Patch to add support for partial indices

[Note to list maintainer: I sent this yesterday but it got stuck. I've
resent it now. If you find it in the queue, you may safely delete it.]

This patch re-enables support for partial indices. This includes:

* Updates to documentation
* Added pg_dump support
* Corrections to the grammer to accept the CREATE statement
* Update old broken code each index type
* Fixes to VACUUM so that it handles indices in general better
* Updates to the optimiser and cost estimators to handle them
* Initial steps to removal of EXTEND INDEX statement
* Patches the regression tests

Note that Tom Lane has already taken the portion that deals with vacuum as
part of his new vacuum code.

As for the extend index statement, it was decided to remove it since it was
not really flexible enough, so in general you'd be better off just dropping
the index and recreating it.

For further discussion, see the pgsql-general list. The patch against
anonymous CVS as of a day ago or so.

Many thanks to Tom Lane for tips and pointers during the initial revisions
of this patch.

http://svana.org/kleptog/pgsql/partial-indices-7.2.patch
--
Martijn van Oosterhout <kleptog(at)svana(dot)org>
http://svana.org/kleptog/
> It would be nice if someone came up with a certification system that
> actually separated those who can barely regurgitate what they crammed over
> the last few weeks from those who command secret ninja networking powers.