[ZODB-Dev] ZODB and new style classes?

[Mike C. Fletcher]

Am I sure it's invalidation of objects that's the problem:
	No, not really.  That's the conclusion Shane came to, and given my 
understanding of ZODB's internals, it sounds about right.  More 
precisely, it seems that some of the objects were holding onto 
references across the invalidation (the only _v_ variables (used for 
cyclic references) were being explicitly reset, so that shouldn't happen 
with functional invalidation AFAICS).

	Oh, and it's a single-threaded app at the moment, so I'm not sure how the 
transaction-boundary problem could be showing up.


I can try to boil down the problem to a test case some time, (time being 
in short supply around here at the moment).  I've got 8 or 9 levels of 
abstraction on top of the database (including the ones in the GUI), so 
it will take a bit of work to get a minimal sample drilled out.

ZODB4 Readiness:
	I'm distributing it to "internal" clients for testing of the interface 
(not actual use), but will not be delivering "to-be-used daily" code for 
a month or two.  Hopefully that time frame will see a stable ZODB4, if 
not, sucks to be me :) .

Enjoy,
Mike


Jeremy Hylton wrote:
...
> closely.  Are you sure that it's a problem with invalidating objects?  It
> sounded to me more like a problem with many of your threads not hitting
> transaction boundaries regularly.  Now it may be that the explicit sync
> method isn't working correctly in ZODB 4, and your code is the first code to
> exercise it.
> 
> If you've got the time, it would be very helpful if you could boil down the
> problem of a simple repeatable test case.  Then you'd force me to fix it,
> and we can add it to the standard test suite to make sure it continues to
> work in the future.
> 
> 
>>Of course, that requires using ZODB4, which isn't a recommended activity
>>at the moment I'm told :) .
> 
> 
> You put the smiling their, so I think you've got the right idea, but I feel
> like I should clarify.  There's no problem with using ZODB 4, but I wouldn't
> trust it for production code and I certainly wouldn't distribute code using
> it to other people yet.  I'm very happy to have non-Zope users experiment
> with and test the ZODB 4 code.  There are a lot of internal changes, and we
> have a number of significant API changes planned for the near future.  Such
> helpful to get feedback from a wide range of users.
> 
> Jeremy
> 
...
_______________________________________
   Mike C. Fletcher
   http://members.rogers.com/mcfletch/