Re: Possible bug in pg_settings/pg_depend

On Thu, Jan 13, 2011 at 2:04 PM, Joel Jacobson <joel(at)gluefinance(dot)com> wrote:
> 2011/1/13 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
>> Yes, probably.  It's certainly possible to have the same linkage occur
>> with different deptypes.  We don't try hard to avoid dups because they
>> don't matter.
>
> "with different deptypes", yes, but in this case there were two
> linkages of the same deptype.
>
> Just seems a bit strange I only found one such in the entire database,
> smells like some kind of bug, but might not be, I dunno, just thought
> it was worth investigating a bit, but if you're sure about it I of
> course trust you.

Instead of trusting him, you could investigate why it happens. A
quick test shows this eliminates both dependencies:

drop rule pg_settings_u on pg_settings;

It appears that both of the dependencies in question are from that
rule and pointing to pg_settings.name, and it looks like that rule
mentions the name column of pg_settings twice. With a little further
experimentation you can probably tease out whether each of the two
mentions produced a separate dependency... my guess is "yes".

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company