Re: [WIP] cache estimates, cache access cost

On 06/14/2011 07:08 PM, Tom Lane wrote:
> I concur with Robert's desire to not push experimental code into the
> main repository, but we need to have *some* way of working with it.
> Maybe a separate repo where experimental branches could hang out would
> be helpful?
>

Well, this one is sitting around in branches at both git.postgresql.org
and github so far, both being updated periodically. Maybe there's some
value around an official experimental repository too, but I thought that
was the idea of individual people having their own directories on
git.postgres.org. Do we need something fancier than that? It would be
nice, but seems little return on investment to improve that, relative to
what you can do easily enough now.

The idea David Fetter has been advocating of having a "bit rot" farm to
help detect when the experimental branches drift too far out of date
tries to make that concept really formal. I like that idea, too, but
find it hard to marshal enough resources to do something about it. The
current status quo isn't that terrible; noticing bit rot when it's
relevant isn't that hard to do.

--
Greg Smith 2ndQuadrant US greg(at)2ndQuadrant(dot)com Baltimore, MD
PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.us