There is actually a bug report that I opened for a similar case of
"phantom" modification. We shouldn't be calling validateFor* methods
in this case:
http://objectstyle.org/jira/secure/ViewIssue.jspa?key=CAY-355
Andrus
On Sep 12, 2005, at 2:47 PM, Kevin Menard wrote:
>
> On Sep 12, 2005, at 2:30 PM, Mike Kienenberger wrote:
>
>> It's not modified from the database point of view, but it's modified
>> from the Java perspective. A's toMany list is different, and those
>> changes eventually have to be propagated to other DataContexts.
>>
>> At least, I'm guessing that's the justification.
>>
>> I could see an argument for the other side too, though.
>
> Interesting. Maybe we should have two classes of modified then?
>
> I'd suggest that the case I'm talking about should never cause
> validation to occur, but I really don't know that that'll never be
> necessary. If I had an easy way to check for the type of
> modification I'm speaking about though, I could bounce out of the
> validation routine easily enough and then there's no problem.
>
> --
> Kevin
This archive was generated by hypermail 2.0.0 : Tue Sep 13 2005 - 12:53:55 EDT