Re: search_path wildcard?

2024-05-22 Thread Pavel Stehule
st 22. 5. 2024 v 21:13 odesílatel Ron Johnson 
napsal:

> On Wed, May 22, 2024 at 1:58 PM Tom Lane  wrote:
>
>> Ron Johnson  writes:
>> > That would be a helpful feature for administrators, when there are
>> multiple
>> > schemas in multiple databases, on multiple servers: superusers get ALTER
>> > ROLE foo SET SEARCH_PATH  = '*'; and they're done with it.
>>
>> ... and they're pwned within five minutes by any user with the wits
>> to create a trojan-horse function or operator.  Generally speaking,
>> you want admins to run with a minimal search path not a maximal one.
>>
>
> Missing tables when running "\t" is a bigger hassle.
>

what is hard on \dt *.*

or you can define own

 dtall = '\\dt *.*'

:dtall

The problem is not on search path, but maybe on design backslash commands -
but there should be some level of consistency

Regards

Pavel


Re: search_path wildcard?

2024-05-22 Thread Pavel Stehule
st 22. 5. 2024 v 19:54 odesílatel Ron Johnson 
napsal:

> On Wed, May 22, 2024 at 12:53 PM David G. Johnston <
> david.g.johns...@gmail.com> wrote:
>
>> On Wed, May 22, 2024, 10:36 Ron Johnson  wrote:
>>
>>> This doesn't work, and I've found nothing similar:
>>> ALTER ROLE foo SET SEARCH_PATH  = '*';
>>>
>>
>> Correct, you cannot do that.
>>
>
> That would be a helpful feature for administrators, when there are
> multiple schemas in multiple databases, on multiple servers: superusers get 
> ALTER
> ROLE foo SET SEARCH_PATH  = '*'; and they're done with it.
>

It can be pretty dangerous, because you don't specify order of schemas

Regards

Pavel


Re: search_path wildcard?

2024-05-22 Thread Ron Johnson
On Wed, May 22, 2024 at 1:58 PM Tom Lane  wrote:

> Ron Johnson  writes:
> > That would be a helpful feature for administrators, when there are
> multiple
> > schemas in multiple databases, on multiple servers: superusers get ALTER
> > ROLE foo SET SEARCH_PATH  = '*'; and they're done with it.
>
> ... and they're pwned within five minutes by any user with the wits
> to create a trojan-horse function or operator.  Generally speaking,
> you want admins to run with a minimal search path not a maximal one.
>

Missing tables when running "\t" is a bigger hassle.


Re: search_path wildcard?

2024-05-22 Thread Tom Lane
Ron Johnson  writes:
> That would be a helpful feature for administrators, when there are multiple
> schemas in multiple databases, on multiple servers: superusers get ALTER
> ROLE foo SET SEARCH_PATH  = '*'; and they're done with it.

... and they're pwned within five minutes by any user with the wits
to create a trojan-horse function or operator.  Generally speaking,
you want admins to run with a minimal search path not a maximal one.

regards, tom lane




Re: search_path wildcard?

2024-05-22 Thread Ron Johnson
On Wed, May 22, 2024 at 12:53 PM David G. Johnston <
david.g.johns...@gmail.com> wrote:

> On Wed, May 22, 2024, 10:36 Ron Johnson  wrote:
>
>> This doesn't work, and I've found nothing similar:
>> ALTER ROLE foo SET SEARCH_PATH  = '*';
>>
>
> Correct, you cannot do that.
>

That would be a helpful feature for administrators, when there are multiple
schemas in multiple databases, on multiple servers: superusers get ALTER
ROLE foo SET SEARCH_PATH  = '*'; and they're done with it.


Re: search_path wildcard?

2024-05-22 Thread David G. Johnston
On Wed, May 22, 2024, 10:36 Ron Johnson  wrote:

> This doesn't work, and I've found nothing similar:
> ALTER ROLE foo SET SEARCH_PATH  = '*';
>

Correct, you cannot do that.

David J.