Question says it all. Working with an old codebase that is using NHibernate 1.2. What do I stand to gain, and what will I lose, or experience in pain, as part of this upgrade? Is the total effort worth the benefit, and if so, what are those benefits? And while I'm at it, would it make sense to just move to Fluent NHibernate?
This really depends on what nhibernate functions you are using. Take the release nodes and see if anything is in there you are using. Most functions are still working the same in 3.2. You have to do some upgrades in your mapping files.
I would do one change at the time. So get it working without Fluent, and decide if it is worth doing going to fluent. You should also have a look a QueryOver and the usage of linq.
One of the most interesting new features in NH 3.2 is that you can now map with code, instead of with HBL files. This is the same feature that FNH has had for years, but integrated into the main NH project.
I have not used this myself in NH (I'm a big fan of FNH Automapping), but my reading indicates that many FNH users consider this much less mature than the FNH equivalent.
One area where FNH is still miles ahead of NH is the aforementioned Automapping. NH just does not do that yet (though there are some misleading blog posts which imply that it does).
Probably not of much value for your current project (you've already gone through the pain of setting up your HBL files), but could be very valuable for new projects. We use it in our project, and it allows us to almost change our object model at will, and let Automapping rebuild the database for us.
Lastly, as I write this there is no official release for FNH with NH 3.2. I've been experimenting with the latest FNH builds (1.3), and it's "not quite there yet". But it probably will be soon.
Nhibernate 3.2 has a mostly functional Linq provider, which is nice if you like Linq.
来源：https://stackoverflow.com/questions/8388500/what-to-be-aware-of-when-upgrading-nhibernate-from-1-2-to-3-2