2009/5/12 Krzysztof Kozmic <[email protected]>

>  Well, as I said - I'm quite new here, so I may be discovering things,
> that are obvious for the experienced ones of you :)
>
> Thanks Roger for your explanation. I know that having the two-way inverse
> association would workaround the issue, but there are still situations where
> it's not welcome to introduce the child-to-parent association in the code.
>
> Now that I know how NH works this out, I'm thinking if it would be
> appropriate (feasable) to introduce a patch that would change its behavior
> to merge the insert pet with update pet.
>

The test you should write should be enough complex to prevent ugly breaking
changes.
Try it with a complex graph with complex cascade.
In your example you should try to solve even another case.
A Animal (Pets) can exists even without a owner (Person); that is what you
are doing using unidir.
Now create a new Person and add it 3 Animals :
1) a persistent animal changing its owner
2) a NO persistent Animal (a new one)
3) a persistent Animal but without a previous owner

then save the new Person.
For Animal you should try it using:
<subclass>
<joined-subclass>
<subclass>+<join>
<union-subclass>

Let us know which is the result.

The test is more important than the fix. Right test mean right fix; wrong
test mean wrong fix.

-- 
Fabio Maulo

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"nhusers" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/nhusers?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to