Re: Single client performance on trivial SELECTs - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Single client performance on trivial SELECTs
Date
Msg-id 23195.1302877433@sss.pgh.pa.us
Whole thread Raw
In response to Re: Single client performance on trivial SELECTs  (Noah Misch <noah@leadboat.com>)
List pgsql-hackers
Noah Misch <noah@leadboat.com> writes:
> On Thu, Apr 14, 2011 at 05:10:41PM -0400, Tom Lane wrote:
>> FWIW, mysql used to use gperf for this purpose, but they've abandoned it
>> in favor of some homegrown hashing scheme.

> Do you have any details, like when mysql did this?  With a quick look, I'm
> failing to find confirmation that mysql ever did use gperf.  (Drizzle has
> replaced the mysql homegrown scheme with gperf, apparently in 2009, though.)

Now I'm not sure.  The evidence that I based that comment on is that Red
Hat's mysql RPMs have had "BuildRequires: gperf" since forever --- as
far back as I can find in their package CVS anyway --- and I doubt that
would've got put in unless it were really needed to build.  But I can't
find any mention of gperf in either the current mysql releases or the
oldest mysql tarball I have, which is mysql 3.23.58.  So if it really
was used, it was a very very long time ago.  If the Drizzle guys went
in the other direction, that would suggest that they don't have any
institutional memory of having rejected gperf, so maybe that
BuildRequires is just a thinko after all.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: Foreign table permissions and cloning
Next
From: Tom Lane
Date:
Subject: Re: [COMMITTERS] pgsql: Rename pg_regress option --multibyte to --encoding