Re: matview scannability rehash (was Re: Drastic performance loss in assert-enabled build in HEAD)

Kevin Grittner <kgrittn(at)ymail(dot)com> wrote:

> I think I need to review the whole thread again to make sure I
> wasn't too quick to concede the point.

On a fresh reading of this, I think a large part of what is at
issue here stems from a bad name for the new bool field I added to
the RelationData structure -- instead of rd_isscannable it should
probably be called something like rd_ispopulated.  The current name
led to some fuzzy thinking on my part when it was referenced, and
both the name and a couple ill-considered uses of it probably
contributed to concerns about how it was generated and used.

I will post a draft patch today to see whether concerns abate.
Basically, the name change should help make clear that this is not
intended to be the only way to determine whether a matview is
scannable.  Second, there is at least on (and probably more) direct
tests of this field which should use a function for a scannability
test.  For 9.3, that will just wrap a test of this bool, but it
makes clear what the longer-term intent is, and help ensure that
things don't get missed when patches are written in later releases.
Third, some comments need to be corrected and added.

Hopefully it can help get us all onto the same page.  If not, it
should at least better focus the discussion.

--
Kevin Grittner
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company