Thread: psql patch: new host/port
Folks, Please find enclosed a patch that lets you use \c to connect (optionally) to a new host and port without exiting psql. Cheers, D -- David Fetter david@fetter.org http://fetter.org/ phone: +1 415 235 3778 Remember to vote!
Attachment
On Dec 9, 2005, at 18:10 , David Fetter wrote: > Please find enclosed a patch that lets you use \c to connect > (optionally) to a new host and port without exiting psql. I'm not familiar enough with the psql code to be able to tell, but is this secure? The pg_hba.conf on the new server is enforced, I assume? Michael Glaesemann grzm myrealbox com
Michael Glaesemann wrote: > I'm not familiar enough with the psql code to be able to tell, but is > this secure? The pg_hba.conf on the new server is enforced, I assume? You don't need to be familiar with the psql code to know that it would be pretty stupid if client programs could override the server authentication setup. -- Peter Eisentraut http://developer.postgresql.org/~petere/
On Dec 12, 2005, at 20:33 , Peter Eisentraut wrote: > Michael Glaesemann wrote: >> I'm not familiar enough with the psql code to be able to tell, but is >> this secure? The pg_hba.conf on the new server is enforced, I assume? > > You don't need to be familiar with the psql code to know that it would > be pretty stupid if client programs could override the server > authentication setup. I'm sorry if I wasn't clear. My point was I'm not familiar enough with the code to see if this implementation is secure. I do indeed realize that clients bypassing server authentication is a Bad Thing. Michael Glaesemann grzm myrealbox com
Michael Glaesemann said: > > On Dec 12, 2005, at 20:33 , Peter Eisentraut wrote: > >> Michael Glaesemann wrote: >>> I'm not familiar enough with the psql code to be able to tell, but is >>> this secure? The pg_hba.conf on the new server is enforced, I assume? >> >> You don't need to be familiar with the psql code to know that it would >> be pretty stupid if client programs could override the server >> authentication setup. > > I'm sorry if I wasn't clear. My point was I'm not familiar enough > with the code to see if this implementation is secure. I do indeed > realize that clients bypassing server authentication is a Bad Thing. > The patch is to the client only, not even to libpq, so of course no auth bypass is involved. cheers andrew
On Dec 12, 2005, at 21:32 , Andrew Dunstan wrote: > The patch is to the client only, not even to libpq, so of course no > auth > bypass is involved. Cool. Thanks for the explanation, Andrew. Michael Glaesemann grzm myrealbox com
On Mon, Dec 12, 2005 at 09:20:57PM +0900, Michael Glaesmann wrote: > > On Dec 12, 2005, at 21:32 , Andrew Dunstan wrote: > > >The patch is to the client only, not even to libpq, so of course no > >auth bypass is involved. > > Cool. Thanks for the explanation, Andrew. Is the patch suitable for a re-send to -patches? Cheers, D -- David Fetter david@fetter.org http://fetter.org/ phone: +1 415 235 3778 Remember to vote!
Michael Glaesemann <grzm@myrealbox.com> writes:
> On Dec 9, 2005, at 18:10 , David Fetter wrote:
>> Please find enclosed a patch that lets you use \c to connect
>> (optionally) to a new host and port without exiting psql.
> I'm not familiar enough with the psql code to be able to tell, but is
> this secure? The pg_hba.conf on the new server is enforced, I assume?
No, security is the server's problem.
What's not clear to me about this patch is what's the point. It's
certainly not a feature we've heard any requests for.
regards, tom lane
On Mon, Dec 12, 2005 at 10:19:00AM -0500, Tom Lane wrote: > Michael Glaesemann <grzm@myrealbox.com> writes: > > On Dec 9, 2005, at 18:10 , David Fetter wrote: > >> Please find enclosed a patch that lets you use \c to connect > >> (optionally) to a new host and port without exiting psql. > > > I'm not familiar enough with the psql code to be able to tell, but > > is this secure? The pg_hba.conf on the new server is enforced, I > > assume? > > No, security is the server's problem. > > What's not clear to me about this patch is what's the point. It's > certainly not a feature we've heard any requests for. Informally, I've heard some carping about how you can change DBs on one server, but you have to exit the program if you want to change servers. The change is low-impact because \c continues to work exactly as before when supplied with < 3 arguments :) Cheers, D -- David Fetter david@fetter.org http://fetter.org/ phone: +1 415 235 3778 Remember to vote!