>Sergio,
>
>I think we talk more or less the same thing.
>To me it looks like we could not express / read fine enough in english.

>The dangling datasession was not there because of an error (normaly the app would stop anyway), it was just because of bad reset.

Yeah, I was talking about listeners in general not about your particular report problem :)
Possibility of those datasession reset mishaps is exactly what I had in mind. If additional session object was holding frx data and there were methods in place there to give you what you want then actual changing of session does not have to happen at all! Just like when exchanging data between two forms with private datasession.

Unless this was impossible due to some engine internals that we have no clue about, I believe this would lead to much safer listener subclassing.

Or it is one of those situations when senior celeagues used to tell us;
You don't always have to agree with the way the things are done; sometime you just have to believe that there was solid reason why things were done certain way :)

Cheers :)
Sergio