Hi. I started using Wicket recently, and found have found it excellent. However, although my development had gone very smoothly early on, recently I've started to hit a bit of a wall because of some issues around Wicket models (as well as my decision to switch to Wicket 1.3 from 1.2). I've already RTFM rather extensively (Karthik G.'s book, plus lots of stuff on the wiki) but I'm still very puzzled about a number of points. If anyone could help with some of my questions below, I'd really appreciate it!
My over-arching question at this point is: Is it a requirement that a model's target object be Serializable? I ask because I've been running into a number of different issues and questions regarding models and serialization: 1) I tried to use a LoadableDetachableModel, partly as a way to avoid making my data object Serializable. But I'm wondering if that's not correct. Take for example a data object with gettters and setters for the following fields: id (int), foo (string), and bar (string). Using LoadableDetachableModel, I can load the object on demand by id - and so all I need to do is hold the id in the LoadableDetachableModel, and can discard the object itself. Thus the object doesn't need to be serializable. However, lets say that a form submission by the user sets the foo property to "foo", and bar property to "bar". And let's say further that "foo" is a valid value, but "bar" fails validation for some reason - which will keep the user on the same form page. In this case won't a logic error result? Since I'm not retaining (or serializing) my instance of the data object, but rather loading it fresh from the database each time, won't the "foo" value disappear? Because when the model needs to access the target object, it wouldn't be accessing a "working" copy that had the value set to "foo", but rather a pristine copy from the database that had an old value? 2) How am I to handle the situation where I want to use a new (not yet saved) instance of a data object as the target for a model? Since the object is not yet saved to the database, I can't use a LoadableDetachableModel to load it fresh from the database every time. And I can't just have the Model create a new instance each time either, since it looks like repeated form submissions would wind up discarding old values? Is there any way to handle this situation without making my model target serializable? 3) You might be asking yourself: why doesn't he just make his data object serializable? And I guess I'll do that if I have to. But I'd like to avoid it if possible - because it means a lot of extra work. It might sound trivial to make an object serializable, but it's not. To do that, you need to make sure that its entire object tree, including every object it references, is serializable as well. This will not be a trivial task in my code, as there's numerous objects that I would need to make serializable. I'd like to avoid this if I can. So is it possible for a model's target object to not be serializable? Or is there some way to turn off this requirement? I noticed that in 1.2, when you use DEPLOYMENT mode, you wouldn't get serialization errors. Has that changed for 1.3? I'm fairly certain that I won't be deploying Wicket with clustering (or at least not session clustering like with Tomcat or Terracotta - I might cluster at the database level) so why should I need to be forced to make all my data serializable? 4) If serialization is now a requirement for model data objects, then it would seem to me that it'd no longer be true that Wicket can be used with POJO's. To me a POJO means a java object that doesn't need to extend any particular subclass or implement any particular interface - you can integrate it in to a framework as is without any coding changes. But if model data objects are forced to implement Serializable (and so are all of it's referenced fields ... and all of THEIR referenced fields ...) then they're not POJO's any longer. There'd then be a strict requirement on how those objects must be implemented, right? Can someone please help clear up some of this confusion? As things stand right now, it looks as if I have no choice but to change large sections of my code to make it all serializable, and I'm less than thrilled at the prospect. And if that's the case, then it'd seems to me that there's a fairly hefty cost that I'm forced to incur in using Wicket. I hope that I've just misunderstood things here, and this is not the case. Would appreciate any help and clarifications that people can offer this Wicket noob. TIA! DR ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Wicket-user mailing list Wicket-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wicket-user
