From: | Stephan Szabo <sszabo(at)megazone23(dot)bigpanda(dot)com> |
---|---|
To: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Constraint stuff (fwd) |
Date: | 2000-08-07 23:29:33 |
Message-ID: | Pine.BSF.4.10.10008071629020.52054-100000@megazone23.bigpanda.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
I need to learn how to type headers I think... :(
Stephan Szabo
sszabo(at)bigpanda(dot)com
On Mon, 7 Aug 2000, Jan Wieck wrote:
> > More generally, the PL functions need to be able to deal with
> > recomputing saved plans after an ALTER of a table referenced by the
> > function. I haven't really thought about how to do that ... but it
> > seems like Stephan's idea of a table showing referencers and referencees
> > might help detect the cases where plans have to be flushed.
>
> More more generally, you cannot tell which objects are
> referenced from saved plans inside of a PL function. Might be
> possible for PL/pgSQL, where we could use a specialized
> invocation of the PL compiler to determine. But there's no
> way to do it for PL/Tcl or the like.
> We might end up with some general catalog sequence, bumped
> any time a schema change happens, and require each function
> to forget about all of it's saved plans for the next
> transaction. Ugly, but the only way I see to be consistent.
As a dumb question to help me understand better...
What exactly is saved in the plans and how are the plan saved for a
PL/Tcl function that does something where it generates a query that
you say don't know the table of until run time?
From | Date | Subject | |
---|---|---|---|
Next Message | The Hermit Hacker | 2000-08-07 23:45:42 | Re: AW: LIKE/ESCAPE et al, initdb required! |
Previous Message | Tom Lane | 2000-08-07 20:23:36 | Re: Constraint stuff |