Re: RFC/WIP: adding new configuration options to TOAST

On 4 November 2015 at 20:48, Bill Moran <wmoran(at)potentialtech(dot)com> wrote:

> On a related note, since I've been digging in this area of code I feel
> comfortable commenting that pluggable compression isn't terribly difficult
> to implement. The most significant work will probably be in actually
> implementing the various algorithms, not in making them pluggable. I've
> been considering making that my next project.

IIRC the biggest roadblocks for pluggable compression were around
binary upgrade and around the per-Datum or per-tuple space cost of
identifying which algorithm is used. But I guess that's getting
off-track.

> Overall, I don't feel like you're actually disagreeing with
> anything I'm doing ... you're just wishing I was already further
> along.

Indeed. I'm concerned about the extra complexity exposed to users and
unsure about GUCs as an interface, but I can see being able to
encourage earlier and more aggressive compression as potentially quite
handy.

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services