From: | Andres Freund <andres(at)anarazel(dot)de> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Make planning via GEQO repeatable |
Date: | 2009-07-16 21:40:28 |
Message-ID: | 200907162340.28283.andres@anarazel.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Thursday 16 July 2009 23:04:58 Tom Lane wrote:
> Andres Freund <andres(at)anarazel(dot)de> writes:
> > Query planning via GEQO currently can yield a different plan on every
> > invokation of the planner due to its non-exhaustive nature.
> > This often can be inconvenient because at times there may be a very
> > bad plan. It also makes it very hard to reproduce a problem with GEQO.
>
> Applied with some editorialization. Mainly, I didn't see the point of
> preserving the ability to have nondeterministic planning, and I
> especially didn't care for having that still be the default behavior.
> So I just made it unconditionally initialize the seed. It would of
> course take only minor tweaking to do things differently.
Nice.
Mainly I did not have the guts to change the behaviour completely...
archive.org has a copy of the dead link to the comp.ai.genetic FAQ linked at
http://web.archive.org/web/20051226001402/http://www.cs.bham.ac.uk/Mirrors/ftp.de.uu.net/EC/clife/www/location.htm
which is the same as the one referenced by alvaro in
http://archives.postgresql.org/pgsql-docs/2009-07/msg00004.php
If considerered relevant enough, you can update the link...
Andres
From | Date | Subject | |
---|---|---|---|
Next Message | Alvaro Herrera | 2009-07-16 21:47:08 | Re: Launching commitfest.postgresql.org |
Previous Message | Jaime Casanova | 2009-07-16 21:31:24 | Re: Review: support for multiplexing SIGUSR1 |