[Zope] Is data.fs corrupted?

[Stefan H. Holek]

On Fri, 10 Nov 2000, Chris Withers wrote:

> Chris McDonough wrote:
> > 
> > I'm going to take a guess.  It would appear AFAICT that the ZDiscussions
> > product installed an object which it hung off the root object (the
> > application object) when it was first imported.  
> 
> Don't rememebr seeing anything like this in ZDiscussions when I was
> looking over the code...
> 
> > The point being that I'd bet we'd find a programming error staring at us in
> > this sequence of events if we looked hard enough.  Though it's tempting to
> > call it a "ZODB corruption" problem, if you really get down to the lower
> > levels, it's likely an application-level problem.
> 
> Maybe it's a Zope-level rather than ZODB level problem? Maybe some of
> the ZClass machinery or product registration machinery is screwing
> things up?
> 
> Given that people have reported this from more than one product now, I'd
> be inclined to think it's not random bits of bad code in products. Maybe
> there's a method/pattern which used to be okay/recommended which is now
> causing problems?

FWIW, I can also report this problem with ZDiscussions. In my case it
stemmed from not properly deleting the ZDiscussion product before moving
my Data.fs from 2.1.6 to 2.2.1 (I think). The symptom then was beeing
unable to create ZSQL Methods (!) because of a missing
../Products/ZDBase/ZDiscussions.py file which I "fixed" by installing
ZDBase again. I can however NOT import the ZDiscussions.zexp
anymore because of "Duplicate Class Ids".

Oh well,
Stefan