On Mon, Jul 16, 2001 at 01:20:01AM -0400, Tom Lane wrote:
> Martijn van Oosterhout <kleptog@svana.org> writes:
> > This patch re-enables support for partial indices.
>
> Applied, with some further changes by myself. Many thanks for tackling
> this ancient TODO item!
>
> A few comments for future reference:
>
> 1. Apparently you didn't run the regression tests; if you had, you'd
> have found at least one more bug (transformIndexStmt was broken for
> indexes generated implicitly by CREATE TABLE). The regression tests
> are a good thing to use for backend code development.
Oops. I kept thinking that I should but somehow never actually did it. It's
make regress, right?
> 2. I didn't like the way you did the SELECT of the partial index
> expression in pg_dump; it'd break down the first time a quote mark
> showed up in a relation name or compiled qual expression. What I
> did instead was
True. I guess I was naively assuming that the compiled form wouldn't have
any quotes in it. The compiled form doesn't actually show the actual
relation names.
> 3. A needed feature not yet implemented is for psql's \d commands to
> show the predicate when displaying info about a partial index.
I'll look at that. Can't be too difficult with the new get_expr function :)
Now, if that's all in CVS, I can rediff my remove-extend patch and submit
it, unless you beat me to it :)
--
Martijn van Oosterhout <kleptog@svana.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.
