Misleading sentence about default privileges - Mailing list pgsql-docs

From PG Doc comments form
Subject Misleading sentence about default privileges
Date
Msg-id 162392083110.690.13436519749099977740@wrigleys.postgresql.org
Whole thread Raw
Responses Re: Misleading sentence about default privileges  (Bruce Momjian <bruce@momjian.us>)
List pgsql-docs
The following documentation comment has been logged on the website:

Page: https://www.postgresql.org/docs/10/sql-alterdefaultprivileges.html
Description:

In the docs
(https://www.postgresql.org/docs/10/sql-alterdefaultprivileges.html) it
states:

> You can change default privileges only for objects that will be created by
yourself or by roles that you are a member of.

Yet, altering the default privileges `for role`'s that I am a member of
(i.e. `target_role` in docs), does not affect privileges granted on objects
created by other members of said role.

Seeing as separating Users (roles with log-in privilege) from Roles
(containing concrete grants, unable to log in) seems a common, and
recommendable pattern, I believe the statement is quite misleading.

For an example of expected behaviour, see this Stack Overflow question:
https://stackoverflow.com/questions/56237907/why-doesnt-alter-default-privileges-work-as-expected

The only scenario I can think of where the statement makes sense seems quite
foreign to me: 
Scenario: I, say `role_a`, have log-in, and am also a member of another
Role, say `role_b`, which also has login. Only objects created directly by
`role_b` (i.e. not any of its members) are affected.

I suggest adding something like the following to the documentation:

" Note that only object created directly by _*target_role*_ , i.e. not any
of its members, will have privileges granted. "

pgsql-docs by date:

Previous
From:
Date:
Subject: change float4 to floating point about type of autovacuum_vacuum_insert_scale_factor
Next
From: PG Doc comments form
Date:
Subject: 24.1.5.1. Multixacts And Wraparound