There is a few regressions reported for 3.3.0: NH-3124 - a suggested fix from Julian is available NH-3125 - a fix is already merged There are a number of other issues reported on 3.3 but I haven't checked if they are regressions, or old problems simply reported for the latest version.
How long do you think we should hold off releasing 3.3.1? In my mind 3.3.1 would consist only of regression fixes. As for coming major releases there is a number of existing pull requests for further improvements of the linq provider, and some other bug fixes and minor features. Also, we have Fabios suggestions of finally going for ISet and .Net4. There was some discussion on the previous thread concerning Net4 versus Net4.5. Someone correct me if I'm wrong but the way I've perceived is that Net4+ for NH without dropping Iesi.Collection would be a pain. So instead of deciding upfront to go for Net4.5 it seems more reasonable to first aim to fix the issues that make NH with .Net4+ problematic. When that is finished, one can discuss at which point it is acceptable to require Net4.5. (I think it would be bad if we get scope creep and don't release until 2015 or so...) Do we have any sort of estimate on the effort required to drop Iesi.Collection? Amount of work and/or schedule. And perhaps most important - volunteers? If work is started on using ISet for NH4 perhaps it should be done in parallell, to have the option of releasing a NH3.4 with further minor feature work. /Oskar
