Re: Temporary tables and miscellaneous schemas - Mailing list pgsql-general

From Sean Chittenden
Subject Re: Temporary tables and miscellaneous schemas
Date
Msg-id 20031027211748.GT86551@perrin.nxad.com
Whole thread Raw
In response to Re: Temporary tables and miscellaneous schemas  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Temporary tables and miscellaneous schemas  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
> > Um, I forget whether or not this was given any credence or anyone
> > weighed in on it, but what about having two modes for psql?  An
> > admin mode which hides nothing and is the default for superuser
> > connections, and a user mode which is the default for non-DBA
> > connections.
>
> I thought that would be likely to create more confusion than it
> solves.
>
> To take just one problem, the newbies who could use the "friendly
> user" mode are very likely the same ones who do all their work as
> postgres, because it hasn't occurred to them to create any
> unprivileged users.  They won't get the benefit of it if we make it
> act as you suggest.

Hrm, well, two flaws with that argument being:

1) Users who (ab)use DBA accounts aren't likely the ones with
   gazillions of pg_temp_* tables and probably don't even make use of
   temp tables or care about pg_toast.  No harm, no foul, as the
   feature isn't likely used.

2) Queries that are written by a DBA and given to a user will still
   work when executed by the user, so the confusion is limited to a
   \command not showing the same results that a DBA sees.

Seeing extra info if your prompt is '#' and not '%' shouldn't surprise
anyone.  Few complain about tab completion in shells not listing
programs that aren't readable by the current user.

eg:

    % /usr/local/bin/root_only_cmd[TAB]
    *system beeps, root_only_cmd_here isn't executable by $USER*
    # /usr/local/bin/root_only_cmd[TAB]
    # /usr/local/bin/root_only_cmd_here


-sc

--
Sean Chittenden

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Temporary tables and miscellaneous schemas
Next
From: Tom Lane
Date:
Subject: Re: sequences: cache_value, log_cnt, is_cycled columns