It will be with a test recreating the situation outside your system
http://nhforge.org/blogs/nhibernate/archive/2008/10/04/the-best-way-to-solve-nhibernate-bugs-submit-good-unit-test.aspx

On Thu, May 5, 2011 at 1:14 AM, Mark Kharitonov
<[email protected]>wrote:

> Hi Fabio.
>
> I guess I my description of the problem is not clear. There is no code
> to clear the table. But there is a code to delete the contents of
> unreachable collections, invoked from FlushCollections method
> (somewhere in the NH2 code base, I do not have the class name at this
> moment, but it is unambiguous).
>
> Like I described, the AssociatedEndpoints is a noop property.
> Debugging the code I have noticed, that all the AssociatedEndpoints
> collections become unreachable from within the Commit statement and
> since they cover all the Endpoint_Protocol entries as a result of the
> Commit all of them are deleted.
>
> Is it more clear this way?
> Thanks.
>
> On May 5, 1:42 am, Fabio Maulo <[email protected]> wrote:
> > NH does not have code to clear a table, neither NH2 nor NH1.
> > Since NH2.1 *you* have an option to clear table/s representing a
> hierarchy
> > but nothing to clear a table representing a many-to-many table.
> >
> > That said... somebody else have cleared your tables but you can continue
> > hoping in a bug somewhere else ;)
> >
> > On Wed, May 4, 2011 at 4:36 PM, Mark Kharitonov
> > <[email protected]>wrote:
> >
> >
> >
> > > Anyone?
> >
> > > On 1 mai, 17:09, Mark Kharitonov <[email protected]> wrote:
> > > > Dear ladies and sirs.
> > > > I am not sure, if this is a bug, I am doing something wrong or this
> is
> > > > a bug in NH 2, but not NH 3 or later.
> > > > Anyway, we are using NH 2.
> >
> > > > I have two types of entities, here are the mappings:
> >
> > > >   <class name="Protocol, Entities" lazy="true" table="Protocol">
> > > >     <id name="Id" column="Id" type="string" >
> > > >       <generator class="assigned"/>
> > > >     </id>
> >
> > > >     <bag name="AssociatedEndpoints" table="Endpoint_Protocol"
> > > > access="noop">
> > > >       <key column="ProtocolId"/>
> > > >       <many-to-many column="EndpointId" class="Endpoint" />
> > > >     </bag>
> > > >   </class>
> >
> > > >   <class name="Endpoint, Entities" lazy="false" table="Endpoint">
> > > >     <id name="Id" column="Id" type="int" >
> > > >       <generator class="native"/>
> > > >     </id>
> >
> > > >     <bag name="SupportedProtocols" table="Endpoint_Protocol"
> > > > lazy="true">
> > > >       <key column="EndpointId"/>
> > > >       <many-to-many column="ProtocolId" class="Protocol" />
> > > >     </bag>
> > > >   </class>
> >
> > > > At some point, the database contains two endpoints and a dozen of
> > > > protocols - each endpoint is bound to each protocol. Meaning, the
> > > > Endpoint_Protocol table contains two rows per each protocol.
> >
> > > > Now, the server is restarted and one of the endpoints is refreshed.
> If
> > > > I throw away all irrelevant stuff (which has really nothing to do
> with
> > > > the problem), the relevant code is:
> >
> > > >       using (var session = GetSession())
> > > >       {
> > > >         var endpoint = FetchEndpoint(session, data.Id);
> > > >         using (var transaction = BeginTransaction(session))
> > > >         {
> > > >           transaction.Commit();
> > > >         }
> > > >       }
> >
> > > > Unless I do not understand something really basic, this code should
> > > > have no effect whatsoever on the database. And yet, after Commit
> > > > returns there the Endpoint_Protocol table is empty! Only the
> relations
> > > > were deleted, the entities themselves are fine.
> >
> > > > After debugging it, I have noticed this:
> > > > * Commiting flushes the entities and the collections found in the
> > > > session,
> > > > * While flushing the entities, it marks all those collections, which
> > > > are values of the mapped properties of some flushed entity - these
> are
> > > > the reachable collections.
> > > > * When flushing the collections, the unreachable ones are deleted
> from
> > > > the database.
> >
> > > > What happens, is that the fetch endpoint is reachable - of course. It
> > > > maps the SupportedProtocols property, hence all the protocol entities
> > > > are reachable as well. However, an AssociatedEndpoints collection
> > > > linked to each protocol entity is not reachable, because no mapped
> > > > property of Endpoint references it (recall thenoopaccess). At least,
> > > > this is how the  PocoEntityTuplizer extract the reachable values of
> an
> > > > entity - AssociatedEndpoints is not reachable according to it.
> >
> > > > Hence I witness these collections being deleted.
> >
> > > > Note, that once I remove the access="noop" making it the default
> > > > access and add the respective private property to the Protocol
> entity,
> > > > everything works just fine.
> >
> > > > What am I doing wrong?
> >
> > > > Thanks.
> >
> > > --
> > > 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.
> >
> > --
> > 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.
>
>


-- 
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