Robert,
* Robert Haas (robertmhaas@gmail.com) wrote:
> It probably makes sense to start with indexam api changes, since the
> other one depends on that one.
I've looked the patch over, but the problem is that the patch doesn't do
terribly much by itself, and Tom's already commented on things he
doesn't like about it. Alot of the patch here:
http://archives.postgresql.org/message-id/4A5ADFE6.6060507@enterprisedb.com
is changing heap_hot_search_buffer to have heapTuple passed in as a
pointer rather than a local var, which makes for many '.' to '->'
changes. The rest is just splitting index_getnext into two pieces.
To be honest, I'm not convinced I'm qualified to review this patch
anyway, but I was giving it a go. Of course, it strikes me that this
patch has had a fair bit of feedback from committers already, and
anything I try to add will likely end up being wrong anyway. :)
Thanks,
Stephen