autovacuum and reloptions

So I gave up waiting for someone else to do the reloptions patch for
autovacuum and started work on it myself. What I soon discovered is
that on first blush it seems a lot easier than I had expected.

On second look, however, the problem is that I seem to be
forced to declare all the autovacuum-related options and their parsing
properties in reloptions.c. This seems very ugly. I'd very much prefer
to be able to declare the options in autovacuum.c and let the rest of
the code just ignore them.

However, parseRelOptions seems inclined to barf about options it doesn't
know about. Maybe that's fine with the current usage, but I think it
would be better to leave the options in StdRdOptions alone, and have the
autovacuum options defined elsewhere, which seems to require an API
change to parseRelOptions -- though I'm not sure what's appropriate.

AFAICS this is completely uncharted territory -- the current code
understands only fillfactor as a valid option. If we were down the
route of just adding the new options just like fillfactor is currently
dealt with, the API would get really ugly very quickly.

It seems to me we should provide a way to "register" valid options, so
that autovacuum.c could inform reloptions.c what are the valid keys that
a normal option parsing should just ignore (and, conversely, what
options should it ignore when parsing for autovacuum). Thinking more
about it, it seems to me that the treatment that fillfactor currently
gets should be ripped out in favor of being registered too, somehow ...

Before we waste too much time thinking how this registering is to be
done, does anybody think that the current approach is OK and thus I
should just add the autovacuum options directly into StdRdOptions and
default_reloptions?

--
Alvaro Herrera Developer, http://www.PostgreSQL.org/
"El Maquinismo fue proscrito so pena de cosquilleo hasta la muerte"
(Ijon Tichy en Viajes, Stanislaw Lem)