Re: Things I don't like about \du's "Attributes" column - Mailing list pgsql-hackers

From David G. Johnston
Subject Re: Things I don't like about \du's "Attributes" column
Date
Msg-id CAKFQuwbmkPtLvXwOnwb=FZxL3iow2HJDQFTuV9kzODkW2sZhgA@mail.gmail.com
Whole thread Raw
In response to Re: Things I don't like about \du's "Attributes" column  (Robert Haas <robertmhaas@gmail.com>)
List pgsql-hackers
On Tue, Jul 16, 2024 at 8:00 AM Robert Haas <robertmhaas@gmail.com> wrote:
I'm starting to have some doubts about whether this effort is really
worthwhile. It seems like what we have right now is a patch which uses
both more horizontal space and more vertical space than the current
implementation, without (IMHO) really offering any clear advantage.

It's simple enough to only add the mandatory vertical space here (and definitely reduce the width) by leaving the current presentation of "N connections" and "password valid until" unchanged but give them their own line in Attributes like all of the rest.  The choice to move them to their own columns seems like where the contention lies.

e.g.,
regress_du_role0 | yes   | Inherit     | Tue Jun 04 00:00:00 2024 PDT |                0 | 
becomes
No connections allowed +|
Inherit +|
Password valid until Tue Jun 04 ... |

(pending confirmation of where the new inherit label fits slot-wise)

versus:

regress_du_role0 | No connections                                            +|
                  | Password valid until 2024-06-04 00:00:00+03                | 
 
(Which actually look the same because of the automatic wrapping on the current version versus explicit one-per-line in the proposed.)

In short, fix the complaint about comma-separated attributes and leave the rest alone as being accurate reflections of the catalog.  In short, from the original message, do "a" without "b".  Tom suggested "either" as well and I agree with Robert that having done both we've made it both wider and taller which ends up not being a great outcome.

Also, address the N connections [allowed] to make it clear this is a static configuration and not some derivation of current state.

Can also deal with Password valid until infinity if a better phrasing can be thought up that doesn't require knowledge of whether a password has been set or not.

David J.

pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: [PATCH] Refactor pqformat.{c,h} and protocol.h
Next
From: Fujii Masao
Date:
Subject: Re: Add a GUC check hook to ensure summarize_wal cannot be enabled when wal_level is minimal