FYI: I discussed this with Mike. As a workaround, I'm going to remove
the fetchLimit on my query, then iterate through the results and remove
all the objects marked for deletion... this should yield the kind of
results my application expects.
Gili
Gili wrote:
>
> In my application I have something like this:
>
> ImageSpecification[id, width, height]
>
> ImageSpecification.getCanonicalInstance() issues a query and if a
> given [width, height] combination already exists in the database,
> returns an instance of that object. So for example, if specification1
> and specification2 both have the same [width, height], specification1 is
> already in the database and I invoke
> specification2.getCanonicalInstance() then I'd expect it to return
> specification1.
>
> Now ... here is where the problems begin. Currently, Cayenne treats
> a DataContext as a scratch-pad (to borrow Mike's terminology), not a
> transaction. If I execute the following code:
>
> context.deleteObject(specification1);
> specification2 = specification2.getCanonicalInstance();
>
> then specification2 will end up equal to specification1. Why?
> Because as far as Cayenne is concerned, specification1 has not yet been
> removed from the DB (we haven't yet invoked commitChanges()) and so it
> is a valid result.
>
> I'm expecting Cayenne to be a bit smarter with queries and treat the
> DataContext as a transaction, not just a scratch-pad. This means that
> when specification2.getCanonicalInstance() is invoked, Cayenne should be
> smart enough to realize that from the point of view of the current
> context "specification1" has been deleted, and so it should not show up
> in the query results. Implementing this change should be easy enough (I
> can contribute the patch).
>
> Is it possible/reasonable to modify Cayenne to act this way? From my
> point of view, it's an ease-of-use issue. The user can explicitly check
> deletedObjects() but he shouldn't have to (is there a benefit to having
> to check explicitly?) What do you think?
>
> Thanks,
> Gili
-- http://www.desktopbeautifier.com/
This archive was generated by hypermail 2.0.0 : Thu Sep 08 2005 - 12:35:07 EDT