Обсуждение: BUG #6565: GEQO documentation discrepancy after 9.0 change?

The following bug has been logged on the website:

Bug reference:      6565
Logged by:          Guillaume Cottenceau
Email address:      gcottenc@gmail.com
PostgreSQL version: 9.1.3
Operating system:   not relevant
Description:=20=20=20=20=20=20=20=20

9.0 release notes contain:

"""
E.8.3.1.4. GEQO

Use the same random seed every time GEQO plans a query (Andres Freund)

While the Genetic Query Optimizer (GEQO) still selects random plans, it now
always selects the same random plans for identical queries, thus giving more
consistent performance. You can modify geqo_seed to experiment with
alternative plans.
"""

However, documentation up to 9.1 (pg website) say:

http://www.postgresql.org/docs/9.1/static/runtime-config-query.html#RUNTIME=
-CONFIG-QUERY-GEQO

"""GEQO's searching is randomized and therefore its plans may vary
nondeterministically."""

I guess this sentence is outdated now?

On Thu, Mar 29, 2012 at 3:36 AM,  <gcottenc@gmail.com> wrote:
> The following bug has been logged on the website:
>
> Bug reference: =A0 =A0 =A06565
> Logged by: =A0 =A0 =A0 =A0 =A0Guillaume Cottenceau
> Email address: =A0 =A0 =A0gcottenc@gmail.com
> PostgreSQL version: 9.1.3
> Operating system: =A0 not relevant
> Description:
>
> 9.0 release notes contain:
>
> """
> E.8.3.1.4. GEQO
>
> Use the same random seed every time GEQO plans a query (Andres Freund)
>
> While the Genetic Query Optimizer (GEQO) still selects random plans, it n=
ow
> always selects the same random plans for identical queries, thus giving m=
ore
> consistent performance. You can modify geqo_seed to experiment with
> alternative plans.
> """
>
> However, documentation up to 9.1 (pg website) say:
>
> http://www.postgresql.org/docs/9.1/static/runtime-config-query.html#RUNTI=
ME-CONFIG-QUERY-GEQO
>
> """GEQO's searching is randomized and therefore its plans may vary
> nondeterministically."""
>
> I guess this sentence is outdated now?

Hmm, sounds like it.  Does anyone think otherwise?

--=20
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Robert Haas <robertmhaas@gmail.com> writes:
> On Thu, Mar 29, 2012 at 3:36 AM,  <gcottenc@gmail.com> wrote:
>> """GEQO's searching is randomized and therefore its plans may vary
>> nondeterministically."""
>>
>> I guess this sentence is outdated now?

> Hmm, sounds like it.  Does anyone think otherwise?

Well, you could still argue that it's nondeterministic, but the
nondeterminism is at least mostly hidden from users.  I've removed
that sentence and adjusted some other places to match.

            regards, tom lane