Re: Extensible executor nodes for preparation of SQL/MED

On Wed, Nov 17, 2010 at 2:13 AM, Itagaki Takahiro
<itagaki(dot)takahiro(at)gmail(dot)com> wrote:
> On Wed, Nov 17, 2010 at 10:51, Itagaki Takahiro
> <itagaki(dot)takahiro(at)gmail(dot)com> wrote:
>> On Wed, Nov 17, 2010 at 03:36, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>>> On Tue, Nov 16, 2010 at 1:28 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>>> I am of the opinion that a run-time-extensible set of plan node types
>>>> is merest fantasy.  We will never have that, so putting in place 5%
>>>> of the infrastructure for it is a waste of time and notational
>>>> complexity.
>>>
>>> I think I agree; and moreover there's been no compelling argument made
>>> why we would need that for SQL/MED anyway.
>>
>> I see. I'll cut useless parts from my patch.
>
> I tested simplified version, but I cannot see measurable performance
> improvement at this time. So, I'll turn down the whole proposal
> to use function pointer calls. I'm sorry for all the fuss.

Wait a minute... I'm confused. Didn't you have a measurable
performance improvement with an earlier version of this patch? If
taking out the "useless" parts removed the performance benefit, maybe
they weren't useless?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company