That's cool for sure but what to do if commit failed?
From: [email protected] [mailto:[email protected]] On Behalf Of Belvasis Sent: Monday, June 22, 2009 6:27 PM To: [email protected] Subject: [nhusers] Re: deep copy aggregate root There you can decide if the update to the db should be done or not. It gives you the entity and it's state and oldstate so you can rollback any changes if you veto the event. 2009/6/22 Yevhen Bobrov <[email protected]> IPreUpdateEventListener? How that can help? From: [email protected] [mailto:[email protected]] On Behalf Of Belvasis Sent: Monday, June 22, 2009 5:47 PM To: [email protected] Subject: [nhusers] Re: deep copy aggregate root Why not use the IPreUpdateEventListener or simply an transaction? 2009/6/22 FrederikGheysels <[email protected]> Well then, why not implement some undo - functionality that is based on reflection & serialization ? Create a BeginChanges, AcceptChanges and CancelChanges on your object. When you display the object using databinding, you call BeginChanges. When you want to cancel the changes, call CancelChanges so that the object's state is reverted ? In the BeginChanges method, you can find all fields using reflection, save their state in a hashtable, etc.... When you Cancel the Changes, get the hashtable, loop over it, and set the values of all the fields back to the value that you find in the hashtable for that field ... Offcourse, this is not so simple, especially with collections (I don't even know if it is possible out of the box, I think that you will have to create a user-type for NHibernate). On 22 jun, 15:39, "Yevhen Bobrov" <[email protected]> wrote: > Frederic, will all respect it is not viable to delay data binding in big > enterprise application with myriad of objects. This will require an enormous > amount of code to implement temporal storage of state for each object or > collection or reference. This what data binding for - to ease manipulation > of object state. Either you will need to duplicate all validation logic and > all of behavior in your intermediate objects or dumb down screen to stupid > "input form with DTO's underneath", as in this scenario you can't touch real > objects before commit. > Delaying modifications to object state is only a viable option for very > small apps. > > > > -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf > > Of FrederikGheysels > Sent: Monday, June 22, 2009 4:23 PM > To: nhusers > Subject: [nhusers] Re: deep copy aggregate root > > Well, you can implement an N-level Undo functionality (like I referred > to in my post), or , you can 'delay' the databinding on your WinForms > app to get a solution for your problem. > My first idea was to implement N-level undo functionality (which I've > done) on my base-entity class, but after a little bit of playing with > it, I thought it was overkill. (Some of my entities have associations > to other entities, etc... so a lot of data would be serialized). > Next to that, I didn't want to make all my entities serializable, so I > came up with another solution; that is: delaying the 2-way > databinding. > This problem is not directly related to NH, although I do find it a > pity that NH commits the changes you made to an entity when you flush > the session, even if you did not call 'SaveOrUpdate' for that entity. > > On 20 jun, 17:35, Yevhen Bobrov <[email protected]> wrote: > > It is. It is not about NH problem but about what happens when changes > > are applied directly to the object and there is need to cancel them. > > This is what happens in usual WinForms application. > > > On 20 июн, 18:15, Fabio Maulo <[email protected]> wrote: > > > > 2009/6/20 Yevhen Bobrov <[email protected]> > > > > > BTW, Frederik talks about such kind of problem here > > >http://fgheysels.blogspot.com/2009/06/winforms-databinding-on-cancell... > > > > That is not a post about a problem, that is a post about a UI solution > and > > > as you can see there is no reference about the persistent-layer and no > > > reference about the usage of Clone. > > > Frederik had the good spirit, even showing the n-level-undo capability > of > > > CSLA (where, another time, the persistence-layer has nothing to do). > > > > We'll wait your example showing your problem in detail; after have it we > can > > > talk about a solution if you can't find it. > > > > -- > > > Fabio Maulo > > __________ Information from ESET Smart Security, version of virus signature > database 4174 (20090620) __________ > > The message was checked by ESET Smart Security. > > http://www.esetnod32.ru > > __________ Information from ESET Smart Security, version of virus signature > database 4174 (20090620) __________ > > The message was checked by ESET Smart Security. > > http://www.esetnod32.ru __________ Information from ESET Smart Security, version of virus signature database 4174 (20090620) __________ The message was checked by ESET Smart Security. http://www.esetnod32.ru __________ Information from ESET Smart Security, version of virus signature database 4174 (20090620) __________ The message was checked by ESET Smart Security. http://www.esetnod32.ru __________ Information from ESET Smart Security, version of virus signature database 4174 (20090620) __________ The message was checked by ESET Smart Security. http://www.esetnod32.ru --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
