Re: [PoC] Improve dead tuple storage for lazy vacuum

On Wed, Nov 30, 2022 at 2:28 PM Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com>
wrote:
>
> On Fri, Nov 25, 2022 at 5:00 PM John Naylor
> <john(dot)naylor(at)enterprisedb(dot)com> wrote:

> > These hacks were my work, but I think we can improve that by having two
versions of NODE_HAS_FREE_SLOT -- one for fixed- and one for variable-sized
nodes. For that to work, in "init-node" we'd need a branch to set fanout to
zero for node256. That should be fine -- it already has to branch for
memset'ing node128's indexes to 0xFF.
>
> Since the node has fanout regardless of fixed-sized and
> variable-sized

As currently coded, yes. But that's not strictly necessary, I think.

>, only node256 is the special case where the fanout in
> the node doesn't match the actual fanout of the node. I think if we
> want to have two versions of NODE_HAS_FREE_SLOT, we can have one for
> node256 and one for other classes. Thoughts? In your idea, for
> NODE_HAS_FREE_SLOT for fixed-sized nodes, you meant like the
> following?
>
> #define FIXED_NODDE_HAS_FREE_SLOT(node, class)
> (node->base.n.count < rt_size_class_info[class].fanout)

Right, and the other one could be VAR_NODE_...

--
John Naylor
EDB: http://www.enterprisedb.com