Re: Object translation mechanism is, umm, broken.

On Fri, 2011-06-03 at 14:24 +0100, Dave Page wrote:
> On Fri, Jun 3, 2011 at 1:45 PM, Guillaume Lelarge
> <guillaume@lelarge.info> wrote:
> >
> > Hi,
> >
> > On Thu, 2011-06-02 at 08:54 +0000, Dave Page wrote:
> > > [...]
> > > This commit:
http://git.postgresql.org/gitweb?p=pgadmin3.git;a=commitdiff;h=4b8ae9d82c3b879eb02dda60908ed9590a674ec4
> > > (particularly the code in pgObject.cpp) is causing problems. I keep
> > > running into objects for which we now get messages like "Refreshing
> > > unknown object of type Replication", which I believe used to work just
> > > fine. I've just hit it with the Slony replication cluster object, and
> > > now have a sneaking suspicion it's also going to affect all the other
> > > slony object types.
> > >
> >
> > Right.
> >
> > > Aside from the fact that the code is incomplete, it's a serious
> > > modularity violation. We shouldn't have a giant conditional statement
> > > in a parent class that has knowledge of each of the different classes
> > > that might be derived from that object. Instead, each of the child
> > > classes should override a common function.
> > >
> > > Can you look at this please?
> > >
> >
> > Working on it right now.
> >
>
>  Thanks - I appreciate it.
>

So I finally have something. Can you get a quick look at it to tell me
if that's what you were looking for? I guess you'll be particularly
interested in pgadmin/schema/pgObject.cpp.

Thanks.

BTW, still two issues on this (big) patch:
 * Greenplum objects are taken care of, but I have no way to check them
 * EDB objects still have no support. I can add that, but can't check
   the code


--
Guillaume
  http://blog.guillaume.lelarge.info
  http://www.dalibo.com