Re: [HACKERS] trouble with rules

> >
> > jwieck(at)debis(dot)com (Jan Wieck) writes:
> > > I've only noticed while browsing through the code why he did
> > > comment out those things. He's comparing memoy addresses of
> > > nodes, what doesn't work any more after copyObject(). If he's
> > > not available right now, we must fix that part.
> >
> > Is there more to do than using equal() instead of a plain pointer
> > compare?
> >
> > There might be --- for example the collapsing-UNION problem I mentioned
> > yesterday is a case where using equal() allows an overly aggressive
> > optimization. Where are these comparisons and what are they for?
>
> rewriteHandler.c 1691 and 2908... and rewriteManip.c 175, 403
> and 1068. Now that I've looked closer I see that it are
> assignments. All of them have to do with sublinks and
> lefttree-aggregate issues. Shouldn't be too hard to figure
> out what's right and it will give us some additional queries
> for the rule system checks.
>
> So can someone please tell me how INTERSECT/EXCEPT works?

Are the regression tests he supplied installed yet. Should be samples
in there.

>
> I'll deuglify the code while working on it then :-}. It's
> really hard to read (must have been written in a 120 char
> wide window or so).

Yes. Just run pgindent on any files you want, or I will do it if you
tell me where. I ran it on some optimizer file. I can easily do a
whole directory if no one else is working in there.

--
Bruce Momjian | http://www.op.net/~candle
maillist(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026