Forwarding message from ibatis-user-java mailing list, as per Clinton's suggestion.
Philippe > I'm not sure adding the functionality in the DAO would be > very efficient. Maybe I'm missing your point, but consider > the queryForList() method, it would require iterating on the > results to create a proxy for each item in the List... It > would be much more efficient if the proxying mechanism (or > whatever) was applied upon bean creation. > > Personally, I think that having an external bean > instantiation mechanism would make iBatis a much more > flexible framework... For example, it could help unit tests > for statements (create test beans to test set and get calls), > all sorts of caching, initialisation or security processes > could be applied but most of all, statement result classes > could be defined as interfaces. To me, the last point is the > most interesting... > > Would anybody else find it useful to have interfaces instead > of classes when mapping statement results? What does the > iBatis dev team think? > > Thanks, > Philippe > > > -----Original Message----- > > From: Clinton Begin [mailto:[EMAIL PROTECTED] > > Sent: January 15, 2005 11:23 AM > > To: ibatis-user-java@incubator.apache.org > > Subject: Re: Bean Factory is available? > > > > > > Hi Tak, > > > > No, that's not something we support. But that sounds like > > something that would be best done by your DAO class. An > > alternative would be to put a dynamic proxy behind the > > SqlMapClient interface and proxy your classes on the way out > > of the queryForXxxxxx methods. > > > > Cheers, > > Clinton > > > > > > On Fri, 14 Jan 2005 08:08:34 -0500, > > [EMAIL PROTECTED] > > <[EMAIL PROTECTED]> wrote: > > > > > > Hi, > > > > > > Is there a way to set bean Factory to resultMap instead of setting > > > bean class directly? I mean, > > > <resultMap id="result" class="my.model.Person"> > > > to something like > > > <resultMap id="result" class="my.PersonFactory"> > > > > > > so that I could create proxy on top of my class. > > > > > > thanks, > > > Tak > > > > > >
