| From: | Peter Eisentraut <peter_e(at)gmx(dot)net> |
|---|---|
| To: | Adam Brightwell <adam(dot)brightwell(at)crunchydatasolutions(dot)com> |
| Cc: | Stephen Frost <sfrost(at)snowman(dot)net>, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: superuser() shortcuts |
| Date: | 2014-11-20 21:52:14 |
| Message-ID: | 546E628E.3040602@gmx.net |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 11/5/14 5:10 PM, Adam Brightwell wrote:
> Attached is two separate patches to address previous
> comments/recommendations:
>
> * superuser-cleanup-shortcuts_11-5-2014.patch
Seeing that the regression tests had to be changed in this patch
indicates that there is a change of functionality, even though this
patch was previously discussed as merely an internal cleanup. So either
there is a user-facing change, which would need to be documented (or at
least mentioned, discussed, and dismissed as minor), or there is a fault
in the tests, which should be fixed first independently.
Which makes me wonder whether the other changes are indeed without
effect or just not covered by tests.
> * has_privilege-cleanup_11-5-2014.patch
I don't really see the merit of this patch. A "cleanup" patch that adds
more code than it removes isn't really a cleanup. If you want to make
the APIs consistent, then let's make a patch for that. If you want to
make the error messages consistent, then let's make a patch for that.
There is other work going on replacing these role attributes with
something more general, so maybe this cleanup should be done as part of
that other work.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andres Freund | 2014-11-20 22:03:41 | Re: superuser() shortcuts |
| Previous Message | Tom Lane | 2014-11-20 21:47:29 | Re: pg_multixact not getting truncated |