Re: psql's \dn versus temp schemas

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Sat, Sep 18, 2010 at 6:35 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>> On Sat, Sep 18, 2010 at 3:11 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>> This is at least inconsistent and at worst wildly misleading. ISTM
>>> we ought to adopt some combination of the following ideas:

>> I vote for this combination:
>>
>>> 3. Don't show either pg_temp_nn or pg_toast_temp_nn schemas, not even
>>> for the current backend.
>>
>> and
>>
>>> With any of 1-3 we could also consider adding a rule that \dn+
>>> doesn't hide them.

This approach makes sense to me too; I'd be inclined to hide pg_toast as
well under the same rules. In all of these cases, the schemas are not
meant to be referred to explicitly. I think that the original
motivation for letting \dn show the backend's own pg_temp_nn schema
was that there were cases where you needed to refer to it by name.
Since then, we invented the "pg_temp" alias mechanism, which seems to
remove most of the need for that.

> Or perhaps another option would be to make \dnS display these. Not
> sure whether I like that or not.

Hmm. If we had a \dnS option, what I would sorta expect it to do is
show the "system" schemas pg_catalog and information_schema. The toast
and temp schemas seem like a different category somehow. On the other
hand, if we did it like this, then the S and + modifiers would be
orthogonal which is a nice property.

Anyone else have an opinion?

regards, tom lane