I'm noticing that the mapping file for copletinstancedata specifies that AttributesFieldHandler should be used. I haven't used Castor much, but it seems I could make a simple change to that class to make sure it gets only the attributes from copletinstancedata when necessary. If so, would that fix the problem? I could then make the copletinstancedata getAttributes and getAttribute methods use the lists in copletinstancedata and copletdata. That would minimize the amount of code I'd need to change in other classes.
I guess the real question is, does Castor retrieve the data through the AttributesFieldHandler class when building the XML? Ralph Carsten Ziegeler said: > Hmm, I really can't remember everything. Now one problem was that if > you save the objects to XML, you can't save all available attributes > for the instance. You have to filter and only save the attributes that > are part of the instance and not those of the coplet data. > > An additional problem is removing of attributes etc. What do you do if > the attribute you want to remove is not an attribute of the instance? > Do you delegate it to the data object? And so on. > Now, rethinkink everything, perhaps the simplest solutions is to leave > everything as it is and only add a "get value from all attributes method", > which is just a getter method that first looks in the instance and then > in the coplet data object. But for setting and removing everything could > be left unchanged. > > HTH > Carsten >
