Dieter> What you see is very strange.

That's strangely heartening. ;-)

    Dieter> I could only imagine that the transaction you commit is not the
    Dieter> transaction the modifications are registered with.

    Dieter> This could, e.g., happen when you modify the objects in a thread
    Dieter> different from where the transaction is committed (the default
    Dieter> transaction manager manages a transaction per thread).  However,
    Dieter> you code does not give a hint that this is indeed the problem.

That's not the case here.  Some SpamBayes apps use threads, but my
train-to-exhaustion script does not.

For more information about ZODB, see the ZODB Wiki:

ZODB-Dev mailing list  -

Reply via email to