RE: drop table and pg_proc

> This is just one instance of the generic problem that we don't enforce
> referential integrity across system catalogs.  Since this issue has

Wouldn't be easy to do for views (rules) anyway - table oids are somewhere
in the body of rule, they are not just keys in column. Also, triggers are
handled by Executor and we don't use it for DDL statements. I think it's ok,
we have just add "isdurty" column to some tables (to be setted when some of
refferenced objects deleted/altered and to be used as flag that
"re-compiling"
is required) and new table to remember object relationships.

Guys here, in Sectorbase, blames PostgreSQL a much for this thing -:)
They are Oracle developers and development under PostgreSQL makes
them quite unhappy. Probably, work in this area will be sponsored
by my employer (with me as superviser and some guys in Russia as
developers), we'll see.

Vadim