Re: psql: Add role's membership options to the \du+ command - Mailing list pgsql-hackers

From Pavel Luzanov
Subject Re: psql: Add role's membership options to the \du+ command
Date
Msg-id 5c5650e8-6452-c250-89de-1c5d1930130b@postgrespro.ru
Whole thread Raw
In response to Re: psql: Add role's membership options to the \du+ command  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-hackers
On 24.01.2023 20:16, David G. Johnston wrote:
Yeah, I noticed the lack too, then went a bit too far afield with trying to compose a graph of the roles.  I'm still working on that but at this point it probably won't be something I try to get committed to psql.  Something more limited like this does need to be included.

Glad to hear that you're working on it.

I'm not too keen on the idea of converting the existing array into a newline separated string.  I would try hard to make the modification here purely additional.

I agree with all of your arguments. A couple of months I tried to find an acceptable variant in the background.
But apparently tried not very hard ))

In the end, the variant proposed in the patch seemed to me worthy to show and
start a discussion. But I'm not sure that this is the best choice.


Another thing I did with the graph was have both "member" and "memberof" columns in the output.  In short, every grant row in pg_auth_members appears twice, once in each column, so the role being granted membership and the role into which membership is granted both have visibility when you filter on them.  For the role graph I took this idea and extended out to an entire chain of roles (and also broke out user and group separately) but I think doing the direct-grant only here would still be a big improvement.

It will be interesting to see the result.
-- 
Pavel Luzanov

pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: New strategies for freezing, advancing relfrozenxid early
Next
From: Robert Haas
Date:
Subject: Re: New strategies for freezing, advancing relfrozenxid early