Re: RESULT_OID Bug

On Wed, Jul 27, 2005 at 12:56:15AM -0400, Tom Lane wrote:
> Michael Fuhr <mike(at)fuhr(dot)org> writes:
> > Could this be platform-specific?
>
> Seems that way. I tried it on HPUX 10.20/HPPA/gcc 2.95.3.

No luck on FreeBSD 4.11-STABLE/i386/gcc 2.95.4. The box that does
have a problem is Solaris 9/sparc/gcc 3.4.2.

Can anybody else reproduce the problem?

> My guess is that the behavior is related to plpgsql's caching
> of plans for functions, and as such should be driven by the
> backend's history not the whole database's history. But it's
> just a guess.

Another test case that's been consistent for me:

Session 1: connect
Session 1: create table and function
Session 1: call function; returns oid

Session 2: connect
Session 2: call function; returns NULL

Session 1: exit

Session 3: connect
Session 3: call function; returns NULL

Session 2: exit

Session 3: exit

Session 4: connect
Session 4: call function; returns oid

Session 5: connect
Session 5: call function; returns NULL

Any suggestions? Would it be useful to attach gdb to one of the
backends? If so, what should I be looking for?

--
Michael Fuhr
http://www.fuhr.org/~mfuhr/