t53327_1 https://github.com/hackorum-dev/postgres.git

In a checkout you already have, add the fork once:

git remote add hackorum https://github.com/hackorum-dev/postgres.git

then, for this patchset and every later one:

git fetch hackorum t53327_1 && git checkout t53327_1

Patchset v1 (message #1) is on t53327_1

Jump to latest
#1Peter Smith
smithpb2250@gmail.com

Hi.

Make the "option" part consistent in the synopses of all Server Applications:

1. "[option]" should be "[option...]"
2. It should be first

PSA patch for the same.

======
Kind Regards,
Peter Smith.
Fujitsu Australia

Attachments:

t53327_1
v1-0001-DOCS-option-should-be-first-and-have-ellipsis.patchapplication/octet-stream; name=v1-0001-DOCS-option-should-be-first-and-have-ellipsis.patchDownload+3-4
#2Andreas Karlsson
andreas.karlsson@percona.com
In reply to: Peter Smith (#1)
Re: DOCS - Server Applications [option] should be [option...]

On 2/17/26 2:08 AM, Peter Smith wrote:

Hi.

Make the "option" part consistent in the synopses of all Server Applications:

1. "[option]" should be "[option...]"

Agreed.

2. It should be first

I am not sure that is an improvement. I would rather move things with
less signal towards the end. Also how many examples have we of each?
Most seem to be on the format below which I feel indicates [option...]
is last (but before order dependent arguments like dbname.)

pg_dump [connection-option...] [option...] [dbname]

Andreas

#3Peter Smith
smithpb2250@gmail.com
In reply to: Andreas Karlsson (#2)
Re: DOCS - Server Applications [option] should be [option...]

On Wed, Feb 18, 2026 at 7:32 PM Andreas Karlsson <andreas@proxel.se> wrote:

On 2/17/26 2:08 AM, Peter Smith wrote:

Hi.

Make the "option" part consistent in the synopses of all Server Applications:

1. "[option]" should be "[option...]"

Agreed.

2. It should be first

I am not sure that is an improvement. I would rather move things with
less signal towards the end. Also how many examples have we of each?
Most seem to be on the format below which I feel indicates [option...]
is last (but before order dependent arguments like dbname.)

pg_dump [connection-option...] [option...] [dbname]

Thanks for the feedback

The scope of my patch is only the Server Applications [1]https://www.postgresql.org/docs/devel/reference-server.html. Maybe it is
still correct that it is not much of an improvement, but you are
citing examples from Client Applications.

======
[1]: https://www.postgresql.org/docs/devel/reference-server.html

Kind Regards,
Peter Smith.
Fujitsu Australia

#4Andreas Karlsson
andreas.karlsson@percona.com
In reply to: Peter Smith (#3)
Re: DOCS - Server Applications [option] should be [option...]

On 2/18/26 10:05 AM, Peter Smith wrote:

On Wed, Feb 18, 2026 at 7:32 PM Andreas Karlsson <andreas@proxel.se> wrote:

I am not sure that is an improvement. I would rather move things with
less signal towards the end. Also how many examples have we of each?
Most seem to be on the format below which I feel indicates [option...]
is last (but before order dependent arguments like dbname.)

pg_dump [connection-option...] [option...] [dbname]

Thanks for the feedback

The scope of my patch is only the Server Applications [1]. Maybe it is
still correct that it is not much of an improvement, but you are
citing examples from Client Applications.

Ah, now I see why it is put first in many of the commands. It is to
allow writing the following:

initdb [option...] [ --pgdata | -D ] directory

So I guess I will have to change my mind and say that putting it first
is the least bad option.

Side note: that looks like a bug, shouldn't it be the below?

initdb [option...] [[ --pgdata | -D ] directory]

Andreas

#5Peter Smith
smithpb2250@gmail.com
In reply to: Andreas Karlsson (#4)
Re: DOCS - Server Applications [option] should be [option...]

On Thu, Feb 19, 2026 at 8:32 AM Andreas Karlsson <andreas@proxel.se> wrote:

...

Side note: that looks like a bug, shouldn't it be the below?

initdb [option...] [[ --pgdata | -D ] directory]

You will have noticed that I am trying to keep each of these recent
threads/patches narrowly focussed on just one specific problem at a
time.

Last year I had tried fixing many of these same problems at once but
that thread went in too many directions at once, and got bogged down
eventually forgotten about. I hope this focussed thread strategy has a
better chance of success.

~

But yes, I agree with you. The Option description [1]https://www.postgresql.org/docs/devel/app-initdb.html#APP-INITDB-OPTION-PGDATA says "This is
the only information required by initdb, but you can avoid writing it
by..."

Feel free to post a thread / patch for this, or else I can add it to
my list and get to it sometime.

======
[1]: https://www.postgresql.org/docs/devel/app-initdb.html#APP-INITDB-OPTION-PGDATA

Kind Regards,
Peter Smith.
Fujitsu Australia

#6Daniel Gustafsson
daniel@yesql.se
In reply to: Peter Smith (#1)
Re: DOCS - Server Applications [option] should be [option...]

On 17 Feb 2026, at 02:08, Peter Smith <smithpb2250@gmail.com> wrote:

Make the "option" part consistent in the synopses of all Server Applications:

1. "[option]" should be "[option...]"

Only if the application can take multiple options. pg_controldata can for
example only take a single parameter so [option] is correct there.

2. It should be first

Why? I think it makes more sense to list [option...] after required parameters
just like how pg_upgrade does it. We might not be consistent as is, but I'm
not sure it's a net improvement to always list it first as opposed to what we
have.

--
Daniel Gustafsson

#7Peter Smith
smithpb2250@gmail.com
In reply to: Daniel Gustafsson (#6)
Re: DOCS - Server Applications [option] should be [option...]

On Tue, Feb 24, 2026 at 1:48 AM Daniel Gustafsson <daniel@yesql.se> wrote:

On 17 Feb 2026, at 02:08, Peter Smith <smithpb2250@gmail.com> wrote:

Make the "option" part consistent in the synopses of all Server Applications:

1. "[option]" should be "[option...]"

Only if the application can take multiple options. pg_controldata can for
example only take a single parameter so [option] is correct there.

You are correct. Thanks for pointing it out.

2. It should be first

Why? I think it makes more sense to list [option...] after required parameters
just like how pg_upgrade does it. We might not be consistent as is, but I'm
not sure it's a net improvement to always list it first as opposed to what we
have.

This patch/thread was primarily about #1, which turned out to be
mistaken. The #2 re-ordering part was just done at the same time, but
got it no support

So, I am withdrawing this patch.

======
Kind Regards,
Peter Smith.
Fujitsu Australia