Re: CREATEROLE and role ownership hierarchies - Mailing list pgsql-hackers

From Mark Dilger
Subject Re: CREATEROLE and role ownership hierarchies
Date
Msg-id A4FCDB8C-9965-4977-90C9-9EFFE2FEF002@enterprisedb.com
Whole thread Raw
In response to Re: CREATEROLE and role ownership hierarchies  (Andrew Dunstan <andrew@dunslane.net>)
Responses Re: CREATEROLE and role ownership hierarchies  (Andrew Dunstan <andrew@dunslane.net>)
List pgsql-hackers

> On Feb 1, 2022, at 1:10 PM, Andrew Dunstan <andrew@dunslane.net> wrote:
>
> The whole 'NOFOO WITH ADMIN OPTION'
> thing seems to me a bit like a POLA violation. Nevertheless I can
> probably live with it as long as it's *really* well documented. Even so
> I suspect it would be too complex for many, and they will just continue
> to use superusers to create and manage roles if possible.

I agree with the sentiment, but it might help to distinguish between surprising behavior vs. surprising grammar.

In existing postgresql releases, having CREATEROLE means you can give away most attributes, including ones you yourself
don'thave (createdb, login).  So we already have the concept of NOFOO WITH ADMIN OPTION, we just don't call it that.
Inpre-v8 patches on this thread, I got rid of that; you *must* have the attribute to give it away.  But maybe that was
toorestrictive, and we need a way to specify, attribute by attribute, how this works.  Is this just a problem of
surprisinggrammar?  Is it surprising behavior?  If the latter, I'm inclined to give up this WIP as having been a bad
move. If the former, I'll try to propose some less objectionable grammar. 


—
Mark Dilger
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company






pgsql-hackers by date:

Previous
From: Thomas Munro
Date:
Subject: Re: Windows crash / abort handling
Next
From: Nathan Bossart
Date:
Subject: Re: make MaxBackends available in _PG_init