| From: | Bernd Helmle <mailings(at)oopsware(dot)de> |
|---|---|
| To: | KaiGai Kohei <kaigai(at)ak(dot)jp(dot)nec(dot)com> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: [PATCH] [v8.5] Security checks on largeobjects |
| Date: | 2009-06-29 18:52:52 |
| Message-ID: | C4F0921DE0ED6A41BE5DC618@[192.168.1.119] |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
--On 29. Juni 2009 08:32:29 +0900 KaiGai Kohei <kaigai(at)ak(dot)jp(dot)nec(dot)com> wrote:
> Yes, it intends to assign an identifier string not only numeric
> large object identifier. The identifier string can be qualified
> with a certain namespace as follows.
>
> E.g)
> SELECT lo_open('my_picture01', x'40000'::int);
> SELECT lo_create('pg_temp.my_musid02');
>
> In the later case, the new largeobject will be reclaimed after
> the session closed due to the temporary namespace.
I'm not sure about the usefulness of this. While having an identifier for a
LO is nice, i believe most users store additional metadata about objects
within their own tables anyways, linking the LO there. Also i doubt there
is much need for temporary large objects (at least, i have no idea about
this....).
It might be interesting to dig into your proposal deeper in conjunction
with TOAST (you've already mentioned this TODO). Having serial access with
a nice interface into TOAST would be eliminating the need for
pg_largeobject completely (i'm not a big fan of this one-big-system-table
approach the old LO interface currently is).
--
Thanks
Bernd
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2009-06-29 18:53:07 | Re: pre-proposal: permissions made easier |
| Previous Message | Josh Berkus | 2009-06-29 18:41:47 | Re: pre-proposal: permissions made easier |